after much hair pulling, my solver correctly solves the river, with results matching both simple postflop and piosolver. :)
Aug. 10, 2021 | 10:12 a.m.
That might explain the really weird stuff I was seeing! I'll try adjusting that...
July 27, 2021 | 4:38 p.m.
I'm trying to write a really simple monte carlo CFR solver. It works with Kuhn poker, which isn't a very big accomplishment but it's a start :P
I've now turned to trying to solve a river. Right now, the tree that I'm building is 100% range for IP and OP, 0 money in the pot, 0 rake, IP can check, raise to 2 or 6 chips, OP can check, raise to 2 or to 6 (so for instance, check, raise to 6, call is part of the tree, but check, raise to 6, raise to 12 is not).
This should in theory be a very tiny tree to solve. I am just not experienced with any existing solver, and I need a solution to check my results against. Would anyone be willing to run this for me? I'd need to see the individual solutions for every hand, not just aggregate stats, to help in debugging.
I plan on actually buying a solver to check my results against once the project is a bit more mature, but for now it seems premature. Any help would be massively appreciated :)
PS: I know the free version of Simple Postflop should be able to handle rivers, but again, since I have no experience using solvers, I'm afraid I'd end up checking against a different solution than what I expect!