Expectimax search
When some of what happens next is your choice and some is luck, look ahead by taking the best of your options and the average of luck's.
7 approaches
Techniques
Fourteen ways to build a player, each explained with a small example before it meets Drop7. Read one, then follow the link to the approaches that tried it.
Heuristic evaluation
When the search has to stop before the game ends, a hand-written formula guesses how good the board it stopped at is.
11 approachesQ-learning and value learning
Keep a number for 'how good is it to do this here', play, and after each step nudge that number toward what actually happened plus the best number available next.
10 approachesN-tuple networks
Look at the board through many small windows, look each window's contents up in a table of learned numbers, and add the numbers.
5 approachesNNUE evaluators
A small neural network built so that when one move changes a few cells, only the affected part of the first layer is recomputed.
4 approachesPolicy gradients
Skip the value; keep a probability for each action, play by sampling, and after a good outcome raise the probability of what you did (lower it after a bad one).
4 approachesEvolutionary optimisation
Try many slightly different versions, keep the ones that did best, make the next generation from them, and repeat; no gradient needed.
6 approachesMonte Carlo tree search
Instead of examining every branch to a fixed depth, spend each unit of thinking where past playouts say it is most worth spending, and grow the tree lopsided.
3 approachesRollouts and policy iteration
Judge a move by playing it out with some ordinary policy many times and averaging; then use that judgement to improve the policy, and repeat.
10 approachesOracles, teachers and distillation
Let a program that can see hidden information make labels, then train a program that cannot see it to copy the labels, and only ever measure the copy.
9 approachesRisk and survival objectives
When the game ends the first time you go broke, the average payout is the wrong thing to maximise; you also have to care about the bad tail.
7 approachesAfterstates
Value the board after your move has resolved but before luck deals the next thing; then you never need to give the evaluator the action, and you never split a value seven ways by what comes next.
3 approachesConstructive planning
Build something on the board that pays nothing now and a lot a few moves later, and use the rise as the trigger.
6 approachesDeterminized planning
Sample several concrete futures, search each one deeply, then combine their advice without letting the search spend information the player will not have.
2 approaches
Every technique group also has a page under Approaches listing the strategies that used it and how each one fared.