"Challenge" vs "Difficulty"

I was thinking about Challenge vs Difficulty (particularly in cooperative board games) recently. At a glance, they sound like the same thing, but if you think about it, they're really not.

Definitions

I looked up "challenge," and the definition in this context is: "a difficult task, especially one that the person making the attempt finds more enjoyable because of that difficulty," while "difficult" is simply defined as "hard to do, requiring much effort." For the purposes of this post, I will sum those up as follows: 

Difficulty in a game is the extent you're unlikely to succeed.

Challenge in a game is the extent to which overcoming the difficulty is fun or rewarding.

I've found that these terms come up a lot in the context of cooperative games, where the players need some sort of AI or algorithm to play against rather than the cunning and guile of a human opponent. A good cooperative game is challenging, it sets a task for the players, and they have fun trying to overcome obstacles to accomplish that task. One of the nice things about these games is that an optimal set of plays is not clear, and the whole point is to make that optimal set of plays, or close enough, that you achieve the goal of the game in the time allotted.

However, some cooperative games aren't challenging so much as just difficult. In those games, the chances of losing even with an optimal set of plays is too high, so it doesn't feel as fun or rewarding when you win.

The Components of Difficulty

There are 2 major things that can make a game more difficult - they both reduce your likelihood of winning: Increased depth, and decreased fairness, where depth is the amount of good play required to win, and fairness is the degree to which that good play determines the outcome.

• In a fair game, the outcome is less often dominated by factors outside your control. If you play well, you'll win more often.
• In an unfair game, too often the outcome is outside the player's control. There's a significant chance you'll lose even if you were to play optimally
• A deep game has a high skill ceiling. You have to build up to the point where you can hope to play near-optimally
• A shallow game has a low skill ceiling. You can be confident your play is near-optimal without too much time, effort, or study

In order to examine this more closely, I thought I'd make a graph of Depth to Fairness. Plotting a game on that graph could allow us to visually see some of these relationships and make sense of them:

WAIT! Why am I using UNfairness on the bottom axis instead of Fairness? Difficulty comes from both depth and unfairness. A deeper game is more difficult to win because it requires better play. A more unfair game is difficult to win because despite good play, you might lose due to chance. So both of those things can increase difficulty. I use Unfairness for the axis so that moving away from the origin in either direction makes the game harder.

Now we can locate games on the graph, draw a vector to them from the origin, and the length of that vector relates to how hard the game is to win: 

Magnitude = Difficulty

"Challenge," then, is how fun or rewarding it is to overcome that difficulty. In this graph, the slope of that vector (rise over run: depth over unfairness) relates to the amount of challenge the game has to offer. The deeper the game, or the fairer the game, the more rewarding it is to overcome its difficulty, and therefore the more or better challenge it presents.

Practical Application

Let me lay some ground rules on what this graph applies to, how it can be helpful, and what it's limitations are.

For one thing, you'll notice that there are no numerical values on these axes. I am not sure how these aspects could possibly be measured! Also, those gray, labeled boxes are completely arbitrary, and can be deleted or redrawn wherever you'd like, so a single data point on this graph is meaningless. This graph only allows us to make relative comparisons between multiple games. As soon as you get 2 data points, you can begin to compare them and see, in a qualitative sense, which is more Fair, which has more Depth, which is more Difficult, and which is more Challenging.

This information might work best for 1-player games, solo modes, and co-operative games, when the "opponent" has a consistent skill level. It does not make as much sense for a multiplayer game, where the difficulty depends on your opponent's skill level as well - although you might be able to gain insight into questions like "how hard/challenging is it to play various different games vs Steve?" If the opposition is fixed as that particular opponent for each game, then I think the model will still allow comparisons, which could be fun to do amongst a group of friends who play vs each other a lot. 

Similarly, I think this could be used to compare factions or characters in an asymmetric game - not in general, but if you look at various factions/characters vs a particular faction/character, then you might be able to glean some useful balance (or strategic) information about matchups.

Something I think might be a stumbling block in reading this graph is that we're not talking about the outcome of a particular match here. We're talking about the difficulty of the game in general -- overall win rates, not whether you win or lose this instance, or this play of the game.

A Worked Example

Pandemic is a popular cooperative game for 2-4 players, and you can adjust the difficulty from "easy" to "expert" by using 4-6 Epidemic cards in the deck. The game is obviously harder with more Epidemic cards, but I've also observed that the game gets harder the more players you have due to logistical concerns. Let's take a look at some of the configurations you can play Pandemic in, and plot them on this Difficulty chart:

The way I drew that I'm saying that 2p-expert and 4p-easy are approximately the same difficulty, but in 4p it's more rewarding to overcome that difficulty

Good Difficulty vs Bad Difficulty

We have learned that Difficulty comprises two aspects: Depth and Unfairness. One of those increases Challenge, and the other does not, but both increase difficulty. I don't think I'm out on a limb saying difficulty that increases challenge is "Good Difficulty," and difficulty that does not increase challenge is "Bad Difficulty." 

There do exist cooperative games where added difficulty does not seem to increase the challenge. I remember when Ghost Stories came out, it was the talk of BGG.con that year, and all the buzz was that the game was really difficult. After finally playing Ghost Stories I remember thinking "I'm not sure a higher chance you can't win is the kind of 'difficult' you want in a game." Of course, that hasn't stopped the game from being incredibly popular over the years!

I've recently gotten to try The Princess Bride Adventure Storybook Game, and it strikes me as a game with a low skill ceiling. It's fairly easy to figure out what to do, and you can go through the motions to do it, and then you see if you were able to finish in time. You can reach the point where you can achieve optimal or close-to-optimal play with some confidence, but even with optimal play, often whether you win or lose a chapter seems to come down to chance. In this game's defense, I've only played it with 2 players, and it's possible that, like Pandemic, the game is more challenging (or at least harder) with more players.

To be fair, that's always the case with cooperative games -- given optimal play, there's still a chance you'll lose the game. Most (all?) cooperative games have that dynamic, lest they become "solved," and that's probably a good thing. But perhaps each game has an Unfairness threshold, a percentage chance that you lose anyway, even with optimal play, that is acceptable. Under that threshold, the game could be considered "Fair" (or fair enough). Above that threshold, the game could be labeled "Unfair." In a good cooperative game, the amount of Unfairness would be closer to that threshold, whereas in a bad game, the Unfairness might be much higher than the threshold. Unfortunately, I don't have a good feeling for where that threshold should be, and it is almost certainly different from game to game.

Conclusion

The best case scenario in my mind is a deep/fair game, where winning correlates to amount of good play, and the chance you lose even with good play is within acceptable limits. This offers more challenge, and it's therefore more rewarding and more fun to overcome the difficulty. A bad game in my mind is one that is shallow and unfair, where winning does not take much good play, but does require a lot of luck.

I'll try to keep this difference between "challenge" and "difficulty," and this Depth-to-Fairness ratio in mind when designing, especially when working on a cooperative game. For example, if I ever finish up Alter Ego, maybe I'll apply these lessons to ensure that to the extent the game is difficult, it provides "good difficulty," not "bad difficulty." It's important for games to be challenging, not just difficult!

Post Script on Effort

Another thing that could be considered is when a game requires more effort: doing lots of simple math, moving pieces around, grinding actions, that sort of thing. Busywork. Since these things don't require any particular skill (except maybe stamina to endure them), I'm not sure they make a game harder to win, they just make it more annoying to play. They could be said to reduce the Challenge, because they make it less fun and rewarding to overcome the difficulty of the game, but they do not themselves contribute to the difficulty. Therefore I don't think we need to consider Effort in this model.

Designing with competitive vs NON-competitive play in mind

No catchy title this time, just wanted to talk about how we should design games that hold up to competitive play… AND to NON-competitive play.

I've said before that all things being equal, games are better if they hold up to competitive play. That is to say that they don't break down when a player "tries to win." For many games, that's not strictly necessary, for example, many party games are not really played "to win," but just to facilitate a fun time. That's well and good, but my point still stands: the game would only be better if it did not unravel when one or more players do play competitively.

