The Candidate-Move Method: How to Calculate Less and See More
Many chess players describe calculation as if the goal were to see everything. Faced with a complicated position, they begin one line, abandon it, test another move, return to the first and eventually choose under time pressure. The problem is rarely a complete inability to calculate. It is a failure to decide what deserves calculation.
Strong players do not examine every legal move to equal depth. Research on chess expertise points in the opposite direction: experts identify relevant parts of a position earlier, focus on fewer promising ideas and use pattern recognition to guide search. Calculation remains essential, but it works best after the field has been narrowed.
That is the purpose of candidate moves. Instead of asking, “What happens if I move this piece?” you first build a short list of serious options, then compare them under the same standard. The method is simple enough to use in a tournament and structured enough to improve through review.
Step 1: Read the position before choosing moves
Candidate generation should begin with a position scan, not a move. Check four things:
What changed on the opponent’s last move?
Are either king, queen or any loose pieces immediately vulnerable?
What forcing moves exist for both sides—checks, captures and direct threats?
If there is no tactic, what is the position asking for: development, a pawn break, a better piece, simplification or prevention?
The first question is especially important. An opponent’s move may open a file, abandon a defender, block a retreat square or change the value of an exchange. If you evaluate the board as though nothing changed, you can calculate accurately from the wrong starting point.
The scan should be brief. Its purpose is to define the problem. “My king is under pressure” is more useful than “the position is complicated.” “The d5-square is available, but my back rank is weak” already suggests what candidate moves must address.
Step 2: Generate a small, diverse shortlist
In most non-forced positions, aim for two to four candidates. A useful shortlist often contains different types of moves:
A forcing move: check, capture or concrete threat.
An improving move: activate the worst-placed piece or seize an important square.
A structural move: a pawn break, exchange or transition into a favorable ending.
A prophylactic move: restrict the opponent’s plan before pursuing your own.
This is not a requirement to find one move from every category. If you are in check, the legal replies define the list. If a tactical sequence is forced, positional alternatives may be irrelevant. The categories matter because players often generate three versions of the same idea and mistake quantity for comparison.
Avoid adding a move merely because it looks attractive. Every candidate should have a sentence behind it: “This wins the pinned defender,” “This exchanges my bad bishop,” or “This stops ...e5.” If you cannot explain what the move is trying to achieve, it probably does not yet deserve calculation time.
Step 3: Order the candidates by urgency
Calculate forcing moves first when they can change the position immediately. A check may force only one reply and settle the question quickly. But do not confuse “forcing” with “good.” A useless check that drives the enemy king to a safer square should be rejected before it consumes a full variation.
Next examine moves that answer the opponent’s threat or create an irreversible change: pawn breaks, sacrifices and exchanges. Quiet improving moves usually come later because their consequences depend more on the opponent’s freedom of choice.
This ordering prevents a common error: spending ten minutes refining a positional plan before noticing a tactical shot that makes every plan irrelevant.
Step 4: Calculate against the opponent’s best reply
For each candidate, ask what the opponent would play if they knew your intention. Do not begin with the reply you hope to see. Look first for their checks, captures and threats, then for defensive resources such as counterattacks, intermediate moves and simplification.
Keep the line narrow. Candidate-move calculation is not a tree that must grow forever. Follow the most critical reply until the position becomes stable enough to evaluate. If the opponent has two genuinely different defenses, calculate both; if five moves lead to the same quiet structure, do not spend equal time proving each one.
At the end of the line, stop calculating and evaluate. Compare material, king safety, piece activity, pawn structure and practical difficulty. A move is not superior because its line is longer or more beautiful. It is superior because the resulting position is better—or easier to play under the actual clock conditions.
Why experts appear to “see” moves faster
Eye-tracking studies help explain why this process improves with skill. In one experiment, experts chose better moves faster than intermediate players and used fewer visual fixations. Later research found that experts distinguished relevant from irrelevant regions of a position during their earliest fixations, while novices needed longer before their attention became selective.
A think-aloud replication of Adriaan de Groot’s classic work reached a similar conclusion. Grandmasters and international masters searched more quickly than lower-rated players, but the overall results still supported pattern-recognition models of expertise. The advantage was not unlimited brute-force calculation. It was faster access to useful chess information.
For improving players, this means candidate selection can be trained. Annotate positions before turning on an engine. Write your shortlist, the purpose of each move and the opponent’s strongest reply. Only then compare your process with the game and computer analysis. The engine’s first choice matters less than discovering why your shortlist omitted it.
Step 5: Perform a final blunder check
Once you have chosen a move, briefly look at the board as if it has already been played. What checks, captures and threats would the opponent have? Did the moved piece stop defending something? Did you expose a pin, back-rank problem or intermediate move?
This final check is not a new calculation session. It is a safety filter. If it exposes a serious problem, return to the shortlist; if not, play the move before doubt consumes the rest of the clock.
A practical time rule
Use deeper candidate analysis when the decision is irreversible or the evaluation may swing sharply. Spend less time when one move is forced, several moves transpose, or the position can be handled by a familiar principle. Time management is part of move quality: a nearly perfect decision that leaves no time for the next five positions may be strategically expensive.
The goal of the candidate-move method is not to make chess mechanical. It is to protect creativity from disorder. First understand the position, then generate a small set of different ideas, calculate the opponent’s best resistance and compare the resulting positions. You will not see everything. You will become better at seeing what matters.
Comments