Tag: randomness

Logic vs Intuition

I was watching a Lewis Pulsipher game design video where he talked about players who like to use Logic vs Intuition in games. He broke things down as follows

Logic Players:
-Tend to balk at randomness
-Tend to like to “figure things out”
-Tend to want to come to a definitive conclusion
-Tend to like more serious “thinky” games

Intuition Players:
-Are okay with randomness
-Tend to go with heuristics or instincts
-Tend to be OK with fuzzy conclusions that may be incorrect
-Tend to prefer “beer n’ pretzel’ style games

I wanted to push back on this a bit an mention that I like to think about this in terms of Keith Burgun’s forms. Kieth defines puzzles as interactive systems that have “correct” solutions and games as those that feature much more ambiguity in terms of your approach. Basically this means that it is not ever confirmed whether your move was “best” or correct in a Burgun game (as it would be when you have solved a puzzle), but you can have some sense of whether it was good, or whether some moves are better than others through a developed intuitive/heuristic sense of the game system.

These Burgun games – systems that rely on heuristics/intuition rather than calculation – are what I am most interested in playing and creating. For starters, you can circumvent the calculation/busywork involved in taking a purely logical/conclusion based approach, and secondly, because your “solutions” are ambiguously correct, there is a lot of room for continual development of your internal heuristic framework as you continue to engage with the game system!

Commandment #5: Ambiguity

This post is part of a series of articles detailing the BrainGoodGames Design Commandments. You can see the full list here.

A strategy game is a delicate thing. Humans are excellent at using heuristics to reduce complex problems to simple rules of thumb, and the formation of those heuristics (learning) is a huge part of the enjoyment. However, if a game features too much rules complexity, the player will spend their time trying to internalize the rules rather than engaging in higher level strategic thinking (more on this in Commandment #1 here). On the other hand, the less inherent (rules) complexity a game has, the easier it is for human minds to “solve” it, at which point it becomes more like a puzzle than a game. A classic example of a game that is highly solvable is tic-tac-toe.

Games do have a secret weapon to employ to help resist solvability, while maintaining some degree of elegance: Ambiguity. Interestingly, one of the most prevalent and oldest forms of ambiguity in games is simply another player! You are unable to know definitively what other players will do in a game (although some games are designed to test your ability to guess), so even in the case of a very deterministic system like Chess, your opponent contributes to a very high degree of unpredictability/unsolvability (such that we have played Chess for many many years without “solving” it).

However, in a single-player strategy game, we do not have the luxury of all this free ambiguity; designers must thoughtfully add it themselves. The most natural solution may seem to be to come up with an intelligent AI opponent to play against. This has a few drawbacks. Firstly, in a practical sense, playing against an AI opponent in a strategy game frequently has players casting about for ways to exploit the patterns in the behavior of the AI, rather than trying to form a deep understanding of the game system itself. Secondly, if we think of an AI as a set of rules, adding even the simplest AI to our game has a huge ballooning effect on the rules complexity of our game. Players may not need to know about the AI behavior, but they can learn about it, and doing so pushes against learning how the core systems of the game interact strategically.

Fortunately, there are many more techniques available to a game designer to add ambiguity to their games. Dice, cards and coins provide convenient metaphors so that the player can understand the possible outcomes and their likelihood without having a big reference table. Procedural generation is another fantastic tool for providing new non-rules information to the player while they have time to strategically react. The enemy movements in Militia are yet another example of providing new information for future turns. Each enemy chooses from ~4-8 possible moves, within simple rules, which has the effect of mixing up the board for next turn, but in a way that can be understood and planned for.

This board is a product of the board that came before it plus some simple enemy movement rules.

There are absolutely ways in which introducing too much ambiguity, or ambiguity that the player cannot react to or plan around can reduce the learning potential of your system or make it unsatisfying to play. I’ll be going into more detail about my take on the different forms ambiguity can take in later articles, but for now you can check out Keith Burgun’s excellent article on the subject.

Advance Wars Game Design

Today I was watching a cool video about Advance Wars map design. Makes me want to play some more Advance wars :).

In particular I really like the way randomness is handled in Advance Wars. For the unfamiliar, units all have 10hp, and when you get in a fight the system will tell you you will do 52% damage for example. This means you have a 80% chance to do 5 damage and 20% chance to do 6 damage. There are two reasons this is cool:

1) The variance is within a narrow range (i.e only one point of damage different out of 10). This means that you can normally account for both possibilities (unless you are far behind, which is fine), and often either result is fine, and will provide the same general outcome for the particular localized battle while…

2)  The different outcomes most often result in minor differences in the amount of hp surviving units are left with. This is moreso Input Randomness (more desirable) for later turns than output randomness in my opinion. (I still have to write an article outlining my thoughts on the distinction here).

In addition, Advance Wars even attempts and largely succeeds in playing with an effective and player-influenceable information horizon with fog of war (and some additional ambiguity with weather effects). Overall a very strong design achieved with a relatively low complexity in terms of unit types, terrain types and rules.

