Happy new year to everyone!

This is my graph at the start of 2022:



42% of the challenge done.

WR by the stack depth:



The WR doesn't seem to change much based on the stack depth. The main reason I prefer to short stack is the lower std dev and the ability to play more hands/hr.

WR by position;



WR by the number of fish at the table:



You can notice that more fish = more rake/100 hands, but also higher WR, which is expected.

If anyone has any input on this, please let me know.

My main goal for 2022 is to reach 200nl and make profit in those games.

Cheers!

Nice work, OP. Interesting finding with STD dev by stack size. Best of luck next year.

Sorry this might be a dumb question, but is this the total for this year? Or since you started in 2018? Amazing results either way. Happy New Year 2022

How many more hands an hour do you get in while short stacking Vs regular stack sizes? I would've assumed it can make some difference, but didn't think it would be big enough to compensate for the different winrate. Gl with the grind this year.

Your graph reminds me of Nanonoko's back in the day, smooth as silk.
GL in 2022!

Keep going! Consistency is key, you’ll get there!

Optimal win rates

Few years ago I made a goal for myself - to have better win rate in every single area than the top 20 regs in my pool.
It's a pretty ambitious goal that might not even be reachable, but chasing this goal seems like a good way to constantly improve in all areas of poker.

In December 2018 I compared my win rates to those of the top 20 regs for the first time. I looked at the 52 areas that I deemed most important and found out that my win rate was better than theirs in 34 of the 52 areas.
This gave me a score of 65%, which I was happy with.

After that I started focusing on improving in the "red" areas, in which I had suboptimal win rates.

In September 2019 I made another comparison, to see if I've improved and that time my win rate was better in 53/68 areas, which gave me the score of 78%.

Now, in January 2022 I compared the win rates again, for the period since the last comparison. And the results were these: (I've decided to only reveal the win rates for the spots in which I'm leaking)



This time my win rate was better than theirs in 66/78 areas, which gave me the score of 85%.

My main leaks seem to be in some CO spots and especially in my play vs limpers. Any advice on this topic would be helpful. Would some of you share your win rates when facing 1 limper?

According to the comparison I seem to be leaking vs limpers the most from the SB, so I ran some extra sims on this. I'll from now on be using more or less this exact range to isolate from SB vs a 45/7/4 rec (the average rec in my pool):



Do you guys think this seems fine? Recs tend to limp almost identical ranges from all positions, so their position shouldn't really impact my strategy.

Thanks. Best luck to you too!

It's since the start of the challenge. Thanks.

While shortstacking I've been playing up to 1600 hands/hr on the fast tables and 800-1000 hands/hr on the normal cash tables. Even if I'm only playing 20% more hands/hr this way it should still be more profitable after the RB. I've done the math. Lower std dev matters too, as it lets me play longer sessions and move up faster.

Thanks to both. I worked hard on it.

sick OP, glgl on your journey to 100k $

Good luck man!

Looks fine for the most part. I'd always raise the AQo is like the only change I'd make. Did you run a solve for this? It gets weird when they don't ever 3b you. Also with regards to your shown winrates for these filters spots you need to understand there is going to be some confidence interval applied to these and the sample sizes might be small. You'd want to look at sample size and std dev for the filter to get a better idea of what your win rate "range" is for those spots. I'm sure you know that, but just as someone with a 10k hand sample is basically going to have a +/- 20bb/100 winrate, in some filtered spots you might have a very low hand sample and therefore your "winrate" might be x +/- some large value.

You've also made the argument in the past that you've folded profitable hands to increase your $/hr while lowering your $bb/100. This is also going to factor in with these goals. Also, for what it's worth I have the average population's limping ranges from each position for 200z on ignition and the limping ranges between UTG and BTN differ reasonably enough to have a different strategy.

^ Thanks for the reply. For that chart I ran a sim and nodelocked the average IP limping range of recs in my pool + how they respond to raises. In GTO we should surely isolate much tighter and more polarized than this. Raising vs a limp is pretty much the same as 3betting in theory, while vs actual recs it plays more like an open.

I understand that there's going to be some confidence interval in the win rates, but I do have a 1mil+ sample on both me and the top regs, so the results should be somewhat relevant. I for sure seem to be leaking vs limps to some degree (unless that's mainly due to my short stack strategy).

It's interesting that in your sample the limping ranges between UTG and BTN actually differ significantly. Could it be that a decent portion of those limpers aren't complete fish? Because when regs/nits limp, their ranges do actually tend to be extremely strong from EP in my experience. I was, however, creating a strategy specifically vs passive fish here, which (at least in my pool) don't seem to be very positionally aware:



It's hard to get a very big sample on these guys, though, as they don't tend to stick around for very long. There's 49 of them in the alias above and it's still not a huge sample.