I've stumbled across a new observation that is related to this, and may be even more important. Not only should a game strive to hold up to competitive play, but it must also hold up to NON-competitive play! By this I mean simply: if one, some, or (worst case) all of the players do not aggressively pursue the winning condition, the game must not stop working, and it must still progress toward an end.

Case Study: Apotheosis

This has come up on my current design, Apotheosis. In every playtest of that game, at least one player (myself, if not my friend Dave) would treat the game as the race it was intended to be. We both strive to do tier-2 adventures as quickly as possible, as they're much more efficient then tier-1 adventures, and we press as hard as we can toward reaching the win condition (the end of a track). As a result, the duration of the game was always acceptable, and I thought the game was in pretty good shape.

The other day I had a test with 3 other players on Tabletop Simulator, and I decided to sit out and help facilitate instead of playing because TTS is kinda fiddly, and I thought it would go faster that way. This turned out to be fortuitous because it revealed what I'd consider a fatal flaw in the game: all three players went for even advancement to obtain the "consolation" level-ups I'd added, effectively spending too many turns building up rather than pressing to get to those tier-2 adventures and racing up a track. As a result, perhaps not unexpectedly, the game dragged on and ended up taking about 2 hours -- fully twice an acceptable duration!

Thinking about this problem is what turned me on to the axiom above: games must hold up to non-competitive play. First time players will not necessarily notice that you're intended to push up the tracks as fast as you can. In fact, the current incentives kind of suggest the opposite. And many players just play games to explore their systems, and don't try doggedly to achieve victory. Therefore, it is definitely appropriate to address this game-dragging problem in some way.

Brainstorming solutions 

My first thought was to remove or reduce the "consolation" level-ups. I'd added them to ensure that simply picking one track and ignoring the other two wasn't necessarily the best path to take. To be honest, I'm not sure they were really necessary in that respect, but I did like having a reward for even advancement in a game where the goal is to advance any 1 track to the end. While that might have reduced the problem, it would not have eliminated it, as new or bad players could still dilly-dally too long and make the game drag. This might not be a practical problem, but it's certainly at least a theoretical one. The game should naturally push toward an end, no matter how players decide to play.

My next thought was to make some of the rewards on the tracks "1st come, 1st served" to encourage players to race for those. This might be good to do, but I'd also like to see players get those rewards more often, so limiting them might not be great after all.

Finally, the 3rd thing that came to mind was the biggest, and possibly best solution: add a game end condition that would trigger when players dilly-dally. Such a game timer would keep the game from dragging by definition, if players don't progress the game themselves, it will still come to an end. In general I'd say this is an obvious choice, except in Apotheosis, the win condition is reaching the end of a track. So what happens if the time runs out and nobody has achieved the win condition? How do you decide who wins? In some games it's easy to assess relative progress, and award the game to the player who's closest to winning. But here that is thematically odd because topping out a track is supposed to represent a big, momentous event.

Another option is to say that if time runs out, then nobody wins. This is an interesting thought, however it may be out of place in this type of game, and it's likely to make for a bad first play experience if players all lose in the first game.

Settling on a solution? 

What I have decided to try is this: add a game timer (the king will return, and once he does, your opportunity to steal the throne will be gone!), and if you win before time runs out, great (the king returns to find you on his throne, controlling his army, or backed by a demon, or with his court turned against him)! But if nobody has won by the time the king returns, then the player with the best reputation across all of the guilds (evenly advanced up the tracks) is the winner.

Tangentially related: game end dynamic 

In addition, I'm considering a variable, slightly random game length, something like this... When the adventure deck runs out, the king is ALMOST home. Put a King marker on a short (6 space) track that's revealed under where the deck was. Give the triggering player a marker as a reminder, and for the rest of the game, after that that player's turn each round, roll a die. On a 1-3, advance the king marker 1 space. On a 4-5, advance 2 spaces. On a 6, advance 3 spaces. Therefore when the deck runs out, you have 2-6 turns left to win by getting to the end of one of the three victory tracks. This way, if you don't think you can reach the end of any the tracks by the time the king returns, or if you think someone might beat you to it, you can advance your track markers more evenly in case nobody else achieves the win condition in time either. This gives you something to do if you feel you can't win, and it might also extend the tension (until the last minute at least) even if you know you can reach a track end before anybody else.

The game must communicate its dynamics to the players

This brings me to another recent observation, which we could put down as another axiom: The game must communicate its dynamics to the players. I have talked about this in the context of the "Alpha Player Problem" and what I call "Solitaire by Committee", or committee-style cooperative games, but it applies more generally as well. In SbC games, this axiom suggests that it's important to let players know that the game is not about making your own choices and having full agency with some incentive to coordinate with or help the rest of the group, rather in an SbC style game, the whole point is to have a little committee meeting to decide on a course of action, and then do that.

For the new perspective alternate end game trigger in Apotheosis, this axiom would suggest that it's important to make clear the "most evenly advanced" win condition is a secondary condition, and that the primary and most common way to win will still be by reaching the end of a track. Without clearly communicating this, I can see how it'd be very easy for a player to assume both win conditions are equally viable to go after, and I can just see reviewers now complaining that "the game is not well balanced, not all win conditions have the same win rates" (duh, they're not supposed to!)

I'm not exactly sure how to go about that communication outside of explicitly stating it in the rulebook, which is not ideal by itself, because it is too easy to overlook or forget about.

TL;DR summary 

Games are better if they hold up to competitive play, but they MUST hold up to non-competitive play. Don't allow your game to drag on or fall apart if players don't pursue victory as aggressively as you expected them to.

Modular vs Integrated expansions

Having recently done an expansion for Crusaders, and having done 3 expansions for Eminent Domain and one for Isle of Trains as well, I have noticed a particular style I like to use when creating expansions, and I've identified two distinct types of expansions: Modular and Integrated.

Modularity

A modular expansion is one with several distinct modules that can be added in various combinations. These can be good because they allow the players to use the modules they enjoy, and simply ignore the ones they don't care for. It allows players to customize their experience. Undoubtedly, this is something that some players will appreciate.

However, it also requires playing each module at least once to determine if it works well for your group, and it requires some effort to curate the expansion content to provide that perfect customized experience. For a group that loves a game and plays it all the time, one that's willing to put in the effort and plays to find the perfect combination of modules, this could be fine. But in today's market, I wonder if players will put in that time and effort? Or would they prefer to just buy a curated expansion, where the "best combination of modules" is the only thing provided?

Another down side to a modular expansion is that because of that modularity, the pieces of the expansion may feel bolted together and disparate rather than feeling like a seamless experience.

Integration

An integrated expansion is one where the new parts play off of each other and off of the original content. When done well, this type of content fits seamlessly with the base game, and it can be difficult to differentiate the new stuff from the old.

This loses the opportunity for customization provided by a modular expansion, but it gains the curated feel, and doesn't require any work on the part of the player to get to the "best" configuration.

Another down side to integrating an expansion is that it may be more difficult to remove the content when you want to play the game with new players. For an insular group who plays the same game many times, an integrated expansion may be preferable, while for a group where new players come and go with each play, it may be more difficult to pop the expansion content in and out.

My personal preference

It's probably clear in the above paragraphs that modular expansions aren't my preferred format. Well integrated, expansion content fits together with the base game in such a way that it feels like it was always there, or like it belongs. I find something attractive in the thought that the expansion integrates so well that a new player might not be able to identify the new content from the old.

Looking at the expansions I've done, I think it's clear I've attempted to go for integration over modularity:

In the Eminent Domain expansions, you simply add the new tech cards, shuffle the new planets into the deck, and the Fleet, Mining, and Political Influence tiles tie it all together.

When I first played Isle of Trains: All Aboard (which hasn't been published... yet) with the publisher, they were skeptical going into it that the game really needed an expansion. After we played, they said they were surprised how well integrated the expansion was, and that it was not obvious the expansion content wasn't just part of the base game in the first place.

