An oddity in Jonathan Little's Preflop charts

[primrose]

I was looking at Jonathan Little's Exploitative Poker Hand Charts. One thing that struck me as odd is the bottom right corner of the RFI range on the BN:



Raising 32s is a bit wild, but what I really want to know is, why 32s over 42s, 52s, and 63s?

32s makes two straights (A2345 and 23456) so it's essentially a double-gapper, same as, e.g., 96s, which we're also raising. And one more than e.g. 84s or 73s, so it makes sense that we're not raising those. However, 42s and 52s and 63s also make two straights each, so why are we opening 32s and not those?

It might just be "solver randomizes and we don't so we had to round some up to 1 and some to 0" but I think there's probably more to it. I asked GPT and it pointed out that 25s is basically not better than 23s because the 5 will never kick anyway. I guess that makes sense. But 36s is a little better. And even so, that still doesn't explain what 32s has above the others. But it might make sense to not be dominated or something?

Idk, this kind of narrow edge case is interesting to me even if it's super niche; curious if anyone has an idea.

 

[monkeytilter]

I would guess because GTO can play 32s for a slightly higher ev than 42/52/63 (..and that's not all about making the best hand post flop). So 32s has made the cut when JL's team has applied some (arbitrary) threshold when choosing which hands to include/exclude.

 

[sandy358]

I was looking at Jonathan Little's Exploitative Poker Hand Charts. One thing that struck me as odd is the bottom right corner of the RFI range on the BN:

View attachment 386251

Raising 32s is a bit wild, but what I really want to know is, why 32s over 42s, 52s, and 63s?

32s makes two straights (A2345 and 23456) so it's essentially a double-gapper, same as, e.g., 96s, which we're also raising. And one more than e.g. 84s or 73s, so it makes sense that we're not raising those. However, 42s and 52s and 63s also make two straights each, so why are we opening 32s and not those?

It might just be "solver randomizes and we don't so we had to round some up to 1 and some to 0" but I think there's probably more to it. I asked GPT and it pointed out that 25s is basically not better than 23s because the 5 will never kick anyway. I guess that makes sense. But 36s is a little better. And even so, that still doesn't explain what 32s has above the others. But it might make sense to not be dominated or something?

Idk, this kind of narrow edge case is interesting to me even if it's super niche; curious if anyone has an idea.

Interestingly enough, GTOW never opens 32s, as well as other low suited 2s from the BTN at any stack size. I suppose it is a minor mistake, JL's ranges are generally alright.

 

[primrose]

It's not necessarily a mistake; the charts are supposd to be exploitative, so it's conceivable that there is some reason why 32s plays a bit better against humans than against the solver.

 

[dannystanks]

It’s an exploit chart so it wouldn’t matter a whole lot between 32s and 42s, if you know your opponent folds a lot you can rfi any two cards.