AI full-game example

See the whole game, move by move

This is a static presentation sample; live scores are generated from the current cloud book and engine.

S91Game score
Red S 91Black A 8467 plies4 key momentsCloud book + browser engine

Board

第 34 回合
S 91A 84
楚河汉界

Final position · Red has just moved

Moves

67 plies
========
1.兵七进一
炮2平3
2.炮二平五
象3进5
3.马二进三
车9进1
4.马八进七
车9平2
5.车一平二
马2进4
6.马三退五
马8进9−65
7.车九平八
卒9进1
8.炮八进四
士4进5
9.车二进四
卒1进1−120
10.炮五平六
卒1进1
11.车二平六
炮8退1
12.炮六进六
炮8平4−310
13.炮八平五
车2进8
14.马七退八
炮4退1−82
15.马八进七
卒1进1
16.马七进八
马9进8
17.马八进七
马8进7
18.马五进四
炮4平3
19.兵七进一
马7退6
20.车六进一
马6进5
21.车六退二
马5退6
22.士六进五
车1平2
23.车六平九
前炮进2
24.车九进二
将5平4
25.马四进五
车2进2
26.炮五平六
车2平3
27.车九平七
炮3进3
28.车七退二
炮3进2
29.炮六退五
炮3退1
30.炮六进一
象5退3
31.车七平六
将4平5
32.炮六平五
炮3平2
33.帅五平六
炮2退4
34.马五进四

Commentary & report

AI

Five dimensions

Opening95
Middlegame89
Accuracy92
Conversion90
Opponent88
Key moment

The position turns sharply on move 24; this is a good place to review the recommendation.

View example moves

1. 兵七进一 炮2平3 2. 炮二平五 象3进5 3. 马二进三 车9进1 4. 马八进七 车9平2 5. 车一平二 马2进4 6. 马三退五 马8进9 7. 车九平八 卒9进1 8. 炮八进四 士4进5 9. 车二进四 卒1进1 10. 炮五平六 卒1进1 11. 车二平六 炮8退1 12. 炮六进六 炮8平4 13. 炮八平五 车2进8 14. 马七退八 炮4退1 15. 马八进七 卒1进1 16. 马七进八 马9进8 17. 马八进七 马8进7 18. 马五进四 炮4平3 19. 兵七进一 马7退6 20. 车六进一 马6进5 21. 车六退二 马5退6 22. 士六进五 车1平2 23. 车六平九 前炮进2 24. 车九进二 将5平4 25. 马四进五 车2进2 26. 炮五平六 车2平3 27. 车九平七 炮3进3 28. 车七退二 炮3进2 29. 炮六退五 炮3退1 30. 炮六进一 象5退3 31. 车七平六 将4平5 32. 炮六平五 炮3平2 33. 帅五平六 炮2退4 34. 马五进四

Xiangqi Blunder Checker | Chinese Chess

Use AI analysis to find missed tactics, blunders, and turning points, then understand why a move made the position worse for Chinese chess (xiangqi).

Check Blunders →

A blunder checker should not merely say that a move was bad. It should show where the mistake happened, why it happened, and how to avoid the same pattern later. The most important nodes are usually where the score changes sharply, a threat appears, or an advantage disappears. This page is built for Chinese chess (xiangqi).

Author: Sachess Development Team · Updated: 2026-08-28 · 3 Sources

Highlights

  • Use score changes to locate blunders and questionable moves.
  • Compare the played move with engine candidates to find steadier choices.
  • Classify missed tactics, greed, slow tempo, and other error patterns.
  • Useful for reviewing real games and checking your own analysis lines.

Steps

01

Load the full history or a FEN with moves so the page can replay the position before and after each move.

02

Watch where the engine score changes the most and mark those nodes for priority review.

03

Compare the played move with engine candidates and decide whether the mistake came from tactics, structure, or tempo.

04

Save the conclusion as an error label such as missed check, greedy capture, weak king, or slow transition.

Details

What counts as a missed tactic, blunder, or questionable move

A missed tactic usually means a forcing resource was overlooked: check, capture, pin, or continuous threat. A blunder is more severe and often causes a clear evaluation drop after one move. A questionable move may not lose immediately, but it makes the position harder or structurally worse.

When checking with an engine, do not look only at the sign of the score. The key signal is the size of the swing and the gap between candidate moves. If the played move is far from the best line and the continuation is concrete, the node deserves separate review.

  • Missed tactics usually involve forcing resources.
  • Blunders often come with a clear score drop or direct tactical loss.
  • Questionable moves may not lose at once but reduce future position quality.

How to turn results into training

Knowing that one move was bad is not enough. Give the mistake a precise label: missed opponent check, captured before calculating counterplay, failed to keep initiative, or ignored back-rank weakness. The clearer the label, the easier it is to notice next time.

Review only the two or three most important mistake nodes from a game. Export those nodes as FEN and practice similar positions later. Over time, that is more useful than skimming the whole game.

  • Prioritize the two or three most important errors in each game.
  • Write specific error labels for later review.
  • Saved FEN nodes can become a personal blunder training set.

Common Pitfalls

Pitfall 1: Looking only at the final losing move

The real blunder often appears before the final collapse, so check the turning point.

Pitfall 2: Treating every score swing as a mistake

Some swings come from deeper search, so compare candidates and lines before judging.

Glossary

Missed tactic

A forcing resource for either side that was not seen.

Blunder

A move that clearly worsens the position and changes the result trend.

Examples

Calculation miss

Capturing before seeing the opponent’s continuous checks can make the score collapse at that node.

Training example

Save similar missed-tactic nodes and drill checks and forcing threats first.

FAQ

How do I know whether a move is a blunder

If the score gets clearly worse and there is a concrete better candidate move, it is worth reviewing as a blunder.

Is blunder checking only about score

No. Score is only the entry point; candidate moves, forcing lines, and position structure must support the judgment.

How many mistakes should I check per game

Start with the two or three biggest turning points. More than that can dilute attention.

Can it check opening mistakes

Yes, but opening checks should use both cloud book and engine because some common moves come from practical experience.

More Tools

Sources

Sachess analysis page Pikafish project Pikafish knowledge wiki

Related Pages