Crusaders: Divine Influence is about to ship from China, so not many have played it yet, but my goal with that one was to seamlessly integrate the new content with the old. I replaced the Influence action with something more involved, and I added new buildings and a few factions. If you know the rules of the base game already, then playing with the new content should be easy to pick up. Learning the game for the first time with the expansion content included will be a bit more complicated than learning just the base game, but I have had some success teaching the game that way most of the times I have played the expansion with strangers.

Conclusion

There might be a temptation to make expansions modular, because as a designer it can be easy to think that's what players want, and it abdicates the responsibility of curating the expansion to the players. My advice for designers is to consider the game you are expanding, and that game's audience. In this day and age, with thousands of new titles coming out every year, is modularity really the best format for your expansion? Or would your audience be better served with a well integrated, curated expansion? For most of the games I work on, I think it's the latter.

Making steps forward ("killing your darlings")

In a few of my games I have started with a specific mechanism or idea that, over the course of the game, has had to drop out, or change significantly. I'd venture a guess this happens in many, if not most, designer's games.

For example, in my first published title, Terra Prime, I began with a "great" idea for a supply/demand mechanism for the value of resources you deliver. Based on Power Grid's fuel market, you would get paid more for delivering Bluium if there hadn't been any delivered yet than you would if there were already some Bluium sitting there on the ships, ready to send back to Earth. Every once in a while, the ships would take off, and the resources on them would clear off, resetting the demand (and therefore the price). Each resource was supposed to have a different curve - maybe Yellium was worth 7 the first time it's delivered, then 4, then 1... so it's super valuable, but goes down fast. On the contrary, maybe Bluium would range from 5 down to 3, and be a lot more steady:

I forget the exact details, but the bottom line is that while I liked the idea, the mechanism just wasn't working in Terra Prime.

In the end I (thankfully) decided to replace the whole thing with a Demand Tile mechanism. Much more simply, each resource had a price, and tiles indicated how many were required. Once all the blue spaces were filled, the demand for Bluium dries up, no more can be delivered. As soon as one of the tiles fills up, it goes away and a new one comes up, which may bring back some demand for Bluium.

This worked a LOT better, and I'm happy that's how I ended up going in that game.

The point of that story is to show how a major mechanism that I really liked ended up changing because it wasn't right for the game. Recently, that dynamic has showed up again on one of my current active designs: Apotheosis.

One of the instigating ideas for Apotheosis was that your workers would level up when played, like the cards in Solforge. The crux of the game was intended to be whether you play out all of your workers before recalling, and leveling up your entire workforce evenly, versus playing just a couple of workers then recalling them and playing them again, and ending up with a few high level workers while the rest remain at low level. That idea being the whole point, survived the first several revisions, and many playtests. However, as of last week, I tried a new version -- instead of leveling up every worker you have played when you recall them, we said you could only level up one worker.

I was hesitant to try this, but I was hopeful it would fix certain issues I was having with the game. As we began the first game with that rule, I missed being able to play an extra worker before recalling in order to level them up, but pretty quickly I could see how certain aspects of the game seemed to be tighter and work better. All told, you're still choosing whether to recall early to level more often, or play more workers first for their effects.

One concern with only leveling one worker at a time was that it slowed things down, but that's just a matter of balancing costs, and could be easily fixed. However, in an effort to try something in between "level one worker" and "level all workers," we tried something that amounted to "level some workers." I had expected that to be better, with some of the benefits of the restricted leveling, and some of the benefits of the original idea. But in practice I found that it didn't really work as well, and it became clear that just leveling once per recall (and maybe a few special rewards for things could get you an extra level-up) is the way to go.

So there you go -- once again I've had to "kill a darling," so to speak, and remove or significantly change an instigating mechanism for a game.

As a side note on this topic, my friend Gil Hova (of Formal Ferret games) has a particularly aggressive auction mechanism from an old, unpublished game of his called Wag The Wolf (one which I thought was good!) that he keeps designing games around. He used it in Battle Merchants, but in the end cut the mechanism. He used it again in The Networks, and again cut the mechanism before the game was finished. And I believe he started his latest game, High Rise, with that mechanism, and eventually cut it from that game too!

The End Is Nigh! Game End Dynamics and how to use them

The Beginning of the End

Historically, I've felt like I'm bad with game end dynamics. I remember playtesting several different variations for Terra Prime, and not liking any of them. The game ends when all the Yellow space hexes are explored. Blah! The game ends when ALL the space hexes are explored. Double blah!

A similar thing happened with Crusaders, which initially had a whole end game phase. Knights ran for safety while a wave of destruction, emanating from Paris, destroyed (and scored) the buildings in play. Fortunately, in both cases I found more appropriate game end triggers, even if I worried for a long time if the one for Crusaders was really the best I could do.

You should have seen the effort I put into the round end trigger for Captains of Industry, trying to make sure players would play as if the game would continue, but still guaranteeing an acceptable number of turns per round!

I got a little lucky with Eminent Domain in that stack/Influence depletion felt like a natural game end condition, and I was able to achieve an acceptable game length range with reasonably sized decks and Influence supply. Even still, some players (especially new players who stay in the early game longer than is good for them) complain that the game is too short. I even included a "3 player extended game" variant, which I don't recommend using, because I knew that some players would prefer it.

In the Escalation expansion, I changed up the game end dynamic of Eminent Domain: instead of simply finishing out the round in which the game end triggers, now you finish that round and play one more full round. I made this change mostly because the original rule worked less and less well the more players in the game, and Escalation added 5th player support. But I consider that new game end dynamic to be better at all player counts, and I encourage players to use it no matter what format of Eminent Domain they play.

What is it about game end dynamics that give me such a hang up? Let's take a look at the different possibilities, and see if we can identify situations where each one would be a good choice or a bad choice:

Game End Dynamics

There are a few different ways that the end of a game can trigger. Some games end immediately when one player achieves a particular goal. Other games give players a set number of turns to accomplish their goals. Some games ensure players get an equal number of turns, other games do not.

Which of these combinations is best will depend a lot on many factors. As a designer, it is very important to choose a game end dynamic that's right for your game, or else you may end up with players feeling underwhelmed at the end of an otherwise great experience!

For this discussion, I'm going to break down different game end triggers and look at them with two major factors in mind:

1. Do all players get an equal number of turns?
2. Do players have certainty whether any given turn will be their last?

To be clear, not every game must ensure an equal number of turns. And it's not always important whether or not you know it's your last turn, and in some games it may be important that you don't know.  Part of the reason for this discussion is to determine what types of games best benefit from those factors, in which cases they don't matter as much, and ways to compensate players for turn order if appropriate.

Here are all the game end dynamics I can think of:


Immediate End

• unequal turns
• uncertainty

Finish Current Turn 

• unequal turns
• uncertainty

Finish Current Round 

• equal turns
• uncertainty

Final Lap 

• equal turns
• certainty

Set Number Of Turns 

• equal turns
• certainty

One More Turn For Other Players 

• unequal turns
• certainty

One More Turn For All Players 

• unequal turns
• certainty
• incentive to end

Let's look at each of these in a little more detail:

Immediate End

Race games are the most common type of game that end as soon as a player reaches the end condition. This includes race themed games, where the goal is to cross a geographical finish line, but it can also include games that are a race to get a certain number of points, like Catan.

These types of games end abruptly, the moment the end condition is met. Sometimes this means a player doesn't even finish their turn. Hansa Teutonica is interesting in that respect -- it's a euro-style game where you do a number of actions on your turn. If you trigger the end game with your 1st action, you're not even allowed to take your 2nd, 3rd, etc. Obviously this means you will use all the actions possible to get points before triggering the end of the game, but if you have two point-scoring actions, and both of them will end the game immediately, then you can only do one of them, and that can feel like you got cheated out of the other action's points.

Sometimes, ending the game immediately when an objective is reached makes a lot of sense thematically, and continuing to play once the winner has been determined could be considered a waste of time. However, this does mean players are not guaranteed an equal number of turns in the game. Depending on the situation, it may also mean that as the end approaches, player's can't be sure whether any given turn will be their last.

