Different solvers have different outputs?
Join Date: May 2021
Posts: 3
I'm quite (completely) new to this and this discrepancy takes all credibility from the outputs - in my mind.
I ran two solvers - GTO+ and PIO (free) - on simple game tree and they not only have different outputs but are complete opposite.
Where PIO is betting 100%, GTO+ is checking 92%.
Parameters for tree I used:
Pot: 6
Stacks: 50
No rake.
Flop: Qs Jh 2h
OOP: 14.3% (linear)
IP: 45.7% (linear)
Single bet for each betting round.
OOP: calls or folds
IP: checks or bets 50%
So, how it is possible that PIO and GTO+ have different outputs? Also it is possible, that my parameters are not identical somehow, but game trees look and are similar in both.
Join Date: Oct 2013
Posts: 3,667
Your test is wrong, both outputs with the same exact settings and ranges should be pretty much teh same
Join Date: Aug 2015
Posts: 5,678
I have GTO+ checking 69% of the time. Solved to a high accuracy just because it's a very simple tree. I can't reproduce your 92% result for GTO+ by solving to a lower accuracy. I entered the percentages you gave to get the ranges.
Pretty sure you made a mistake with Pio just because why would the weaker range be betting 100%? Also there's just very little incentive to bet something like TT when it's WA/WB and folds out most of the worse hands. There are a lot of hands that don't make sense to bet, TT is just the most obvious one to me.
Join Date: Jan 2020
Posts: 817
Piosolver is less forgiving if you forget to put something in the tree building process. Anything blank is omitted. GTO+ will automatically put whatever is selected as the default bet size in all of the boxes, so you likely just made a mistake in Pio.
