Quote:
Originally Posted by TeelXp
- The EV of fold for Kings is equals 0 because you wouldn't lose chips acting that way and the EV of call is positive bacause first player plays with Queens 31.5%.
It doesn't sound right to me: if you are mixing between the 2 actions the solution can only be a Nash equilibrium if the 2 actions you are mixing have the same ev because of the indifference principle.
Obviously there can be a small difference because this is just an epsilon NE, or it may be possible to have bigger differences if one of the action was taken very rarely or if the node was reached very rarely, but this is not the case.
As for the ev of folding we certainly agree and the software does too.
On the other hand, when you call with kings you are sometimes winning against queens, thus getting +30 chips (as compared to folding), but you are also sometimes losing against aces, giving you -10 chips. The indifference principle tells us that opponent is 3 times as likely to show up with aces than with queens, thus making both the call and the fold 0 ev.
The computed solution seems accurate within the nash distance, in fact ip player is betting aces 65.14% of the times and queens 21.66% of the times, ie about 3 times as often. If we use these figures, the ev of calling should result -0.02, still accurate imo, but the software shows an ev of +2.509, which doesn't make sense at all.
I'm adding a screenshot here so that the spot can be reproduced and debugged, the range for both players is QdQh, KdKh, AdAh