Great Stuff ZK. Are you using the 10xbb iso vs limpers strategy in the SB?

Tombos ran some sims with HRC Beta and concluded when deciding between a 5bb iso and 10bb iso. The solver overwhelmingly preferred the latter.

^ My sims prefer a 7bb iso from the SB and 9bb from the BB, 100bb deep.

Iso is basically just a 3bet vs an 1x open. It's best to look at it that way.

Interesting, I think he only gave it two options of 5bb and 10bb in his sims.

At 200BB Deep are you going bigger or smaller?

Bigger, just like with 3bets.

3bets go smaller though

I ran sim for something similar a while back and ran two preflop sims.

1. just with the limping range

2. with limping range and then an estimated locked "3b" range (which was very tight)

in 1. you get a pretty tight iso range

in 2. it stops making sense because if they never raise and fold at least sometimes you kind of go crazy.

In practice I do something ~identical to the chart you showed and think that's a strong baseline.

With regards to the ranges its hard to say and I will state that anonymous games have players doing more weird things sometimes than I think would happen on a named site like PS.

Here is BTN limp (normalized)



and UTG limp (normalized)



In the UTG range you can see it's actually quite a bit more "playable"/stronger hands.

^ Yeah, in anon games I could see regs limp much more often from EP/MP as a slowplay. It might even be a good play, as other regs will tend to iso too wide, thinking the limper is a rec.

It's interesting to see that low pocket pairs are a huge part of player's limping ranges. I didn't know that, so I was probably getting a bit owned there, as I often tend to check AQ type of hands down vs fish on J96 type boards, thinking they won't bluff enough. Betting at some point and making those low PPs fold might actually be the optimal play.

Not according to all the sims I've ever done.

If the 3bets really do go smaller, the deeper we are, wouldn't that mean that we should also be 3betting bigger the shorter we get? So for example 50bb deep we'd 3bet to like 16bb? That would seem very odd.

Tombos made this and it goes over 3b sizings as stacks increase for a HU situation at some point in the video. Not sure where.

https://www.youtube.com/watch?v=MPae2gqkeRw

^ HU is a bit different than BB vs BTN. Same goes for 3betting IP vs OOP. There's always the upper ceiling for any 3b size, which is reached at different stack depths based on the exact situation. For example, the ceiling for the 3bet size HU is reached at around 150bb stack depth, while for BB/SB vs BTN it's reached much later on. My sims consistently show that the optimal 3bet size for BB vs BTN/CO 2.5x open at 200bb stacks is around 20bb:



No rake / 50nl Rake



No rake / 50nl Rake



50nl rake 250bb deep

ZK is it possible your sims have incorrect abstractions? I'm not technical at all but I've never seen 20bb 3bets advocated when 200bb deep. I asked Upswing and they said you go smaller when deeper and GTO Wizard goes smaller when deeper as well.

Also I don't think it's as straight forward and linear as just going bigger when short and smaller when deeper. I think there are SPR thresholds that make it non linear.

I'm using the Simple Preflop solver for sims. Perhaps the guys you asked are using something else and are getting a bit different results because of it? Overall, my sims seem very consistent and give logical results for all inputs I give them. They 3bet to 11bb SB vs BTN and to 13bb BB vs BTN, 100bb deep. When shorter they go smaller, when deeper they go bigger, until around 250bb stacks, after which they start lowering the 3bet size again.

I don't know what could be wrong with my sims as I allow all the relevant sizes preflop and 4 sizes/street postflop and always solve it to a very high accuracy:



I ran many sims for 150bb - 250bb stacks with different inputs and never got any results that preferred 3bets smaller than 16bb BB vs BTN.

Perhaps your friends are making incorrect abstractions? For example, they might not be allowing small enough 4bets, which would probably make it optimal for BB to 3bet smaller. Or they might not be allowing non-all-in 5bets from the BB, which my sims often use 200bb deep.

Oh Simple Postflop has bad abstractions I think, I might assume Simple Preflop does as well.

I realized this when I was using GTOx/Simple Postflop and they had a donking range on UTGvsBB SRP on A54tt. I looked at every solve from my own, others and GTO Wizard and the donking range was 0%.

Interesting though, hopefully we can figure these discrepancies out.

GTOwiz has a donking range there though. In the NL50 complex sim 13% of range is a pot+ donk. Both general and general 2.5x utilize the small donk.

Hi ZKesic, I'm wondering why the smallest size you gave the solver was 16bb? I'm in the process of solving 300bb NL50 at the moment and it seems to prefer a much smaller 3bet, around 10bb.