Thinking in Bets (part one)
In studying (and modeling) decision-making methods for real-world competitive situations, game theorists have identified two very clear distinctions between games. In this essay, we will discuss the differences between such games and freely comment on the ideas from Annie Duke’s book “Thinking in bets” about real-world decision-making.
.
Complete and incomplete games
To illustrate the concept of complete and incomplete information games, let’s turn to poker and chess’s opposite characteristics.
At any moment, over a chess table, both players are (or can be) completely aware of the status of the game. Pieces do not randomly appear or disappear from the board, there is no third-party interference, climate cataclysms, ways where one player can to deceit another by introducing new pieces, or any kind of on-fly rule changing. After all, there is too little space for luck. Each possible movement in a chess game can be classified as optimal or suboptimal, and judging it as an optimal movement is perfectly possible and computationally feasible.
On the other hand, in incomplete information games a player must make decisions under conditions of uncertainty over time. Fortune (or luck?) has a big role in the outcomes, and any changes in the rules become beyond the expertise of players. It is impossible for a rational player to compute any kind of credible payoff exactly because he can not have full information about the environment. It means that he is unable to create a “model” from reality that can be used to estimate the payoff of his (and adversaries) actions. Poker is a good example of an incomplete information game even though it has fixed rules and has not real-time rule changes.
Incomplete information games have loosely rules, changing scenarios without instantaneous consensus about the outcomes. Diversely, complete information games are much more controlled environments with strictly rules and well defined rounds where players can perform in cadence. It is easy to understand that a lot of real life situations are incomplete information games.
Incomplete information games are challenging not only because you have incomplete knowledge at the decisive moments of playing, but also because it remains “information incomplete” from a future perspective. It is difficult to conclude that a sequence of decisions that led you to a good outcome was just luck, or is a correctly applied strategical/technical procedure. From this, the opposite remains true: you can proceed technically impeccably, making all playbook decisions, and end with a poor result. Diversely, complete information games allow the existence of experts. Ideally, any training system aims to emulate all kinds of key situations in a given knowledge field in order to build the expert individual experience tool set. Such educational methods forge experts as talented and highly cultivated people, which appear like a gifted genius in some fields. From what we discuss until now, complete and incomplete information games require different approaches to be analyzed. In the next sessions, we will discuss incomplete information games in more detail and propose ways to deal with their complexity.