In a game where players make about the same amount of progress each turn, being denied a turn due to turn order can be a real bummer. It feels bad to lose a race when you would have crossed the finish this round as well. But short of simultaneous play, what can you do?

One possibility is to use one of the other end game dynamics. Steampunk Rally is a race themed game that does not end immediately as soon as one player crosses the finish line. Instead, you finish out that round of play, and the winner is the player who not only crossed the finish line, but who moved the farthest beyond it.

Maintaining the immediate end condition though, it may be wise to compensate later turn order players with some kind of boost during setup. In a way, this looks like the staggered starting blocks for a foot race. The runners in the outer lanes look like they get a head start. Is that fair? Well, if you look at the oval-shaped track and straight finish line, you may realize that the outer lanes are longer than the inner ones. That "head start" simply ensures that all racers have the same distance to run.

You have to be a little careful though, turn order advantage can be a tricky thing to balance. If you dole out too much compensation, you just end up unbalancing the game in favor of the last player!

I'd venture a guess that the reason behind using an immediate game end is most often thematic, and as such, it probably doesn't matter whether players know it's their last turn. However, since players won't be assured an equal number of turns, some kind of turn order compensation is probably in order.

Scythe, by Jamey Stegmaier, is a very popular, highly thematic game with Euro-style mechanisms. Scythe uses this Immediate End dynamic, and it does not offer any turn order compensation. Jamey says the idea is to give a player a strong incentive to end the game without threat that the other players will be able to make their big move as their final turn.

In a game like Scythe, scoring opportunities take several turns to develop, and in some cases you can score a fair number of points in one multi-turn play. So I personally feel that equal turns, or some kind of compensation, would be in order.

Furthermore, when game action takes multiple turns to develop, I have come to prefer some certainty as to whether I'll have another turn or not at the end of the game. If I know that this is my last turn, I'll do what I can to get a few final victory points, or shore up my position. I won't invest in a course of action that won't pay off for a turn or two. On the other hand, if I'm sure I'll get at least 1 more turn, I can consider starting such a multi-turn plan. If I don't have any certainty either way, then things get a little foggy. Is it the type of game where I can reasonably tell that another player is likely to end the game after my turn? Or will I be completely surprised by the game end? Is it the type of game where a player should be expected to watch opponents' positions closely enough to make that judgement?

Finish Current Turn 

Many games allow a player to finish their turn after they've triggered the game end. Other than timing (or thematic) oddities, these games are very similar to Immediate End games when it comes to game end dynamics, and they suffer from the same down sides: unequal turns, and uncertainty as to whether it's your last turn or not.

My first game, Terra Prime, uses this dynamic because in that game, players get a large number of micro-turns to sort of simulate simultaneous action. Because the number of turns is high, and the amount of stuff you get done in any given turn is small, I did not feel it necessary to ensure players got an equal number of turns. I did give a bit of compensation for turn order in the form of starting money, but not because of the unequal turns! The compensation in Terra Prime is because the early turns of the game favor the first couple of players in turn order.

For players who felt cheated out of finishing their multi-turn plan, I awarded points for resources and colony markers on your ship at the end of the game. This was a compromise that I feel worked well. It allows partial credit for a big score you might have been about to make, and at the same time it gives you a way to grab a couple of points if you don't think you'll get another turn: just pick up some resources or a colony marker.

Finish Current Round 

My preferred genre is the European style strategy games. Many of those ensure that players get the same number of turns, which only seems fair because those types of games often have a strong aspect of each player doing their own thing, and then comparing their performance against the other players. That doesn't mean there's not interaction in the game, but it does mean that an equal number of turns is appropriate for a fair comparison -- of COURSE I can do more than you if you give me more time to do it in!

Since all players get the same number of turns, turn order compensation is not needed, unless (like in Terra Prime above), seat order favors turn order at the beginning of the game.

Splendor and Century: Spice Road are good examples of the Finish Current Round game end, and both are well regarded as entry level "gateway" games. In Century: Spice Road, the game ends when one player has bought a certain number of scoring cards (like 5 or 6). It's fairly easy to see how many scoring cards each player has bought, and since resources are open information, you can see if a player is able to trigger the end of the game on their next turn or not. So you can tell pretty well if another player COULD end the game before you go again, but you don't know for sure if they WILL end it or not. So there's uncertainty in the game end.

Splendor is the same way, but in my opinion it feels worse. Splendor ends once a player has scored 15 points, and while you collect cards in front of you, many of them are not worth points, and the ones that are generally score only a few points at a time. Like Century: Spice Road, you can look at each player's tableau and add up their score to see if they're able to end the game on their next turn, but that takes significantly more work. Also, since players can have hidden cards in hand that they can build, you might not know for sure whether they can afford a card that could end the game or not. Since Splendor is such a light, quick game, that kind of record keeping does not feel appropriate, and so many players simply don't do it. Therefore, more often than not, I see games of Splendor end abruptly, surprising several players at the table. I find this dynamic to be a real turn off.

Puerto Rico is a classic euro-style game (one of my favorites) that uses the Finish Current Round game end dynamic. In that game there are three different end game triggers, and while you can make a pretty good guess as to whether the game will end this round or not, it's not always clear. However, in my experience that does not detract from the experience very often, perhaps because you get a large number of turns in the game, and because you get to participate in the roles chosen by each player. I have seen instances of the Mayor role occurring before Builder in what turns out to be the last round of the game, so players cannot man a big building bought at the last minute. Maybe because the phase order is in the players' control that has become a feature more than a bug.

My own Crusaders also uses the Finish Current Round game end, and I struggled with whether that was the best format or not. Crusaders ends when the victory point ("Influence") pool has run out, representing that the Templar factions have become so powerful (as a group) that King Philip freaks out and has them disbanded. This seemed like a good thematic way to end the game, and a good mechanical one as well, I was able to tune the vp pool to allow the game to last about the right number of turns. However, I worried that players would not pay enough attention to the dwindling pool of Influence tokens, and therefore not recognize whether they would have another turn or not. Especially as player 1 in a 4 player game, you can be left to choose a course of action without any real certainty whether or not you'll have another turn. That's my least favorite part of that game.

And as I mentioned above, Eminent Domain's original game end rule used this dynamic. There was some uncertainty as to whether you'd get another turn or not, but the fewer players in the game, the less likely it would really surprise you. I was ok with this dynamic originally, but by the time Escalation came around I had soured on the uncertain aspect of the game end, so I changed it to the next category, the Final Lap. 

Final Lap 

Railroad Tycoon is one of my favorite games for a number of reasons. That game is divided into game turns, each consisting of 3 rounds of player turns. The game is about delivering cubes on the board to the cities that want them, and the game end triggers when a certain number of cities are empty of cubes. When I first played the game, I wondered why the rules indicated that after triggering the end of the game, you don't just finish out the game turn, but you play one more entire game turn (3 more rounds of player turns) before stopping play. It seemed extraneous to me.

Now I understand that there's a real strategic benefit to knowing exactly when the game will end, and exactly how many actions you have left. Once the game end triggers, you can plan out your remaining actions to maximize your points, which is fun, especially after spending 2+ hours building up your position.

As I described above, I adopted this end game dynamic for Eminent Domain, and I'm very happy with the results.

While the previous category (Finish Current Round) may be appropriate for lighter, shorter games where you can fairly easily see if the game will end before you get another turn, the Final Lap game end seems more appropriate for longer, heavier, or more strategic games, or games where plans take a few turns to develop and resolve.

Set Number Of Turns 

Many games, especially euro-style games, do away with thematic or variable game ends and simply let you play for a specific number of turns. This has the benefits of both an equal turns for each player (so likely fair), and certainty as to when the game will end (good if you like to plan). But on the down side, it can feel arbitrary and un-thematic.

Shipyard is one of my favorite games where you build ships, collect stuff to load them up with, and then take them on shakedown cruises in canals that you've built. The game uses a rondel mechanism for its action selection, and many of the actions resolve using other rondels. Just like that main mechanism itself, the game is all about planning ahead, therefore knowing exactly how many turns you have left is welcome.

One More Turn For Other Players 

Concordia is a euro-style game, but it eschews the Set Number Of Turns and even the equal turns for all players dynamics. Instead, the game offers a 7 point incentive (a decent turns score) to end the game, and then gives each other player one last turn. This incentive is meant to encourage players to actually trigger the game end rather than let the game draw out too long.

In Concordia players get a large number of turns to play cards, and once in a while you basically skip your turn to get your cards back. Over the course of the game, players end up with different numbers of cards in hand, and therefore skip a different number of turns. So equal turns does not seem to be important in this game. But as a forward planning game, Concordia's game end does offer the certainty of knowing whether it's your last turn or not.

This game end dynamic could be good for any game where an equal number of turns is not necessary, either because individual turns are low impact, or because players get so many that being short one doesn't have a big impact on the outcome, but is strategic enough that it's worth knowing whether you have another turn coming or not.

Often, an incentive to trigger the game end (like Concordia's scoring bonus) is necessary in this type of game, otherwise it could lead to a game of chicken, with no player wanting to trigger the game end.

One More Turn For All Players 

Ticket to Ride is a well known, excellent, gateway game that ends when any player gets down to 2 train pieces left. In that game, ALL players get one last turn, even the player who triggered the game end (incidentally, that's why it triggers with 2 pieces left instead of 0, so you can lay track on your last turn).

This is similar to the previous case (One More Turn For Other Players), except your incentive to trigger the game end is that you get the last turn. This dynamic also has unequal turns, and certainty in the game end, but with the built in incentive to end the game, additional incentives (like Concordia's scoring bonus) are usually not necessary. Like the previous entry, this dynamic could be good if strategic planning is possible, but it's not important to have equal turns.

Getting into the game

Yesterday I met David and Hoss at the game store and I got the opportunity to play one of my favorite games: Shipyard, by Vladamir Suchy. David was new to the game, so we needed to go through the rules. They took a while to get through, but if you're familiar with that type of game, you'll understand that you kind of need to know everything before you can reasonably do anything.

Complicated rules vs a complicated game

The rules of Shipyard aren't really complicated to tell you the truth -- you choose an action that isn't the one you did last turn, and isn't blocked by another player, and you resolve it. Optionally, you can buy a bonus action as well. There are a few details like the fact that you can do the bonus action and the main action in any order, and that you slide some tiles around (which indicate the actions) and you get income for taking an action that's "behind" other players' pawns, and of course there are details about how each of the actions resolves.

I have run into the same issue in the past teaching another favorite: Tzolk'in: the Mayan Calendar, by Simone Luciani and Daniele Tascini. The rules to that game are simple as well: either place workers (paying money), or remove them (resolving their actions). There are details such as the increasing cost for each worker you place in one turn, the fact that after a round, the wheels will advance, moving your worker to a different (usually stronger) action, the one-time power you have to move the wheel 2 spaces, and of course there are details about how each of the actions resolves.

The more you know...

In both of those games there are a large(ish) number of different actions or effects that can result, and technically you need to sort of know what each of them is so you know whether or not you want to perform them. But in both Shipyard and Tzolk'in the actions are grouped:

The first wheel in Tzolk'in provides food or wood. The farther along you are, the more food or wood you can get. The next wheel provides resources. The farther along you are, the more/better resources you get. The third wheel allows you to advance technology and build stuff. The farther along you are, the more or better you get to do that. Etc.

I've universally heard from players that Tzolk'in is "so complicated" because there are so many different action spaces. Maybe I've learned to sort the information in a useful way, but to me it doesn't seem very complicated at all. Obviously that's not true of all players!

In Shipyard, the actions are similarly grouped onto several different areas, and for the most part they all resolve the same way. If you want rail cars, ship parts, or canal tiles, then you simply buy them for $0/$1/$2 as shown on the board. The rest of the actions are on rondels of their own... If you take the green action, you advance the green rondel, you can pay to advance it more, then you take the item shown. If you take the brown action, you advance the brown rondel, you can pay to advance it more, then you take the item shown. If you take the employee action, you advance the employee rondel, you can pay to advance it more, then you take one of the employees there. Yes, there's 1 rondel where you can't pay to advance it, and which requires a little more description. And yes, there are a couple of details regarding the employees (you can't have both copies, some cost an extra $1 as shown on the tile, and the 3 Level II ones have a prerequisite of the matching Level I), but for the most part the information is compartmentalized.

A few turns into Shipyard, I asked David how he was enjoying the game. He said at that point he 'got it' and was enjoying it, but that he almost bailed on the game after the long rules explanation! He hadn't wanted to play after the teach, and only went ahead because he felt like we'd invested that time. And this is a guy who's played a fair number of games -- he's one of my regular playtesters!

Getting into the game

There are tips and tricks to teaching a game. Certain ways to organize and present the information so that it makes sense. Paul Grogan (of Gaming Rules! fame) has adopted a potentially controversial stance for demoing a game at conventions (and perhaps he does it when teaching at home as well) -- he ONLY tells people what they need to know RIGHT THEN, and nothing more. This is a neat idea, one I've toyed with myself at times, though I've encountered a fair bit of resistance when trying to teach that way. Many players don't want to choose an action without knowing the consequences, even if it's a learning game. I myself have a pet peeve for when a game asks me as a player to make a choice without giving me enough information to decide which option is better for me, and this forces that dynamic on all players as they learn, with the logic being "see what happens, and when you play for real you'll be better informed." Also, when playing a deep game with your friends, having a new player play this way kind of sours the experience for the experienced players, so to play this way every time would be kind of a bummer for the teacher.

This novel approach can work, but I think it works better on modern games than it does on some of the more complicated games from a decade ago. Newer games seem to limit options in the early game, or give you a player power that nudges you (sometimes very strongly) toward one option over another. Many older games are more like a sandbox, with clear strategies that exist and emerge through game play, but myriad options in the early game, with no direction except your own forward planning. All those options can lead to a lot of strategy space, and a lot of depth, but at the cost of some accessibility. To an extent it requires the player to understand what they're in for, lest they be overwhelmed.

This sandbox nature is something I'm finding to be "old fashioned" about a lot of really good euro-style games. Personally, I enjoy the freedom of strategy and the depth provided by these types of games, but with the rate at which new players are coming into the hobby, and the rate at which publishers are churning out new games, we need to start finding ways to get people actually playing the games without requiring lengthy rules explanations. We've already seen a few attempts at minimizing rules, or removing them altogether:

Jamey Stegmaier's legacy euro-style game Charterstone was originally intended to have no rulebook. In the end he found that he needed at least a bare bones rulebook to express the core mechanisms in the game, and while I haven't played it, I suspect additional rules may be added to that book via sticker (as Pandemic Legacy did) as you play through the campaign.

Friedmann Friese's ambitious Fast Forward series (Fear, Flee, and Fortress) come without a rulebook at all. They're just stacks of cards which you're supposed to tear open, set on the table, peel the top card off and read it, following directions as they're given. This is a pretty neat experiment. I've played two of those games, and while the cooperative Flee seems like a decent group puzzle, the competitive Fortress seemed a bit lacking for my taste (I'm not really the target audience). But more importantly, I felt like there were some problems with the "no rules" format -- we came across a few timing or rules questions when things weren't crystal clear on the cards (and there's not a ton of space on the cards for rules text), and when that happens, there's nowhere to look for answers. Also, with the amount and intricacy of some of the rules given on cards, it seemed like a hard sell to call it "accessible" to a complete non-gamer.

Easing players into the game

What we need is for sandbox-y games with strategic depth that players can get into with minimal up-front rules explanation. Games that can be taught using The Grogan Method, if you will, without feeling like you're making choices at random and seeing what happens. Can this be done without losing the depth these games have? I think that's the job.

The question is... how?

