I get why you bought in for the min, but that had to mean you would be facing wider call ranges and more multiway pots.
At key <100, you can be getting called wider and still be at a range disadvantage. With min buyins, you should be at that point more often.
Also, without the blinds pressure of a tournament, you might tend to win a lot of small pots and lose too many of the big pots.
Yes, it’s really felt like this has worked consistently better with a max buy in than with min, though min allowed me to test more aspects of this particular Hammer variant.
New test run idea… no good idea for a name yet. Will involve normal opening ranges, but then something a bit odd after that. I might also run it at a variety of levels if it does turn out to finish up chips in the initial run.
Post Flop Blindfold results:
Tables played: 1/2 NL Holdem 9 max
Hands played: 1,004
Chips won: $3,092
BB/100 hands: 153.98
Mechanics: Tight pre-opening ranges ranging from {AA-JJ, AK, AQs} under the gun to {AA-22, AK-A7, Axs, KQs-K9s, QJs-Q9s, JTs-J9s, T9s} on the button, with pot sized raises and no limping (except from blinds). With the UG hands I would typically 3!, 4!, 5! and keep putting in raises until I was all in. On looser tables with lots of weak pre-flop raises and shoves, I extended that to TT, 99, and AJs, and sometimes a bit wider if isolated against someone often jamming very wide.
Post flop, I just rolled a six sided dice to decide what to do (though I failed the run a bit, and failed to do this quite a few times… but was still doing it the vast majority of the time).
I kept the pre flop ranges tight, as I wanted pretty good hands to play blind post flop. I didn’t end up liking the run very much, but it gave me some good ideas about some possible random polarization systems that I might try at higher levels. Having so many players post flop so regularly makes it harder to come up with simple post flop rules that feel very effective.
Anyway, it was the first test run in a while. I feel right now like I won’t do another low stakes run for a while… but who knows, mood can be a fickle thing.
Results summary:
If you’re still looking for ideas, try a contrary style. Limp when you would normally open, check when you would bet, call instead of raising, raise instead of calling, and so on.
Is this inspired from the Seinfield TV episode where George becomes successful by doing just the opposite of everything he would normally do things?
No sorry. I have seen people here who check when they make a hand and bet when they don’t. It’s more common than one might think.
I think I’ll kick off a Pacifist run, never betting, only calling or folding. I’m not sure I’ll have the personality to be able to endure the boredom, but I’m curious at what level I’ll be unable to do this and still win.
I expect 1/2 will be a breeze, as just by calling some of the frequent, silly all in moves made with random hands, a very strong positive return should be generated, and in general you often have a few out of control players that just can’t stop bluffing post flop, also. I suspect it will remain playable a quite a bit past that, as well, but I’m not sure I’d be able to win with it even on high stakes??? Anyway, curious to see how the BB/100 rate here compares with other strategies played on similar levels.
I think I’ll only try it on 9 max tables (though I may change my mind here). I’ll try to eventually do at least 2 runs (and hopefully 3 or more) on progressively higher levels.
So it will really be nothing but doing the best I can at call/fold decisions.
Ouch… that was like an extended session with the dentist as a child, after a spell when I hadn’t been brushing or flossing very well. I may not have the psychological prerequisites for this one, LOL.
OK, have started an elite stakes run on 100k/200k. It’s off to a hard start, getting stacked by El-Jog in the first few hands of the run… but oh well, that’s elite stakes. I’m naming this one GTO bot, as I’m trying to implement some basic GTO ideas in the run (though with any simple strategy, it won’t of course be truly GTO).
About halfway through with the run, and having pretty good results. I’m only up a little, but that feels like a pretty good result, especially since El-Jogador has stacked me twice. That’s of course a challenge for any simplified strategy at these levels… you’ll have a few very competent players against whom simple just isn’t good enough.
But anyway, I’m enjoying this one. It’s a nice departure from my recent playing patterns. After this first 1,000 hand run is completed, I might just launch into a second, to see what kind of staying power it has, before posting the detailed mechanics of the strategy.
OK, results in for the first 1,000 hand test drive in quite some time. I think I’ll revert to just playing however I feel for a bit, but I still enjoyed this run. I think it was my first time trying a relatively one dimensional strategy at 100k/200k.
Here are the results. Again, I’ll defer revealing the mechanics until after I’ve done one more test run using the same strategy.
GTO Bot results:
Tables played: 100k/200k NL Holdem 9 max and 6 max
Hands played: 1,006
Chips won: $20,651,134
BB/100 hands: 10.26
Almost done with the second run, with about 50 of the 1,000 hands left to go. I’m running just above break even at this point, which is reasonably satisfying, as this second run got off to a pretty bad start.
I must say, I find this a lot easier to play than many others I’ve tried. With some of the styles I feel like I’m locked in a straight jacket, but here the games just zoom by, and I could easily give this another run or two. Still, since I’ll be posting the mechanics as soon as I finish this run, I’ll probably steer clear of it for a while, unless I give it a run at much lower stakes.
I will go ahead and post part of the mechanics right now. This is the first run to have some specific opening ranges for heads up play (since that gets pretty common at this level as you start new tables).
