Concrete analysis
Engine evaluation, candidate moves and principal variations show where the position changed and what tactical resource was available.
Training Methodology
From engine output to a decision you can train.
ChessForgePro is organized around a simple loop: diagnose the recurring problem, practice it deliberately and check whether it improves in later games.
EvidenceDiagnosisPracticeProgress
The training loop
Every recommendation needs evidence and a next action.
Rating and win rate describe outcomes. The training loop looks at the decisions that produced them.
- 01
Collect game evidence
Games, color, result, time control, clock data and opening moves establish what actually happened.
- 02
Evaluate decisions
Stockfish analysis and position features identify costly decisions, missed resources and critical moments.
- 03
Find repeatable patterns
One mistake is an example. Repeated mistakes across games become a training priority.
- 04
Build a short practice route
Forgy connects the strongest evidence to puzzles, opening lines, timed decisions or review positions.
- 05
Check transfer to games
Later games and verified exercises show whether the targeted decision is becoming more reliable.
Move evaluation
Stockfish provides evidence, not the entire lesson.
Opening context
Known course moves and opening positions are interpreted as repertoire knowledge. A playable book move should not become a puzzle only because the engine slightly prefers another plan.
Position meaning
Threats, checks, captures, loose pieces, king safety and move purpose are used to turn a numerical swing into a concrete explanation.
Priority model
A frequent repair matters more than one spectacular mistake.
Training priority is shaped by several signals, not a single average.
SeverityHow much the decision changed the position.
FrequencyHow often the same weakness appears across games.
RecencyWhether the pattern is still present in current play.
TrainabilityWhether there are useful positions or opening lines ready to practice.
TransferWhether the error becomes less common after training.
Confidence matters
Small samples should produce cautious advice.
ChessForgePro uses confidence labels because two games cannot support the same conclusion as fifty. Filters for period, color and time control also change which evidence is relevant.
Explainable actions
You should be able to inspect the position.
Recommendations link back to games, opening lines or exercises. Move explanations can be previewed on the board so the reason is visible rather than hidden behind a score.
Ready to use the loop?
Start training