Milfgard March 6, 2015 at 10:41

How to win computer games [3]: now a complete book

We finally added the book “Play for Victory” by the famous cybersportsman and game designer David Sirlin. Below is a squeeze of the most interesting of the third part and links to full chapters (a lot of letters).

In this part: about the styles of players (find yourself), about the optimal behavior in tournaments, and about what skills are needed to become No. 1 in the game. Plus what to do in the unlikely event that you are already a world champion.

Branches of Choice and Instant Prediction

Player skill is the ability to predict the opponent’s moves. For example, to distinguish between bluff and truth as in poker; represent the direction of development of the enemy’s strategy in Starcraft; Feel the situation when the enemy is going to use a specific technique in a fighting game. From the point of view of the game designer, this looks like the need to lay in the game the possibility of the emergence of "development directions" so that players bring their styles to the game. And sometimes, predicting the actions of the enemy gives the beginner an advantage over the professional.

Why? Because the professional gets used to the fact that all players of his level make a certain counterattack on a certain situation - and a beginner will not do it simply out of ignorance, which can give him an unexpected advantage. In the excellent book The Killer's Diary, the main character carried three combat grenades and one dummy. He threw the dummy last and simply went into the room while everyone was pressed into the cabinets. But one gangster was so unlearned that instead of throwing cover, he started firing back ...

So the game has critical points - moments when the outcome of the game is determined. For example, in Starcraft, there is a well-aimed psi storm over a large army, when the enemy’s attention is diverted to an incident in another part of the map. Or a point of choice in MTG, when the main combination combination is played out. Or a chess move, which gives a good advantage for further implementation. In general, I think you yourself feel such moments perfectly.

The good news is that even in fast fighting games, reaction is not the point . The main thing is the ability to recognize such moments and make decisions in advance. And system registration when you are looking for some kind of pattern in the actions of the enemy. For example, in poker, someone can characteristically look away from the cards for a second at the time of collecting a strong combination.

Again, for example, in the same real fencing, system registration is very important. The roughest example is a beginner, after a bunch of two horizontal strokes, can almost always do a stitching step. Knowing this fact allows you not to “catch the reaction” the moment of the stabbing followed by a counterattack, but to “cache” the block with the counterattack even at the moment the second horizontal strike of the enemy begins. The essence of fencing training is the development of basic patterns of movements to automatism. And at the same time, attempts to get rid of any side patterns that arise with such automation.

Here is what David writes about the 1/60 reaction to a series of shots after training a player:

These are the only two cases where I can generally perform a strike from above ... If I attack you and take several punches, you actually do not need to peer at them. But you know that usually after this I perform a fiery blow. You are completely ready for it, and it is clear to you in advance that you will soon have to make a decision on blocking the upper or lower blow. Yeah, there’s a fiery blow, exactly on schedule. Now you know that now I have got the first chance to execute a blow from above. To fulfill all my goals, there are only two possibilities. Your thoughts do not wander among the thousands of options that you will encounter. You are completely focused on the only approaching moment. You are waiting for him. You filter out all the excess information from the screen. ... The only thing you need to see is how the hit from above begins.

One more thing. If you do not do what would be best in the current situation, and the opponent sees this, he begins to think. And at that moment, while he thinks, it is possible to conduct a standard reception, which in a normal situation - if he had not been discouraged - would not have worked.

Player types

The most “tasty” thing for me in David’s experience is the description of metagame features. The fact is that all meta-communities of players are very similar, and having learned to win in one, you can safely transfer the experience with minor changes to the other. Let's start with the types of players.

The first type is a turtle , a concentration on protection.

During tournament fights, I am famous for my patience and tendency to annoy opponents, Ortiz took these techniques to a whole new height. He plays a practically risk-free game, and is content to somehow achieve a small advantage in life points compared to his opponent. Then he, without any hesitation, “runs away” until the end of the match. (That does not allow an opponent to hit him, and when time runs out, according to the rules of the game, the victory is awarded to the one with the most hit points). His infinite patience is driving even the most serious rivals out of himself, as they despair at least to even the score, time is running out.

“They say my chess games should be more interesting. I could play more interesting - and lose at the same time. ”
Tigran Petrosyan

In GameDev, one must take into account that if protection gives an advantage, then the game will be positional - that is, not very spectacular. In large chess tournaments, both final players often do not make risky moves, but wait until the second makes a mistake - this leads to a huge number of draws, which sharply reduces the interest of the audience.

The next type is the aggressor . These are fast, impulsive players who play very spectacularly. Moreover, aggression does not mean an attack. It means a lot of actions that capture the attention of the enemy. Most often, aggression is a challenge to a counterattack, which the player is waiting for.

Some of Marshall's most brilliant moves at first glance resembled typos.

The aggressor is the best type for the entertainment of the game: many actions, many results. According to my feelings, the younger the audience - the more this style affects the subjective sensations of the game. If aggressively playing optimally - the game is like.

The third type is obsessed . These are those who achieve excellence in one style of play, and try to stick to it. Here is a great description of this type:

Yanovsky was prone to elephants, and his rivals knew about it. He confidently developed numerous areas of the game and positions on the board, which used the strengths of his favorite pieces. The fate of a player obsessed with a certain facet of the game - to know this aspect is better than the rest, even better than the strongest players in the world - but to lose something else, being outside his element. Yanovsky’s rivals learned to offer him situations that allowed him to save his precious elephants in exchange for the loss of other important figures. For many years, American chess players called the two elephants “Two Yans”.

