Advanced strategy · Lesson 10 of 12
Advanced Calculation: Compare Candidates on Equal Terms
Published Updated
Calculation is not the ability to stare further into one attractive line. It is the disciplined comparison of plausible moves. Strong players still lose accuracy by generating a narrow candidate set, analysing one branch much deeper than the others, or stopping when their idea succeeds against a cooperative reply. A reliable process combines candidate diversity, best-defence discipline, clean move-order bookkeeping and evaluation at comparable stopping points.
On this page
Generate candidates by function
Not every position needs one move from every category. The categories prevent tunnel vision. If the first move you noticed is a sacrifice, deliberately seek one improving and one prophylactic alternative before calculating deeply.
- Forcing
- Checks, captures and direct threats that may change the position immediately
- Transforming
- Pawn breaks, exchanges and sacrifices that create a new structure
- Improving
- Moves that activate the worst piece or prepare the main plan
- Prophylactic
- Moves that blunt the opponent’s best reply while retaining your idea
Calculate against the move you least want to see
After choosing a candidate, ask what would make it fail. The opponent may ignore your threat, return material, force a queen exchange or insert a check before recapturing. Search those resources before natural-looking replies. If the line still works, its case is stronger; if it does not, you have saved time that would have been spent polishing a fantasy.
At each branch, reset the forcing-move scan. A quiet move can create a new check or intermediate capture that was absent one ply earlier. This is especially important after pawn breaks and exchanges, when lines change at once.
Compare like with like
Carry candidates to a sensible stopping point: the forcing sequence has ended, the tactical contact is resolved, or the resulting structure can be evaluated. Then compare material, king safety, activity, pawn structure and the next useful move. Do not compare a five-move principal variation for your favourite with a one-move impression of its rival. Equal depth is not always possible, but equal seriousness is.
Italian position — which candidate deserves time first?
Black to move can strike with 11...d5, develop with 11...Be6, or reroute with 11...Ne7. Rank the candidates by urgency. For 11...d5, calculate White’s captures, central advances and any attack on a loose piece. For the quieter moves, identify what White would do with a free move.
Answer — calculate the irreversible candidate first
Keep the tree clean
- Say the move sequence silently or use square names; do not rely on a blurred visual trace.
- When two lines transpose, evaluate the shared position once and note the move-order difference.
- After a long branch, return to the original position and reconstruct it before starting the next candidate.
- Stop when additional depth would be guesswork and compare the stable facts you can see.
Train the error, not only the puzzle
Classify each miss as candidate omission, visualisation error, failure to find the best defence, premature stopping or faulty evaluation of the endpoint. A diet of winning combinations mostly trains recognition of attacks. Add exercises in which the correct decision is a quiet move, a liquidation, or the rejection of a tempting sacrifice. Those positions train comparison rather than confirmation.