

Fuseki training on KGS positions // 8k 3k 1k 2d
8k https://lemmingfuseki.com/pos/4157168333041
3k https://lemmingfuseki.com/pos/5960130441496
1k https://lemmingfuseki.com/pos/7328262315497
2d https://lemmingfuseki.com/pos/1619366004292
Looking for feedback on this kind of fuseki training. Yesterday the thought came up to base this on real human games. So here you go. The move suggestions are from KataGo human net. All the other opening positions on the site are still fully KataGo generated. Your rank is estimated after every (good) move you make. The estimator is quite agressive and your rank may jump wildly.
Let me know what you think.
The position ranks above are OGS ranks, all the other ranks on the site are KGS...
Edit1: the title should read OGS not KGS
Opening Training with Leming Fuseki
Hey all,
I'm working on Go opening training problems.
I imagine them like you see in the screenshot. You get some (random) opening position and (as always) you are supposed to pick the move you think is best. Once you have choosen it you get feedback.
I picked M5 here, that gives me a rank increase bc it is liked by stronger players, stronger than my current opening rank of 4.8k, and I get a suggestion how even better players may play here, the triangles show that, esp the darker ones. My opening rank will increase/decrease based on the moves I pick.
This aims at
- finding your current opening rank and
- help you improve your opening.
It is always just about the first move (for now). You can try this yourself here: https://lemmingfuseki.com/opening What are you thinking? Curious for your feedback.
Edit: The ranks are KGS ranks, as used by the KataGo human net.
Edit: Changed the player rank estimation to be bayesian. Should work much better.