Algorithm: An explicit and finite step by step procedure for solving a problem or achieving a required end... --A Dictionary of Philosophy Game Philosophy seems to be the study of how games work. Past articles in EGG have covered this problem from several angles. Games are seen as Information crunching machines that sputter with static, friction, and waste. They process information In many different ways. They seem to have all manner of variety in how they do things. There is a word that pulls this all together, "Algorithm." Algo what? Sounds like something out of mathematics. what could that have to do with the games played in the hobby? Algorithm is a fancy word that basically means "rules." In other words a set of steps that the players go through to play the game. The steps tell the players what to do to resolve certain situations. "Hmmm ... sounds simple, but so what? Why use a fancy word?" The answer is simple, fancy words often "mean" more than simple words. In this case, algorithm means ANY set of procedures to settle a problem. This breaks out of preset thinking and opens the door for creativity. VARIETIES OF ALGORITHMS: The definition of algorithm says that it stands for a set of steps to solve a problem. It does not say what those steps "should" be. One hundred years of hobby gaming has led to a large variety of "ways to solve problems." As time goes on, more and more sets of rules are created. The idea that there is no one way to do things becomes clearer and clearer. Consider the following examples of fun algorithms... A player makes an argument about what he wants to happen. It is supported by reasons from the matrix. The argument is ruled in or out by a dice role. A Tactics II armor unit moves next to an enemy unit. The ratio of combat factors is calculated and a die roll is made. The combat result is determined by comparing the roll and the ratio on a "combat table." A Knight moves two squares, forward and one to the left. It lands on the enemy Queen and captures it. A row of lead figures moves across a table covered by green felt. It faces another row of lead figures. Dice rolls are made to determine "hits" which lead to the removal of figures. A French army in Picardy advances into Burgundy. A German army also also advances there. Neither has support, so both fail to enter the provence. They "bounce." A player explains his plan to the game master. The DM then play acts out what the enemy is doing. The player is not as good at acting as the GM wants so he decides that the enemy attack. A player types his orders into the computer in the prescribed manner. The machine then "tells" him what happened that turn. There is a lot of diversity here. In fact when one really considers it, there is very little common ground between the way these games "solve a problem or achieve a required end..." This defies a common and persistent myth that many gamers seem to have about games. That being that there is a "perfect" set of rules (algorithm) for games (and that all games fall short of this). The examples all come from proven games whose algorithms work and are fun. They deal with many situations but they have their limits. The machine model of games explains what happens 'when a game exceeds the limits. Unfortunately this feeds into the myth of the "perfect algorithm" since (as the myth explains) a "good" game would not have those limits. This is clearly not true. 'The examples are all from "good" games, and all from games. with limits. It is helpful to look at what limits mean to algorithms. GENERALIZABLE ALGORITHMS: Oh no--another fancy word! What does it mean? An algorithm is a tool like a hammer. Hammers are made to drive nails into wood, but one can be used for other jobs as well. A hammer in use has the property of delivering a sudden bolt of pressure to a small place. This can be "generalized" into cracking open coconuts, knocking gadgets back in place, or threatening people. But generalize too far, and one uses a hammer to knock ice off a car window. The window breaks, and the LIMITS of what the tool can be used for are found. Games are tools that recreate life by use of algorithms. Some algorithms, like hex/combat table games, are very good at handling modern warfare. Tactics II works well. Later games grew out of it. Many of Avalon Hill and SPI games GENERALIZED from the basic Tactics II game. The hex/combat factor algorithm is easily altered so that it fits different combat situations. The limit comes when one tries to use a "combat" algorithm to a peace game. Tactics II type games do not handle Ghandi like tactics at all. Consider a few other limits some standard games have. Diplomacy like games do not reflect the radical shifts that happen in wars. In Machiavelli, France can not conquer all of Italy in a season, but In 1494 that is exactly what did happen. Role Play Games often fall flat when playing out social Interactions (such as courting, marriage, and friendships). They are often handled in a superficial and inconsequential way. This despite the fact that social relationships are what make up the bulk of our lives. Miniatures games make armies with big numbers the raison d'etat of both battles and campaigns. Despite the fact that wars are fought for political reasons. Generalizability then rests on how widely an algorithm can be applied to recreate a situation. Some algorithms are widely generalizable. Tactics II like games are a mainstay of boardgames. Role Play Games are widely applied. Miniatures games are played in many historical and fantasy periods. Even Matrix games are applicable to any number of political/social situations. As algorithms go, these procedures produce satisfactory solutions to many games situations. That