Post script: Reviving old games

I see a lot of reprints of older games coming out lately. Classic eurogames from 10+ years ago, sometimes with updated art, but seldom with updated rules. When I think about the possibility of reviving an old favorite like Shipyard, I wonder if it would really go over well enough in today's market. Sure, hardcore gamers that know they like that kind of thing, or are willing to sit through a long rules teach in order to explore the mechanisms on offer, won't have any problem. But will people new to the hobby be willing to put so much effort into learning the game? Or will they instead opt for something simpler to learn?

Are there any adjustments that can be made to games like Shipyard and Tzolk'in that will allow players to get into the game a bit easier? Can those be made without losing what makes those games so good in the first place? I'd love to hear your thoughts or answers in the comments below!

Exploring variable player powers

Variable Player Powers

Something that seems to have gone over well with the TMG audience, and with a lot of gamers in general, is variable player powers. Often, unique abilities can lend a sense of replayability to a game by giving the player a different feel, or a nudge toward a different strategic path or goal.

But these powers can be a double edged sword, with the potential for a player to feel pigeonholed into a particular strategy or play style as dictated by their player power, or the possibility (even just the perception) of some powers being significantly stronger or weaker than others.

To be honest, I've historically prefered a game where players start out on equal footing, and quickly differentiate their position through game play. I've always liked it when, based on early choices, you develop your own unique player profile for the game. And I especially like it when you can set yourself up with a different player profile the next time you play (replayability!). However, I'm coming around. I get the impression more and more that variable player powers are worth adding to many types of games.

I've just signed a few new games that will be coming out in the next year or two, and I want to add player powers to two of them.

The evolution of player powers at TMG

In order to inform my thinking on variable player powers, it may be useful to look at the TMG games that have utilized them. Here they are, in pseudo-chronological order:

The first TMG original game that had player powers was Ground Floor, from 2012. Designer David Short included Specialty tiles that gave your business a little nudge in one direction or another. They upgraded one of the 6 starting spaces on your player board -- something you could normally do on your own for a cost. With the Specialty tiles, everyone starts with a different space upgraded. This doesn't make a huge impact on the game, but it does make you a little better at one aspect or another, and therefore better suited to utilize that aspect. Presumably a player with an advantage in a certain area would play in a way that takes advantage of that advantage.

When Kings of Air and Steam came along in 2013, we tried something new. We used 2-sided player boards (A/B), where the A-sides were identical, with each player starting on equal footing. We called that the "Basic" game. In the "Normal" game, each of the 7 possible player characters had a unique board with subtle differences in different aspects, as well as 2 special abilities to choose from, and a specific initiative order. It was a lot of fun to come up with, test, and balance the 14 various abilities, and we used the player board differences and initiative sequence (turn order) to adjust for abilities that were too strong or too weak until we felt we'd gotten a fairly well balanced game.

Another TMG game to get variable player powers around that time was Dungeon Roll, and while the other games were complete games without these powers, in Dungeon Roll, the powers play a much larger role in the game. Each hero in Dungeon Roll has a Specialty and an Ultimate ability, one or both of which gets better when the hero gains enough experience to level up. Dungeon Roll shipped with 8 heroes (9 if you count the kickstarter Guild Leader promo), and has since gotten two boosters of 8 more heroes each, as well as the holiday themed Winter Heroes pack, and the Time Traveler promo hero. Kind of like Dominion and Bridge rely heavily on having a new situation each hand (new kingdom cards, or a new shuffle and deal), Dungeon Roll relies on using different heroes to make the experience interesting. However, many of the heroes can offer at least a handful of games before they get boring, and with 30 heroes so far, there's a decent amount of game there. Without the different hero powers, I think the game would get old pretty quickly.

Bomb Squad is a cooperative game from 2013, and designers David Short and Dan Keltner used player powers similar to what you'd see in Pandemic, and to much the same effect. In my opinion, the base game play of Bomb Squad is so solid and fun that the player powers aren't really necessary to play the game over and over, but their effects do add to the experience. Since the game is cooperative, the player powers don't need to be as balanced against each other, they just need to be useful and fun twists on game play to let players feel like they've got an extra, unique way to help the team.

Belfort, released in 2011, never had player powers. But when the Expansion Expansion came out in 2013, it added assistants that you could draft each round, which approximated a player power in that game. Not really the same as a standard variable player power, but worth mentioning the approximation.

My deck learning card game Eminent Domain was also released in 2011 without variable player powers. For many players, the best part of the Escalation expansion from 2014 was the Scenario cards, which altered your starting deck and gave you some technology to start with. They were so well received that I added 5 Base Game scenarios as a promo in Microcosm, and a handful more came in the Exotica expansion. They haven't come to fruition yet, but I talked on my blog about another possible promo item: Emperor Avatars (described here). Emperor Avatars would be a set of two player powers that you would draft before starting the game.

Steam Works, in 2015, was the next big game to have player powers. This time we repeated the A-side/B-side player board idea from Kings of Air and Steam, but rather than having the A-sides being identical, they were unique. The B-sides had even more diverse or crazy abilities.

2015 also brought us Harbour, by Scott Almes. The base game of Harbour is a solid, compact worker placement game with an interesting market mechanism, but like Dungeon Roll, the real attraction for players (I think) comes from the host of player characters in the box. Since Harbour we've set a couple of games in the same universe (with more to come), and one of the things I've tried to do with each of those is maintain that format by providing diverse and colorful player characters with fun abilities.

2015's Dungeon of Fortune is basically a card game version of the press-your-luck dungeon crawl that was Dungeon Roll. Different mechanics, but the same theme and setting. As such, it made sense to add player powers to that game as well, and we tried to model them after some of the heroes in Dungeon Roll.

I wasn't involved in Andy's Bottlecap Vikings (also from 2015), but I know that he added variable player powers in the form of different upgrade powers on your player board in that game.

Variable player powers in today's TMG titles

This year (2017) has brought us a handful of TMG titles with player powers. In Exodus Fleet (which I didn't work on) you choose one of two factions, and that lets you start with a slightly different starting ship, and a unique card or two which could give you a nudge down a different strategic path. I don't think the differences are so substantial that you'd feel forced to pursue a particular strategy though.

I had a big hand in their development of the rest of the 2017 titles, including the variable player powers:

Chimera Station follows the Kings of Air and Steam tack of A-side (identical) and B-side (unique) player boards. That game originally didn't have player powers, and I decided they could be a good thing to add. In Chimera Station, you modify your workers by adding 4 different types of components to them: Brains, Claws, Leaves, and Tentacles. It also happens to be a 4 player game. So for unique player powers, each faction/player is kind of "good at" one of the components. This may be a little heavy handed, pushing you fairly hard toward using a bunch of the type of component that you're "good at." However, you have several workers, and it's still worth getting other components than the one you're "good at," so I think it works out.

When we signed Harvest, the game was simply about growing crops. It didn't have any characters or payer powers. But I decided to set it in the Gullsbottom universe, like Harbour before it. So naturally, it made sense to add player powers like we had in Harbour. At first I thought it would be good to take some of the actual Harbour characters, and see if we could interpret their abilities to make sense in the context of this game. However, we instead made new characters for this game. Designer Trey Chambers, taking a cue from The Voyages of Marco Polo, thought it would be good if the player powers in Harvest were more than just a tweak here, or a nod to a specific strategy there, so his first draft of the player powers were very powerful, very unique, and very diverse! We embraced that strategy for this game, and came up with a set of 8 characters with crazy powers. Mixing that with the A-side/B-side format, I wanted to have the common side be something basic, but I decided this time to make the A-sides playable along with the B-sides. So the basic character, Wil Plantsomdill, has a well rounded spread of starting resources, no particular abilities, but gets 15 points added to his score at the end of the game. This way, you can deal 2 player boards to each player, and they can choose between one character, the other, or Wil, and if I've done my job, the game will be fair.

I was working on Pioneer Days at the same time I was working on Harvest, and I decided to try the same thing with the player powers in that game. Originally there were none, and I decided to add some. Like Kings of Air and Steam or Chimera Station I wanted a basic side and a unique side, and like Harvest, I wanted it to be comparable to the unique side. What I decided to do was to take the designers' proposed standard player power (before we had unique ones), and make it the standard power. It's straightforward, and recommended for first time players, but it's not just vanilla like Wil Plantsomdill is. We called this the Standard Pioneer, and put it on the A-side of each player board. Then on the B-side of each board is one of the unique powers. Again, you can deal 2 player boards to each player, and they have 3 options to choose between: the 2 unique characters, or the Standard Pioneer, and whichever they choose should be competitive.

