Two Candidate Endings Before You Commit

Endgames are planned at the second-to-last rung, and they should be named long before you get there.

Ask a player how they plan the end of a ladder and they will describe the final move. That is one rung too late. Endgames are decided at the second-to-last rung, because that is the first point where the target and the route constrain the same word.

The target is fixed, so the set of words that can reach it in one move is small. Name two of them at the start, and the final choice becomes a preference rather than a search. But the second-to-last word also has to be reachable from the side you are coming from, and that is the constraint everybody forgets. A word that can reach the target is useless if the only corridor to it costs everything you have left.

The practical habit is to name two endings before the middle. Then the middle becomes a question about which corridor is shorter rather than a walk toward a vague direction. When neither ending survives contact, the honest move is to backtrack one rung early. That is much cheaper than backtracking from the edge of the target.

Ladders reward this because endings repeat in shape even when the words change. The same few consonant ends show up as the final hop, month after month. Visit the site on a solved day and count how many of the last two rungs you could have predicted. Then unlimited practice makes the prediction drill cheap enough to run ten times in a sitting.

If the two endings both die, that is not a failure of the plan. It is the plan doing its job early, when there is still budget to change direction. The expensive version of the same discovery happens four rungs later, when the correction costs a route. Naming endings early is cheap insurance, and it is almost never wasted.