Every Go player hits a wall. You play regularly, do problems, watch lectures, and your rank just... stops moving. For some it's the double-digit kyu plateau. For others it's the jump to dan.
The usual advice is "play more games" or "do more tsumego." Those help. But there's a reason players get stuck despite doing both: they keep making the same mistakes without realizing it.
The Plateau Problem
What does a plateau look like from the inside?
You play a game. It feels like it could have gone either way. You're not sure where you went wrong. Maybe the opening felt okay but you lost in the middle game. Maybe you had a good position and it slipped away. The game ends and you think "I'll do better next time."
Next time, you make a similar mistake in a different position. Hard to see the pattern because the board looks different.
That's the problem. Without precise feedback, you can't identify what's holding you back. You end up working on everything a little instead of fixing the specific things that cost you points.
What Changes With AI Review
When you review a game with AI, the vagueness disappears. The analysis shows you exactly which moves lost the most ground. Not approximately, not "somewhere around here," but the specific move and the point cost.
The patterns become visible:
- "I keep losing 5+ points in middle game fighting because my reading breaks down under pressure"
- "My openings are fine, but I consistently play the wrong direction after the first fight"
- "I'm giving away 2-3 points per endgame move because I don't know the correct sequences"
Concrete, measurable patterns that show up across multiple games. Once you can see them, you can fix them.
What to Focus on at Each Level
AI review is useful at every level, but the priorities shift as you get stronger.
Double-Digit Kyu (15k–10k)
At this level, games are often decided by a few big tactical mistakes: missed captures, groups dying unnecessarily, basic endgame blunders. AI review helps you spot these quickly.
Focus on: The 3-5 biggest point losses per game. These are usually tactical: life and death, capturing races, basic tesuji. Fix these and you'll jump several ranks.
Single-Digit Kyu (9k–1k)
The mistakes get subtler. You're not hanging groups anymore, but you're choosing the wrong direction, playing slow moves, or misreading the timing of invasions.
Focus on: Direction of play and whole-board thinking. When the AI says your move lost 4 points, look at why. It's usually a strategic choice, not a reading error. These are the mistakes worth saving to Training Mode for repeated practice.
Dan Level (1d–5d)
At dan level, the margins are thin. The AI might flag moves that lost only 1-2 points, but those small losses add up. Endgame technique, ko timing, and subtle positional judgements start mattering a lot.
Focus on: Patterns across many games rather than individual mistakes. If you review 10 games and the AI flags your endgame play in 8 of them, that's your priority, even if each individual error was small.
The Study Method That Actually Works
Improvement isn't about playing more games. It's about extracting more learning from each game. Here's a method that works:
After every serious game:
- Upload to AI Sensei and let the AI analyze it
- Look at your top 3 mistakes, the moves that lost the most points
- Understand each one. Don't just see the correct move, understand why it's better.
- Save the important ones. Hit "Remember" on positions you want to practice.
- Practice in Training Mode. Drill those positions until the correct response is automatic.
Weekly:
- Review your Training Mode problems. The ones you keep getting wrong are your real weaknesses.
- Look for patterns. Are the same types of mistakes showing up across games?
Monthly:
- Step back and assess. Are the big mistakes from last month still appearing? If not, you've genuinely improved. If so, you need more focused practice on those patterns.
Focus Matters More Than the Tool
The AI is a tool. Powerful, but still a tool. It won't make you stronger by watching it play. It won't improve your reading by showing you variations you don't understand.
What it will do is show you exactly where to focus your limited study time. For most players, that focus is the difference between years of plateau and steady improvement.
Benjamin Teuber, AI Sensei's co-founder, went from strong amateur to German champion while building and using the tool. Not because the AI played for him, but because it showed him precisely where his game needed work.
Start With Your Last Game
You don't need to overhaul your study routine. Just start with one game.
Upload your most recent serious game to ai-sensei.com. Look at the three biggest mistakes. Understand them. Save the most important one to Training Mode.
That's it. One game, three mistakes, one practice position. Do that consistently, and you'll be surprised how quickly the wall starts to crack.
AI Sensei analyzes Go games at every level, from beginner to dan. Free analysis at ai-sensei.com.