We have two upcoming titles that have been announced so far for 2018:

Crusaders: Thy Will Be Done is my next title, and in that game you get a knight faction tile with an ability in it. Most of the abilities have to do with controlling the distribution of your action tokens around your action wheel, which was their purpose. These aren't big effects, but they help you work with the rondel mechanism to pursue your chosen strategy.

I wasn't involved in Downfall at all, but that has player powers too. I think they're pretty low impact though, you get a Leader unit, which is like an upgraded regular unit, and you get a special card which is more powerful than normal. It's a card drafting game, so if you pass your special card and another player uses it, then you get a small owner's bonus while they get the better than usual effect.

Let's think about these games and how they use player powers, and consider what we can learn from them. This will help me design and test player powers for the two new games I mentioned earlier.

Good vs poor player powers

I think the kinds of things that make good player powers are things that either:

• Give you some sort of starting advantage, then you don't have to worry about them again. These are straightforward and good for new players or those prone to being overwhelmed, however they're not ideal since they don't necessarily affect the way you play.
• Give you static end game goals, such as bonus points for certain things, which give you a nudge toward a long term plan. These help drive your choices, but really they just weigh some choices more than others because you want to do things to increase your end game bonus.
• Give you reasons to make choices you wouldn't otherwise make. This is probably the most ideal kind of power, provided it's balanced, but there's a danger of feeling like you're not even playing the same game if the incentive is to do something TOO different than normal.
• Give the you agency to control something you couldn't otherwise.
• Give you an incremental bonus when you do certain types of things, as long as you set them up correctly.

I think the kinds of things that make poor player powers are things that:

• Introduce additional decision points which other players have to wait for, especially during maintenance phases. This can break up the flow of the game, or create timing issues
• Are easily forgotten. If an ability is so seldom used, or so easy to forget you have it, then what is it really doing for you?
• Occur based on chance, such that you might not get an opportunity to use your power, or such that a power could be significantly more or less valuable each game just due to randomness.
• Slow down the game or are hard to keep track of.
• Are directly interactive in a negative way (this is OK in some games, but in most games I work on it's probably not OK).

Are these lists exhaustive? I doubt it. If you have types of powers you'd like to add to either the "good" or "poor" list above, please mention it in the comments!

What do you think? What makes a good player power? What are some examples of successful powers in games, and what are examples of powers that are unsuccessful?

Nuggets of game design advice from designer Matthew Dunstan (Relic Runners, Elysium, Pioneer Days, Professor Evil and The Citadel of Time)

A while ago on Twitter, people were posting in a format such as "for every like this tweet receives, I'll post a nugget of wisdom." As a result, a few threads of potentially useful information appeared, but as is the nature of Twitter, those threads were relatively easy to read in passing, but difficult to reference, copy,/paste, or be otherwise useful in the long term. As such, they're probably effectively lost to the annals of history, only to be stumbled upon via random google search every now and again.

One of my friends, renowned game designer Matthew Dunstan, did one of these threads, stating:

I am going to try this '1 like = 1 nugget of game design advice' thing. Hope you find it useful!  #gamedesign

— Matthew Dunstan (@wisegoldfish) July 5, 2017

I left it open in a tab, but hadn't gotten around to reading it until now, and I realized that the a tweetstorm may not be the best format for reliable access to good advice, so I took a few minutes to copy, paste, and clean up the list. I'm presenting it here (with permission, of course). And I took the liberty of correcting some of Matthews British misspellings..:

Nuggets of game design advice from designer Matthew Dunstan

1. Where possible, start your design with pen and paper - it's much easier and quicker to get started, and forces you to simplify.

2. Ideas may be cheap, but that doesn't mean you have to be wasteful. Give your ideas enough time and effort to find their best form.

3. There are no excuses for not trying a new prototype as soon as possible - most games can be played solo perfectly well. Don't put it off!

4. Find information about new games deeply and often, especially games you don't think you'll like. The best inspiration can come that way.

5. It can be very easy to get caught up with a cool new mechanic or system and forget why the players should care or notice each other.

6. Try to have each decision point in your game have several different consequences - the game will be just as strategic but quicker.

7. When starting out in game design, be very careful of starting too many projects without actually finishing one, even if it feels easier.

8. If you have rounds or upkeep in your game, ask yourself whether you need them. Can you fold these parts into the regular player actions?

9. Having actions in a game which are the inverse of each other (and mixing this ratio) mean players can be in control of the game flow.

10. Game design isn't a solitary pursuit. Even at the start, cultivating communities that you can be a part of and contribute to is vital.

11. Follow @CardboardEdison - they do a great job hunting down game design content so you don't have to!

12. When you start on a new game, you really want to do it. Don't give up when it becomes something more like work, both stages are needed.

13. Game design is like any other skill: to improve you need to practice often. Even if the game doesn't work, you'll get better.

14. For me, the best way of pitching games is at conventions; it also allows you to get to know editors which can lead to opportunities.

15. Write down feedback from playtesters as soon as you get it. You will forget it otherwise, and it shows your testers respect.

16. Good games are easy to design, but the best games require struggle and doubt in their inception.

17. Never underestimate the joy that players get from the physicality of board games, and try and find ways to make the most of it.

18. If you have to design a game with abilities or special effects, first design 1/4 of what you need, make 4 copies of each, and test.

19. If you want to get published, recognize that there is a lot of work beyond a contract. Development, editing and promotion are important.

20. Entering design competitions are a great motivation to finish games, and provide useful feedback (sometimes) - use them!

21. Design progress = # iterations/ time.

22. Watch @britishgaming Game Makers Toolkit on YouTube. While based on video games, very thought provoking for tabletop design as well.

23. Find software that has Data Merge, and read @DanielSolis articles on how to use it. Big time saver for iterating on card games.

24. Have a prototype on you at all times. You never know when you'll have a playtest opportunity (eg. At the film festival this week).

25. Wherever possible, insist on an advance payment in contracts, even if it's small.

26. Constraints-based design is a great way to turn ideas into a first prototype. How can I turn theme X if I only had Y cards?

27. Avoid requiring players to make decisions that effect future turns, but giving them no way of knowing what is a good choice now.

28. Any change to a design that simultaneously solves multiple problems is probably the right change to make.

29. Keeping up with trends can be useful; both by knowing what has already been done and what publishers might be looking for.

30. Try co-designing a game at least once. At worst it's a learning experience. At best you'll make better games, and new friendships!

31. Try to find out what situation or activity best stimulates the game design part of your brain (mine is going for a walk).

32. Whenever you sit down to design, no matter how much time you have, set a goal. It makes the process much less daunting.

33. Be aware of absolute versus relative values. Having a range of 1-4 (points, resources) can be very different to a range of 2-5.

34. In open systems, giving players some direction early is important. Starting quests, special powers etc are only some ways of doing this.

35. The best special player powers are ones that make the player feel powerful, but not restricted to only one strategy.

36. Adjusting the level of randomness is a good way to make a cooperative game more resilient to alpha players.

37. Try to give every game an arc. Even if it's a short game, try and make the start, middle and beginning feel different to play.

38. Some of the best games keep players involved even when it's not their turn, e.g. Catan. Keep your players engaged as much as possible.

39. Collaborating is much like theater sports - designs progress more quickly if you always say yes to different ideas and viewpoints.

40. Once in a while, try designing a game in an unfamiliar genre. It's good exercise for your design brain.

41. Where possible, encourage the publishers of your games to include characters of diverse backgrounds and gender.

42. Don't put off updating your design - the longer you wait the more chance you will miss the opportunity to playtest the changes.

43. Be professional at all times, with publishers, developers, designers and players. You never know where an opportunity might come from.

44. Elegance in game design often comes from simple systems with interesting and deep interactions between players. Interaction is key.

