My solvers says to open 2x from UTG, but 2.5x from BU and 3x from SB.

My understanding is that the underlying reasoning is that early in position we are more likely to face 3bets with more players left to act, so we'd like to risk less.

But couldn't a player compensate by tightening their range to something more defendable to 3bets rather than altering open size? Why not apply a uniform open size and only alter ranges?

No. It's that later position winrates are higher than earlier position winrates. By raising larger IP, you increase your winrate vs BB def. SB raises larger because BB is IP with their discount (posted big blind).

Your solver or your preflop chart?

I use GTOWizard and it shows me recommended open sizing for each position

When you open from early position the strategy revolves around countering 3-bets. Opening small works well because you lose the least when you fold to a raise, and have the option to re-open action.
In late position you're more likely to get called than face a raise, so you can size up. That's why both BTN and SB size up.

If you make the opponent's more passive, (more calls, fewer 3-bets), then UTG tends to size up.
If you make opponent's more aggro, then BTN tends to size down.

But that doesn't explain why we size up in the CO compared to HJ/LJ?

(Don't remember which video this one is from)


I feel like it's not only about the probability of getting 3bet or called, but also about the probability of playing OOP or IP.
We want to size up when it's more likely we get to play IP, because we want to capitalize on our positional advantage. The exception being SB vs BB.

edit: I believe we see something similar with 3bet sizing, where it seems we tend to size down when we are OOP with deeper stacks and size up when we are IP?

OOP 3bet sizing by stack depth from one of your videos:

Another reason probably just has to do how many people are left in the pot. Big raises don't often make sense in multiway pots (pre or post) because of shared burden of defense.

Some evidence for this statement: In a Mississippi AKA BTN straddle game, blinds are (0.5bb / 1bb / 2bb), preflop action goes SB -> BB -> UTG -> ... -> BTN. In these games the opener is guaranteed to be playing OOP so that confounding variable goes away. Yet we still see the optimal RFI sizing increase as we approach the BTN.



(these are the EVs when it folds to them)

Is there no incentive to open the same size from everywhere? Say, always 10 at 1/3?

If we deviate should it be subtle, like 10 everywhere except 15 from button/sb or more drastic like 25 from b/sb

Obviously there are other factors. What do you think is the best approach?

The advantage is that it's easier to remember.

At a 1/3 game you don't need to worry too much about optimal RFI sizing. Probably just adjust it based on how sticky/aggro the table is.

When it comes to preferred 3bet sizes at different stack depths, afaik at certain depths like 150-250 bb deep, preflop solvers actually like to split sizes OOP. Like either very large or very small. Let's say 200 bb deep BTN opens 2.5 bb. If you give the solver an opportunity to choose any size between 8 bb and 20 bb, you start seeing it preferring sizes around 10 bb and less or closer to 20 bb. Rarely something in the between. Most regs don't do that, probably because they never ran any sims for those, or more likely, to keep things less complicated, and I don't blame them. For some reasons the GTOW general sims were not set up like that.

As to the OP, I think from each position one open size preflop strategies are based on which size we get to VPIP the most. Meaning 100 bbd we choose to minraise EP, because then we get to open with a slightly wider range than with a strategy that only opens say 2.2 bb +, and then that size increases slightly by each position towards the blinds.

Solvers ofc never give us any reasonings, but it's probably not bad to assume this has a lot to do with the likelihood of the OR being denied equity (by having to fold vs 3bets). The less other players do that, the wider we get to open. I ran a sim for some of these spots on Rocket some months ago, and I didn't see open raises really sizing up that much when I node locked LP's to have more calls and less 3bets, but the effect on opening frequency was clear. With more accurate sims than what my hardware can do, however, I wouldn't necessarily rule sizing up either, and would love to see some answers to these questions.

Here are the opening frequencies for each position 100 bb deep, taken from GTOW preflop research, 500NL general if I remember correctly. Circled red the highest frequencies for each position (BTN is interesting).

[image]hzcaaLG.PNG[/im..]

The shorter version - from early position you want to play a tighter range for less money because you have positional disadvantage. A smaller open loses less when you get 3bet. From the button you can open wider but you also want to build pots since you have position for the rest of the hand. The SB opens biggest because you are guaranteed to be out of position postflop so you want to either take it down pre or build a pot where your range advantage compensates for the positional disadvantage.

Finding issue with SB advice
I agree with tighter range & less money OOP
I open 10 OOP and 15 in position as standard
Not drastic difference - will also use exploitative sizing based on circumstance.

So, I don’t like jacking it up in the worst OOP

Nobody is giving it to you pre (I guess maybe if you bet big enough) and building a pot only makes sense with AA or KK.

I take SB talk with a grain of salt. It’s vastly different in tournaments, so a lot of mixed advice. Mostly chop in cash, so very little blind vs blind. If I do get involved, it will be cheaply, just like everywhere else OOP.

So many issues come into play: player types, action ahead of you, stack sizes, etc. I’m a raise or fold guy, but I always adapt to circumstances. Like I want in there with a pair of fours if I can get in cheap, but not jacking it up.

If I do come out firing from the SB, you would be wise to fold.

I know this is poker theory and I do appreciate the theory discussion. Just wanted to point out that in soft live games this isn't really relevant. If I'm at a "good" 2/5 table where people will call a $50 raise with the same range they would call a $10 raise with, then I'm playing a tightish range and making it $50 every time. That's going to be immensely more profitable than trying to use "GTO" sizings.

In theory the only reason you put money in the pot is to go after the dead money that's already out there (preflop that means the blinds). In the real world people want to play poker and see flops. Yes you need to beware of regs 3-betting your large opens, but you want the fish to put more money in.

100% valid point for those games. The theory stuff assumes villain is somewhat sizing-aware, which a lot of live rec players simply aren't. If they call $10 and $50 with the same hand, you're leaving a ton on the table by not exploiting that. GTO is a baseline, not a rulebook.