A Discussion on Ambiguity in Games – Part 2

So, we know that we must have ambiguity in our games, and that the amount of it is important, but that’s not all. There are different ways in which ambiguity is put into games. Some of these seem to be more or less satisfying than others. Some of the ways we create ambiguity in games are with input randomness, output randomness, hidden information, time limits (maybe), incalculability or complexity (maybe) and other players. In many strategy games (especially board games – think puerto rico), the main source of ambiguity is the other players at the table. You can look at their board state and have some idea about what they might do, but you can’t know for sure. This form of ambiguity tends to be quite satisfying (at least for me). However, it can sometimes be less interesting if one player is significantly more skilled than the others, or if a player does not take the game seriously and makes seemingly random decisions.

uncertainty

Randomness is also used in many games to create ambiguity however it can often create a feeling that players lack control or agency in the game. If you made all the correct decisions in a game, but you still lost because dice rolls or card draws went poorly you will likely not feel very good about the game experience (there are other factors which might allow you to enjoy the game despite this such as theme or immersion – more on this in another discussion). I like to further distinguish randomness into input and output randomness. I define these as randomness that happens before your decision (input randomness) and randomness that happens after your decision (output randomness). It is possible to get unlucky with either, however since with input randomness you make your decision after the randomness there tends to be less of a feeling that you didn’t control the outcome (and in fact in many cases you did control the outcome). An example of input randomness is in Axes and Acres at the start of every turn your worker dice are rolled and you will have a certain set of faces to use. You make all your decisions after the randomness has occurred allowing you to mitigate the luck of the roll (at least to some extent). An example of output randomness is the dice rolls in Risk. You choose to attack another country and then you roll to see if you are successful or unsuccessful. Output randomness is poor from a strategic standpoint, but I think the reason many people enjoy it is because of the excitement of not knowing what will happen, similar to the feeling people get from gambling. It also prevents the game becoming too deterministic – there is always a chance the outcome could change no matter how far behind you are. Input and output randomness can both be overused, but are also very useful tools for creating ambiguity in a game (especially input randomness).

Hidden information is another form of ambiguity used in many games. For example in Magic: The Gathering each player has a hand of cards that is unknown to the other player. As the player plays cards you gain a little information to make a reasoned guess at what other cards could be in their deck. For example if they play a Swamp on their first turn you know (in all likelihood) that they have black cards in their deck. As they continue to play you may come to tentative conclusions about other cards that would synergize with the cards that you have seen them play. Some information is hidden, but you have some to base your decisions on as well.

treasure-map

Here at BrainGoodGames we have so far focused on making single-player games (and our next game is going to be single-player too). Ambiguity in single-player games is especially difficult. There is no player interaction, it is difficult in many cases to distinguish hidden information from randomness, and output randomness rarely feels good in a strategically deep game. Input randomness is definitely a weapon of choice, but it is a delicate balance preventing the feeling of the game just being a randomly generated puzzle and also leaving control over the fate of the game in the player’s hands.

This brings us to time limits and incalculability which I feel are both related to one another and both create something that seems like it isn’t truly ambiguity, but is perhaps functionally equivalent for our purposes. To explain the ambiguity created by a time limit we could look at the maze example. Solving a simple maze when you can see the entire thing in front of you has no ambiguity. If however you had to solve the maze in 20 seconds you might not be able to determine the outcome in the time allotted. You might then look at the entrance and the exit of the maze and determine the direction you needed to move in. Each time you come to a junction in the maze you could base your decision for which direction to go on whether it takes you closer to the exit or not. Of course this is sort of the trick of many mazes, you sometimes have to move away from the exit to get closer to it. Anyways, the imposed time limit could create some ambiguity in the decisions since you would not have the time to calculate the correct decision.
maze

Another way to go about this is to make the decisions practically incalculable. For example in Chess there is theoretically always an ideal move, but the calculations required to figure it out are practically impossible. Even the most powerful computers can’t compute all the possibilities from the beginning of a game of chess. If a player cannot calculate the best decision they must make a (sometimes reasoned) guess. If you can give the player a lot of information to base their guess on, while still leaving the solution out of reach this can be another form of ambiguity. There are a couple main dangers with this form of ambiguity. The game can sometimes feel like a calculation rather than a game, or the rules of the game can become so complicated that it is difficult to teach and remember them all. If the player cannot keep all of the rules in their head, the decisions they make will not be based on the actual information they are given. It is very unsatisfying to lose because you forgot one of the rules of the game. To go back to the chess example, it is strategically advantageous to calculate as many moves into the future as possible to give yourself the best chance of winning, but it may not be fun or enjoyable to spend time doing this.

empty-chessboard

With Axes and Acres we used primarily input randomness and incalculability as our forms of ambiguity. I think a combination of these different methods may help avoid some of the pitfalls inherent to each method and create a better game overall. There are likely other forms of ambiguity in games, and I would love to know about any you can think of!

Thanks for reading,

Caleb Friesen