45. Early in a design, don't be afraid to make large changes. Exploring more of the design space makes it possible to find the right fit.

46. If you are going to pitch to publishers, know their games, and know how your games could fit their line. Respect their time.

47. Don't spend too much time turning an idea over in your head. It will all change once you play it, and you won't know how beforehand.

48. Even if you're not led by theme in a new design, finding a good theme early ensures a close match between experience and setting.

49. Don't settle for a mechanic or part of your game that is 'good enough'. It will be found out sooner or later.

50. Down time is as important as active  time. Sometimes when you step away from a problem the answer comes when you least expect it.

51. Use the 2:3 rule - whenever  you design a game where you upgrade actions, the upgrade should double its cost and triple its effect.

52. Think about how your game can have memorable moments for a player, things they will remember many weeks and many games later.

53. For majority games, use prime numbers for the number of different majorities, it means that players will be forced to compete.

54. Having a mix of short and longer term goals keeps players having to weigh tactical vs strategic considerations.

55. Beyond making a game with multiple paths to victory, try to make the experience of these paths feel different, not just the label.

Revisiting a classic post: Balancing Game Elements

I've been working on Harvest lately, an upcoming TMG title in the Harbour universe (side note: watch for Harbour on TableTop with Wil Wheaton, season 4!).

Here's the description of Harvest from BGG:

Mind the fields of Gullsbottom! Plant and fertilize your seeds, tend your crops, and utilize the various buildings at your disposal. You'll need to work smarter, not harder, as harvest season is coming to an end! Who will have the best harvest this year? Will it be you?
Each round in Harvest, you first draft turn order (and the benefits that come with it), then send your two workers into town and into the fields. Plant seeds, tend fields, and harvest crops to make room to plant some more! Utilize buildings and magical elixir to amass a bigger and better harvest than your neighbors at the end of five rounds of play.

Harvest is a worker placement game where you first reveal worker cards (spaces that will only be available this round), then draft turn order (the later you go in turn order, the bigger the bonus you receive), then place your workers and take the associated actions. In developing this game I've been working with the designer to figure out the appropriate power level of the actions in the game -- both the standard spaces on the town board, as well as the value of the worker cards, and the buildings you can build. As that's mostly accomplished, lately we've been working on the power level and balance of various characters you can play in the game.

At several points during this process I've been reminded of an article I wrote back in January 2014 called Balancing Game Elements... re-reading it now I think that might be the best game design article I've ever written. It continues to hold true today, as I have been using the same process to balance the elements of Harvest.

You see, a major benefit of working this way, finding an average value for an element (say, the buildings in Harvest) which incorporates all the costs and benefits of that element including opportunity costs, is that it leaves you with only one variable when designing things that interact with that element. This makes it relatively easy to determine things like the value of an action which gives you that element.

By way of example

Without knowing anything about the game it may be difficult to give you a concrete example, but I'll try:

In Harvest, there's a town board that has 3 main areas that offer a variety of different effects or resources. Each of these areas has a "Choose 2" space (letting you get any 2 of the things on offer in that area) which is limited to 1 worker, and a "Choose 1" space (letting you get just 1 thing) which is unlimited. In addition, each round you'll turn up a number of worker cards which have more action spaces on them. The value each of these spaces confers is defined as follows:

Choose 1 space: 1-2 units
Choose 2 space: 2-3 units
Worker card space: 3-5 units

So ideally you'd prefer to take a worker card space first, a Choose 2 space next, and a Choose 1 space only if you had no other option, just based on the value of stuff you would get.

However, the game is not quite that straightforward. A space that's technically worth 5 value might only be worth 3 to you because you can't use all of it's benefits at the moment. So there are plenty of times that a Choose 2 space is just as good if not better for you than a worker card space. Very seldom do I want a player to choose a Choose 1 (default, fallback) space over a worker card space though.

Note that these values are sort of average values, and they may depend on your situation and whether or not you can make full use of the resources you get from these actions.

That said, there are buildings in the game which can confer abilities, one-shot resources, or an end game scoring bonus. There is a wide variety of buildings, with 6 face up to choose from at a time, each supporting various strategies. It's difficult to evaluate exactly how much each of these buildings is worth, which is where my Balancing Game Elements post comes in handy. If you read that post, you know that step 2 in the process involves choosing a desired power level for the elements and designing the elements to be worth about that much. I'm currently choosing to assign a value the buildings in Harvest at "4". This means that I'm targeting an average value of 4, some buildings will be worth a little more or a little less, depending on whether you can utilize them to full effect or not.

One of the things you can get from the town board, an action that's always available, is building a building. So, if the buildings are worth 4, and the default "choose 1" space is supposed to be worth 1-2, and you can use that default space to build a building... then it follows that there should be a cost of 2 to use the build action on the town board. Also, if the worker card spaces are supposed to be worth 3-5, then perhaps one that just allows you to build a building for no cost is appropriate. Simple math, which can be applied because I wrapped all the variables into the valuation of the buildings.

Now, I may be incorrect in that evaluation. If I've over- or under-valued the buildings then that should show up in playtesting a players recognize and either ignore or capitalize on the imbalance. If that turns out to be the case, then I can easily re-evaluate them and adjust the actions that allow you to build accordingly. Let's say I significantly overestimated the value of buildings, and that they should really be worth only 2. In that case the standard build action on the town board should not cost anything, and a worker card that allows you to build should also come with 1-3 value worth of more stuff. Similarly, if I'd under-valued buildings and they are really worth an average of 6, then the standard build action should cost more, and the worker cards that allow you to build might also need some kind of cost.

An alternative to tweaking the actions is to make an editing pass at the buildings, either powering them up or down until they are more closely averaging the targeted value of 4, which would then justify the cost of the town board and the worker cards.

Hopefully that makes sense, and indicates the usefulness of incorporating all of the costs and benefits (including opportunity costs) into the value of the game element itself (in this case the buildings), rather than trying to think about things separately.

Treat the problem, not the symptom

Things go wrong in playtests all the time. When something goes wrong in a playtest, it may just be an anomaly of that particular game -- but if the same thing comes up in multiple plays, then it's probably indicative of some kind of problem with the design. As designers, it's our job to identify these problems and fix them.

This may be harder than it sounds though. Sometimes the issue experienced in a playtest is merely a symptom of a deeper problem. While it may be worthwhile to fix the symptom, what we really want to do is determine the underlying problem and fix that, or else another symptom is likely to rear its ugly head before long.


By way of example I'll illustrate what I mean:

When I was originally testing Eminent Domain, instead of the 5 different role cards in the game there were 6. Produce and Trade were separate cards, not combined into one card. Produce/Trade cycles had the potential to be more lucrative than other scoring paths, but the logistics behind having to take the two separate roles, and having cards for each in your deck, were supposed to balance that out. However, in practice I found that Produce/Trade strategies were frustrating to pursue because it was hard to score well off of the cycle. So hard in fact that players would ignore that strategy altogether.

At first I tried doubling the value of a Produce/Trade cycle, thinking that with higher rewards, players would pursue the strategy. In theory that worked, I could keep increasing the value of trading until it was worth going for. But that solution treated the symptom, not the problem. The more valuable I made trading, the more variable and swingy that strategy (and the game) would become... a player trying to produce and trade might never get their cards in the right order, and when they did they'd get a disproportionately high reward, like a slot machine. The real problem here was that it sucked to have to fill your deck with 2 different cards that were terrible on their own just so that MAYBE you can use them to collect lots of points.

Fortunately one of my players (thanks Wystan!) suggested putting both roles on the same card so that whenever you draw that card, it'd work for either the Produce role or the Trade role - whichever you need at the time. This solution ignored the symptom and went straight for the throat of the underlying problem. As soon as I tried it I immediately knew it was the right solution. And as a result of solving the problem, it eliminated the symptom as well.

So let that be a lesson to you. When something goes wrong in your playtests, don't take it at face value. Make sure to look and see if the issue is merely a symptom of a bigger problem, then do what you can to fix the underlying problem.