Dealing with secret information

Epistemic status : high confidence in the meta level slightly lower in the toy level and low in the object level
Secret information in this context means information known to one player (or party) that is NOT known by the other player (or party).

Large amount of economic behavior such as signaling can be thought of as secret information. In this context the secret information can be thought of as a skill such as ability to do math. The owner of the secret information (often known as the agent) uses this information and try’s to show the principle that they posses the quality, through costly signals. (if the signal were not costly everybody would do it therefore it would cease to be a good signal).

Secret information in board games

Board games use secret information all the time,it’s a great way for the game designer to introduce randomness while not sacrificing agency of the players. Take for example the game poker. Texas hold’em gives players 2 hole cards, these hole cards are hidden from the other players. The fact that you know that they DON’t know what your hole cards are has profound impact on gameplay. a player can see themselves as the guy who 3bet rather than as the guy with pocket aces. The secret information in this case is what I refer to as randomly assigned information. With this sort of information, the player can inform their decision to maximise the information asymetry present, and also try to deduce the information that the other players have. This deduction as one may notice from that video is about narrowing the possibility space to determine what isn’t possible. in a game like chess the possibility space of what has happened in the past is already determined. So such deduction about what previously happened is unecessary. The difference is that in a game like poker, the web of gameplay has impact on what the current game state actually is. For example, there is a big difference between starting 100BB deep, and 3 betting for 6 BB (making a total of 12BB in pot) and then betting half pot on a J46 rainbow and getting called, compared to calling a 4x raise and then potting on a J46 Rainbow and getting called, even though at the end there is the same amount of money in the pot and the same cards have shown up. The web of previous actions is connected to the web of actions after.

Utilizing Secret information (in board games)

When a player is dealt secret information that secret information should dictate their play. When playing magic a common play is to bluff attack with a 2/2 into a 4/4 with the assumption that you have Giant Growth in hand, if you do have giant growth then your opponent suffers greatly, but if you get your bluff attack called they will eat your 2/2 with no compensation. So how can a player get their 2/2’s blocked when they have giant growth but not get them blocked when they don’t? Well they can’t, but a player can employ universe assesment (blogpost to be made later) you attack in the situation where if your attack get’s blocked you were going to lose the game anyway, or you had little chance. And you non-bluff attack in the situation where you have giant growth. Notably one does NOT bluff attack in the situation where the game is slightly favorable but they will fall behind if their bluff get’s called, because in that situation the bluff value is low. Notably they aren’t “soul reading” their opponent, you are looking at the current gamestate and determining “does the value in bluffing outweigh the downsides of when bluffs get called? The answer can often be “no” but is also frequently “yes”.

It is also of note that if you are on the reciving end of a bluff attack, one should consider the state of their hand. examples of such reasoning follow

“my hand sucks, I won’t beat him if he has giant growth anyway, but if he doesn’t have giant growth then the edge I gain from blocking the bluff attack is large especially if his hand is garbage as well, then I actually may steal a win from a game that actually is even”

“my hand and deck are awesome in this spot, I can’t lose unless I make a bad block in this spot if I don’t block I lose nothing, but if I do block I could lose the game”

“My hand is ok, if his hand is the type of hand that would bluff attack, then either his hand is attrocious (and wouldn’t win anyway, he’s just making this attack to cover the spots when he has giant growth) or his hand is medium but he has a bad lategame and needs to push through a good amount of extra damage. his hand could be a giant growth hand with a useless land and a 2 drop, which I do beat if I cast my next spell but I don’t beat if I block, if he has a good hand and a giant growth I lose no matter what, so may as well ignore that possibility…”

The last example while complicated should illustrate how complex the decision making can get when you introduce uncertainty and secret information. Players use their own secret information to make the decision and base their decisions on the information they have rather as well as elimination of possible universes from the opponents possibility space.

Using secret information in real life.

In real life we can use secret information in a number of ways, we can take the same principle from board games that our actions reveal information about the information we have. Signaling in economics is when a person makes a decision that reveals information about the type of person they are. You can use signaling to show for example if you are a person with high IQ, by going to App academy or college straight out of high school.

We can also do the reverse, take signals given to us from other’s sets of information about them and use those signals to guide our behavior. The secret information of somebody’s skill as an engineer is important information to obtain. One could hunt for engineers by offering high compensation to those who are willing to take a 1 month trial run where an employer determines the employee’s productivity. Those productive engineers would be delighted to take such an opportunity, but lower quality ones would certainly consider such trial runs to not turn out well. As such only engineers above a certain caliber would even bother showing up.

In situations where competition is zero sum (or nearly so) one can use the lines of reasoning in the magic the gathering examples, go over what the other side Doesn’t have given their previous actions, and go over what you don’t have given your previous actions. Iteration over this process may help leading to understanding what is and isn’t going to happen.

Thank’s for reading my first blog post, I hope this post was useful to you, comment on my blog about what I could do better. Seriously just comment any comment at all no matter what it talks about would be extremely useful to me for future posting.