There is a popular card game in Chinese community called “Sheng Ji“. The rules are very involved, especially for first-time players (at least for me at the beginning). After some trial and errors, I summarized the strategy in the following:
1. There are only two teams. (Let’s call them A and B.)
2. A round can only end up in two states: Team A wins, or team B wins.
3. Thus, points can only end up in two states: Team A gets all the points of the round, or team B gets all the points of the round.
That says, in order for a first-time player to play along, there is only one measurement he/she has to perform in his/her turn of playing (besides remembering the ranking of the cards):
Measurement: Is his/her team winning this round? The answer can have only three possible results: YES, NO, NOT SURE.
If YES, the strategy is to maximize the points in this round, i.e., throw out all the point cards in hand which are allowed to play in this round (how to determine what are allowed, is out of the scope of this letter).
If NO, the strategy is to minimize the points in this round, avoid playing point cards if possible, i.e., throw out garbage cards.
If NOT SURE, then the ranking of the actions is in the following order:
a. Trying to win this round, for two reasons, to secure the points in this round, and/or to take the lead for next round.
b. If no chance of winning, then take a bet, if bet on the team will win this round, then play point cards, if not, then play garbage cards.
There are more advanced techniques like communication with the teammates and estimate the winning chance, but those techniques are all for better playing the NOT SURE case, which only takes up maybe around 25% of time. So for a first-time player, following the above simple rules should provide him/her an enjoyable playing experience.
P.S., anyone interested in turning this into an algorithm and run a test to see how robust is the strategy?
A physicist's thoughts 