It is a good idea to enable players to deeply develop certain branches either within the game (RPG) or at the level of the learning curve. It is used in racing games such as Asian perfectionists show class.

The fourth type is snakes . These are those who use a ton of off-game techniques to act on game states. Remember how Ostap Bender in Vasyuki complained that the grandmaster had a bad habit of smoking cheap foul-smelling tobacco? This is from the same opera.

Snakes are very good for forums around the game, they give rise to flame and dvizhuha.

Fifth game style - analytical or balanced. This lack of attachment in style, when each decision is dictated by a specific situation. From the outside it looks like a fairly simple game.

And the last one is the style of the beast. These are very risky tricks that always work, because the player knows that the enemy will respond to them. The basis of the style is the ability to "read" the opponent and predict his actions, that is, appeal to the very deep variability of the directions of the game. Let me remind you that there is a way to balance the game when the loser has the opportunity to use more risky tricks or strategies (since he still loses) with a chance to achieve a great result. In this case, the player has the opportunity to minimize the “bad” probability, knowing what the opponent will do - which immediately increases the effectiveness of his game several times. This is not a bluff or random, this is psychology and calculation. Remember the software that teaches the neural network to look for patterns in your rock-paper-scissors game? After 100 games, the software plays noticeably better than you, because I found patterns that you don’t even realize. You can return the probability back to the classic 1/3 to win only by entering true randomness in your actions.

Tournaments

What is important to know:

Opponents play with different strategies and stronger than "in the kitchen", because they do not want to show their capabilities before the tournament.
In addition to the rules of the game, you need to know the rules of the tournament and take them into account. For example, a number of distributions in the first round give a pair of a strong player and a beginner. This is done so that strong players do not fly out playing with each other, but stay for dessert, to the finals.
Good to know the customs of tournaments. For example, if it is customary in a community to communicate in a friendly manner before the game - approx. And if not, then it's time to turn on paranoia, they are probing you.
If winning a series is important, you need to catch the moments when your tactics are being calculated, and change what can be changed in the tournament (a deck in MTG from a sideboard, a character in a fighting game, a race in strategy, and so on).
Tournaments show your weak points well. They will beat right at them.
At the same time, when preparing, you need to hone your best quality to the limit. The game is only one way in strategy. Only one fighting character. Only one style. Roughly speaking, if you are in the 5% best in this tactic, it is better to enter 4% of it, than to train another one, where you are in the top 10%.
The metagame is very important. Most often, a correct metagame prediction gives you a big advantage - knowledge of the most common tactics and understanding of the strengths and weaknesses of the main players.

One more thing. David advises to stop playing in general to win . You should have time for research and development - and for this you need to do very different strange (and often looking crazy) things. Winning means a minimum of risk, working at your best tactics. The game of research is the variability of approaches, but at the same time there are very high risks.

Professional Player Skills

Here is what David thinks you need to download:

An ideal knowledge of tournament competitions and their order is the general skills of working with a metagame.
As much information as possible about the current game.
Love for this game is very useful.
Psychological perseverance and the right attitude towards victory, loss, improvement.
Technical skill (usually dexterity).
Ability to adapt.
Knowledge or ability to other games in this direction.
The ability to identify and predict the actions of another player.
Ability to assess the situation, sober analytical thinking.

Pay attention to the last.

This means that there are a large number of players ... who share common beliefs about what is good and effective and what is not. They represent "worldly wisdom" in the game. But at the very edge of the distribution curve there are several players who have different views on what is right. In their understanding, some well-known tactics cost nothing, because they do not work on high-class players.

In the end, you win and become a famous player. What's next?
And then the most interesting:

You will get influence in the metagame and even influence on the game developers. Your experience may form the basis of the next part (or a patch if you are not lucky).
Your opinion becomes significant. David says that for some players this is a reason to become bastards and denigrate opponents. Which, in general, is not very sporty.
When you teach other weak players, pay attention to several important points. First, you can develop unnecessary habits using techniques that do not work against sophisticated players. Secondly, you have to take on the role of a weak player - for the student to perfect the technique, you need to somehow give him such a chance. The choice between progressive training and "throwing into the river from the bridge" without reducing the class of your game is yours.

What does david do

As you carefully read the book, you can pay attention to the characteristic modus operandi of David:

First, he looks closely at the game and evaluates how varied it is enough to enable a trained player to win. David is unlikely to be interested in games that require grinding the real skill for years to advance by 1%, games where the reaction or dexterity remains important (that is, the player’s initial properties, and not what can be “pumped” by strategy and tactics), games, where to achieve the result you need to pay a lot.
Then he plays this game “for pleasure” with average rivals, so as not to get a lot of setbacks right away with strong players. He masters, feels the taste of victory, sees characteristic tactics.
Switching to the best players is giving up most of the newbie mistakes, completely different behaviors, a different understanding of the pace and what you should fight for.
Then he “swings” the game into a deep understanding of mechanics, engages in a kind of debug. The non-obvious features found are used as an addition to the strategy. This "debug" was especially valuable in the era of arcade machines, when the games were released ready and did not receive patches.
After that, David works at the metagame level - scouts out who plays how, what to contrast with, what features, etc. He meets his opponents, learns their styles, learns to read the characters of such players, watches their games and analyzes patterns of behavior.
And, of course, after that he enters the tournament, trying not to worry too much.