Fable
The chess hall lays out the next few moves on a small board beside it
At the old chess house, a gentleman solved endgames for people. His rules were strict: for each move he examined, he moved only one piece, then re-examined the whole board. His lines were reliable, but more and more people waited outside the door for answers.
The apprentices wanted to help, so each guessed the next move. Some guessed quickly, some guessed off the mark, and the master still had to ask each one why they moved that way, which was more tiring than watching himself.
So they changed their approach: they copied the board onto several small boards and tested out several possible combo chains at the same time. The small boards did not decide victory or defeat; they were only responsible for laying out the suspicious future paths.
The master did not recalculate every small thought step by step. He took these combination moves back to the main board and checked once which prefixes could truly stand; those that could stand were adopted continuously, and those that could not were cut off from the first break point.
Some nights, the small board's combo only worked for one or two moves; other nights, a five-move sequence all passed. No matter how many were adopted, the final game was still decided by the master's main judgment.
The curator once asked how this differed from asking a quick apprentice to write a draft. The gentleman said that the small board is not another player; it is exploring possible n-in-a-row moves in parallel under the same set of position constraints.
Later the chess house fixed this process: generate candidate move sequences, validate them in batches, accept the longest reliable prefix, then continue to the next round. The line outside the door grew shorter, and the game records did not become reckless. The master also stipulated that the small board could only propose candidates and could not secretly alter the main board. If the first two moves of a candidate sequence were already wrong, the rest, however brilliant, was discarded along with them; if the first three moves were all solid, they were accepted in one go, saving the time of repeated scrutiny. The apprentices gradually understood that speed did not come from skipping judgment, but from laying out the drafts that awaited judgment in advance. True authority still rested with the main board; it could simply verify more adjacent possibilities at once. The owner also limited the number of small boards. Set out too few, and they were no help; set out too many, and the master spent more time checking candidates. After several trials, the chess house found the right scale, one where the gains from early expansion outweighed the cost of organizing drafts. The master would also review discarded sequences. If a certain type of candidate was always discarded, it meant the small board's trial method needed adjustment.
The master understood that the root of slowness was not that every step was difficult, but that every step had to wait for the previous round to finish. Only by unfolding several future steps first and then having the same judgment verify them in parallel could the waiting be shortened.