On the button, it is min-raise or fold, opening a very wide range. Rather than listing the hands I’m raising, it will be much faster to list those I’m folding: J2o, T4o- (T4o, T3o and T2o), 94o-, 84o-, 74o-, 64o-, 52o, 43o-, 32o
As the big blind, I’m folding almost the same hands the button is, facing a 2.5 BB raise or larger. If the raise gets larger, I extend the folding range, and smaller I’m adding in hands like 64o, 52o and 43o. Note that raising frequency will also shape this a bit. If I’m facing someone that is only raising 25% or 30% of hands, and they raise to 3x, Q5o is getting tossed. Facing a limp, my raising range varies a lot depending on the mix of limps and raises I’m seeing, but is typically similar to the 3 betting range, but somewhat wider.
BB 3!:
The button is usually calling 3 bets quite wide (unless their frequency is low), continuing with a little over half of the range, mostly with calls (though again the size of the 3 bet will drive the continuation range wider or narrower). Typical 4 bets:
Post flop mechanics are the same as what will be covered in the next post when the run is over.
GTO Bot Run 2 results:
Tables played: 100k/200k NL Holdem 9 max and 6 max
Hands played: 1,010
Chips won: $5,985,102
BB/100 hands: 2.96
So you can see the win rate at this level has been quite a bit lower than many of the other test runs. With better play from your opponents, you will see a lot more volatility in results, and so I think it wouldn’t be rare to have 1,000 hands were you lose money playing like this. But at the same time, my guess is that both of these runs got a bit unlucky, and that the long term result would be a bit better than either run saw.
GTO Bot mechanics:
A few points I’ve realized I missed…
Not sure what size you are using because of these seemingly conflicting accounts. Can you clarify?
BTW, try smaller sizes from EP-HJ and larger from CO and BTN. At 100bb+ (6-max tables), recommended sizes are 2.25 from LJ and HJ and 2.5x from CO and BTN. I modified this a little bit to go 2.5x from 1st 2 positions and 2.75-3x from CO and BTN once effective stacks are 125bb+. I think you will have better results in your GTO-based runs if you adjust your ranges for this opening size scheme. That assumes people are playing back at you enough. If they aren’t, then larger opens are required to build pots appropriately.
The first quote (below) was from the section on heads up play.
The second was from the main section, dealing with play by seat with 3 to 9 players.
I’m not a big fan of smaller opening raises in deep stacked cash games, especially with a pool of weaker players that will continue with overly wide ranges against larger bets (though you’ve kind of said the same thing above).
I’m not clear on why you’d want to size up on the CO and button, and then size down from earlier seats. To me that seems to encourage a lot of calls from players that will have position on you post flop when you raise small early, and you’ll be losing a lot of value in multi-player pots with hands like QQ and JJ. From the CO and button, I don’t see as much need to try and build a bigger pot pre-flop. I’m happy with a larger SPR that will maximize the post flop positional advantage.
I missed that the 1st part was for HU. My bad
The idea is to play slightly larger pots IP and slightly smaller ones OOP. 0.25bb/hand might not seem like much but it makes a large difference in winrates.
These hands represent a small part of your overall range. How many weaker hands are you opening that benefit from playing slightly smaller pots over time? You want to maximize the EV of your entire range with a single open-size.
You aren’t impacting SPR by adding .25-.5bb to your RFI sizes from these positions. I don’t know why you wouldn’t want to play larger pots in position if you have the option to do so. Why give up the extra EV? Especially in tougher games, position is your main advantage. Don’t consent to playing the same size pots IP and OOP if you don’t have to.
As to whether there are weaker players or whether you have a postflop edge, I didn’t take any of that into account. I was addressing a GTO-based bot style of play (how I interpreted this run). IRL, adjust as you see fit as a means to exploit your opponents.
I suppose this is a complex equation, but this seems backwards to me. By raising larger IP, and smaller OOP, it seems like I’m giving my opponent an easier decision in both cases. I’ve never actually believed your primary goal in poker is to give your opponents hard decisions, as I think the EV of various actions trumps that, but I do think that were other factors are close, making decisions harder for your opponent is a good thing.
So what goes into this equation? Facing a larger raise, there will be a smaller range that continues, and facing a smaller raise, there will be a larger range that continues. Raising IP, I mostly want callers, and I’m less interested in just picking up the blinds (though that is still a good result), and so I feel I win more chips by having opponents play post flop more often. If call rates against a larger raise were completely inelastic, then I agree that I’d just make larger raises.
Out of position, especially in early position, raising ranges can’t be wide, so I don’t have a lot of trash hands that do well in multi-player pots. Raising small if you do have a lot of trash makes sense, but you’re then at risk of losing EV to 3 bets, so I’ve always preferred to just go with fairly tight early ranges, and with a strong range I’m quite happy to charge a higher price, and to reduce my positional disadvantage by reducing SPR.
I’m someone that has started pretty late in trying to pick up a “GTO” approach to poker (which these days I think means a variety of efforts to imitate solvers and stronger programs). I haven’t seen too much in the way of pre-flop raise sizes by position in the books and articles I’ve read, though most of what I have seen has been of the raise smaller IP and larger OOP variety. I know there are two camps here… would you have any sources you’d recommend for your camp? Do you happen to know how Pluribus sized pre-flop by seat?
Edit: one additional thought – do you do the same with 3 bets, sizing larger IP and smaller OOP?