is their strength. It is significant that the algorithms that generalize the most make up the basic genres of the gaming hobby. In a way, generalizability is a good way of understanding what a genre is. Simply put, a type of rules that can work for a lot of games is a genre. Matrix games have the potential to become a genre, once more MGs are in circulation. Many games use neat little mechanisms that are specifically for that game. Cute tricks, gadgets, and gimics abound in games. Few of them can generalize beyond the game they appear in. This is not necessarily "bad" since generaling is only one of the limits of an algorithm. A game specific algorithm may be good in one game and a bust In another one. If so, so what, hammers axe still good at opening up coconuts. It is easy to come up with algorithms. Anyone can make a simple game. It may not be very good but it can be done. It is more difficult to make a generalizable algorithm. It is not a planned event. They seem to come along after one does a lot of work making other simple algorithms. Algorithms can be judged "bad" when they step beyond their limits to generalize. Some mechanisms, for an algorithm is nothing more than the mechanisms used in a game, can not even work in one game. It is safe to say these are "bad" algorithms. Most do work in one game, but feel not quite right with the way the world works. This is the next test of how good an algorithm is. VALIDITY/REALISM OF ALGORITHMS: In life we often understand what is going on around us by the algorithms we chose to use to interpret the world. Without such yard sticks events become confusing like a code that is garbled. Games use algorithms In the same way. So what happens when the algorithm In use predicts an event that does not fit with what is observable in the world? The algorithm is labeled "unrealistic" or "invalid." Every day events happen that do not fit Into our preset world views/algorithms. since the world is much bigger than we are, humans often find it necessary to change their algorithm of explanation rather than the other way around. This is not the case with games. If the games algorithm seems incongruous with our life algorithm then we call it "unrealistic" or stupid. Historical games are of course the most vulnerable to this kind of label, but any game can be called stupid. All algorithms are at times not In step with what our life algorithm tells us. it is not therefore bad for a game to not be a perfect simulation of life. Validity seems to be more of a question of how much incongruity Is acceptable before the algorithm becomes stupid. Look at the following examples.. Imagine a game where one side's units are not allowed to shoot at the other side, despite their being identical. It seems unfair, but what is more to the point is "why is the algorithm not applied consitently?" Rule: Internal consistency Is a measure of validity. Imagine a game in which "weak" units are able to do more damage that "strong" units. It sound illogical to call something strong that is not so reflected in the algorithm. Rule: Logical order is a measure of validity. Imagine a Battle of the Bulge game in which the rules make it impossible for the Allies to win. This makes it impossible for historical fact to be "recreated." Rule: Allowing but not mandating know possibilities is a measure of validity. Even completely abstract games like checkers must have a valid algorithm to the point of being logical and internally consistent. It is the last measure that "realism" comes in on. Once a game (even a fantasy game) defines the world in a certain way, players begin to have logical expectations about what will happen next. The game must allow the possibility of that happening to satisfy those expectations. Algorithms are value statements about what the game maker thinks can reasonably happen. The word value and valid do not seem to go together too well but in fact they do. Reasonable values (i.e. values I have reasons for having) are likely to have some validity/realism. They state what is likely to be the outcome of a set of events even before playing them out. For example ... If the Zulus stop to fire fight the British rather than charge, then the algorithm that favors British riflery will determine the battle outcome. Many people shy away from making value statements in games. Of course the statements are made anyway, but people cover themselves by saying, "It's all based on indepth historical research." It Is okay to say that an algorithm is weighted based or a value, especially when one has historical reasons to think that way. In the end though, algorithm validity is a subjective judgement. REAL LIFE AND ALGORITHMS: Life is a mysterious thing. certainly we have internal algorithms/world views to try to understand it all, but when it comes right down to it, we do not know what is going to happen next. The actual algorithms of life are never told before hand. Imagine a game in which one starts off knowing, roughly what your resources are. A plan is made out on what to do with the resources, but the algorithm for figuring out what the chance of success is is hidden. The game would be more about exploring the world than fighting a battle. I do not know any algorithms that would do this, but it sounds like it will have to be something completely different. New algorithms are what this article is all about. Back to Experimental Games Group # 11 Table of Contents Back to Experimental Games Group List of Issues Back to MagWeb Master Magazine List © Copyright 1991 by Chris Engle This article appears in MagWeb (Magazine Web) on the Internet World Wide Web. Other military history articles and gaming articles are available at http://www.magweb.com