You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
On my M2 pro 2023 mac mini, gambit-enumpoly only returns 8 equilibria for 2x2x2.nfg:
build/gambit-enumpoly -V contrib/games/2x2x2.nfg
Compute Nash equilibria by solving polynomial systems
Gambit version 16.2.0, Copyright (C) 1994-2024, The Gambit Project
This is free software, distributed under the GNU GPL
candidate,10,10,10
NE,1.000000,0.000000,1.000000,0.000000,1.000000,0.000000
candidate,10,01,01
NE,1.000000,0.000000,0.000000,1.000000,0.000000,1.000000
candidate,01,10,01
NE,0.000000,1.000000,1.000000,0.000000,0.000000,1.000000
candidate,01,01,10
NE,0.000000,1.000000,0.000000,1.000000,1.000000,0.000000
candidate,11,11,10
NE,0.500000,0.500000,0.500000,0.500000,1.000000,0.000000
candidate,11,11,01
candidate,11,10,11
NE,0.333333,0.666667,1.000000,0.000000,0.250000,0.750000
candidate,11,01,11
candidate,10,11,11
candidate,01,11,11
NE,0.000000,1.000000,0.250000,0.750000,0.333333,0.666667
candidate,11,11,11
NE,0.400000,0.600000,0.500000,0.500000,0.333333,0.666667
There are two equilibria with full support. However, on my older MacBook Pro (2017, Intel Core i5), I get the correct output, with two equilibria with full support.
This discrepancy appears to date back quite some time.
Preliminary speculation is that this will have to do with tolerances, as previously in the development of gambit-enumpoly we noted some irregularities with how tolerances were operating.
The text was updated successfully, but these errors were encountered:
On my M2 pro 2023 mac mini,
gambit-enumpoly
only returns 8 equilibria for2x2x2.nfg
:There are two equilibria with full support. However, on my older MacBook Pro (2017, Intel Core i5), I get the correct output, with two equilibria with full support.
This discrepancy appears to date back quite some time.
Preliminary speculation is that this will have to do with tolerances, as previously in the development of
gambit-enumpoly
we noted some irregularities with how tolerances were operating.The text was updated successfully, but these errors were encountered: