Rec limpers vs. Competent limpers

As someone who does the occasional poker tracking project myself (although usually with regards to game conditions), I have to commend OP for attempting to track his results when using this method. However, rather disappointingly, I doubt it will produce any significant evidence either way. Unfortunately, poker variance and the long term make attempting to prove something using just empirical evidence based on our sole results over the very short terms of live poker very dicey.

I remember when I hit 2000 hours in my 1/3 NL game that I began kicking myself for not keeping track of the amount won/lost with every single hand I played from every position, because no doubt 2000 hours would have given me enough empirical evidence to prove the profit/loss of every hand from every position, right? At the time I was wondering if limping 55 UTG was actually profitable for me in my game, and was so disappointed in myself for not collecting the data that would have proved whether it was. But then I reversed engineered things, and was shocked to discover that whether 55 limped UTG was profitable for me would have boiled down to the results of what happened when I flopped my ~3 sets (the only hands I would be able to continue with postflop OOP in a multiway pot). This also assumed I always got to see a flop, so in reality likely how much I made the ~1.5 times I flopped a set. At the time someone pointed out I could lump 55 in with 66-... thus expanding the sets to somewhere in the neighbourhood of 9 - 18 hands... of course assuming I ran at EV when it came to flopping sets. And furthermore I could lump UTG in with UTG+1 and UTG+2, thus providing a data set of 27 - 54 hands. So on average, the whole profitability of 66- in the first 3 positions would have boiled down to the results of about lol ~40 hands. Sure hope I ran good in those relatively small amount of hands regarding getting some decent stacks in and not getting stacked myself? And to really know whether there was a more profitable method (such as raising large or raising small) I would have had to replay that 2000 hours in the exact same conditions just to get the same handful of small results that I would need in order to compare methods.

The uncomfortable conclusion I came to is that personal empirical evidence gathering might not be too useful when it comes to deciding what method might work best for us in our game. Instead, we might actually just have to go with our instincts, ask ourselves some questions like "do I really think this is a profitable play?" / "decent play?" / "maybe better than the alternatives?" all "for me" and in "my game", and then hope our instincts aren't too far off the truth.

But, good luck to OP with the results collecting and go go go!

Ggoodlucktousall,imoG

40-hands is a reasonable sample for a specific bucket of hole card combos taking a specific action from a subset of positions and flopping a specific type of hand.

The reason you need high 5-figure to low 6-figure hand samples to track things like your overall winrate is precisely because of things like you only get dealt (eg) AA every 200 hands, and who's to say in what position, facing what action, and hitting what flops. So your intuition of what order of magnitude of a sample you need once you hold a bunch of those variables constant is being misapplied.

Now I'm not nearly as much of a statistician as I pretend to be on this site, so I'm not exactly sure of the stdev, much less MOE, much less p-score for different winrates based on a large enough sample where you're playing 66- for a limp from the first 3 seats often enough that you there's a subsample of an average of 40 hands of flopped sets sounds like you could learn a lot from it.

I wouldn't over-extrapolate the lessons just from the bottom-line number you get from that alone, but especially if you had a whole table of data from those hands (with simple binary y/n tick marks for whether you made it to the flop, WTSD, etc) that would be a very data rich endeavor.

Which in my specific case (i.e. profitability of playing small pears in EP based on the results of flopped sets) would have taken me a mere 2000 hours to get the sample size for... i.e. about 4-5 years of recreational play.

Last night I played TT and flopped a set, only for someone else to flop a bigger set, and got stacked. How may thousands of hours of playing TT and flopping middle set would I need to collect in order for the results to suggest that this is likely a +EV situation overall (in spite of last nights results likely skewing this as a -EV spot for... how many more sessions / months / etc.)?

This is kinda my main point to OP. Again, I commend him for attempting to gather empirical evidence... but unfortunately I really doubt it will be useful / he shouldn't hang his hat on results.

GcluelessNLnoobG

Well math doesn't translate to English well and I'm not enough of a statistician to give a good answer here, but it seems plausible to me that a sample of 10 or even 5 would give you a dataset from which you can draw valuable inferences.

The second you start holding a bunch of parts of the poker tree constant, numbers converge quite quickly. I'm working on a large HUD stats project right now, and confidence intervals can produce very narrow confidence intervals with single-digit samples for that specific scenarios.

This is of course assuming you are gathering and analyzing other proximal data. Depending on the parameters, you can make inferences off proximal data a lone. That might seem like an extreme case, but OP isn't even talking about playing baby PPs at all so in a sense this entire conversation is AT BEST proximal...

GG, you make some great points about the limitations of what I am (or in this case was) trying to do. I've had about four more sessions since my last post, and to be honest, I've concluded that it's probably not possible to draw a definitive conclusion because of some of what you've described. In my case, it's been bouncing back and forth. I've been fascinated at how infrequently I can win a single hand with these OOP limps, but when I do win, I will say that it's enough to counter the losses but it hasn't leant itself to any sort of big pot. One of the biggest advantages to being in position is being able to build a pot, but you really can't do that with this. When you get your limp through the pot is relatively small post-flop. At that point you're going to check since everyone is uncapped which means you lose another street of value. Now you've only got the turn and the river to make things happen.

I thought I would do this experiment until I won or lost at least a few hundred in either direction, but it could go on for an extremely long time without me even coming up with the right conclusion. To your point, GG, maybe I luckily smash it once by flopping a straight which skews everything anyway.

At this point it seems to be evening out for me so I'm going to ditch this now in favor of experimenting with other exploits, but I'm glad I tried it and saw that it didn't bring any huge benefits, nor did it drag me into tons of trouble.

Are they really? I mean, they've got to be somewhat capped in most games. I don't know anyone limping JJ+, AK. If someone has a limping range, it means their rfi range is much stronger. And many people still find folds with 28o etc even when they mostly limp. So their range is at least somewhat condensed.

I'm thinking if you have a limping range, leading small would be preferred to checking when you catch a piece of the flop.

In some of these home games I play in, trapping is the norm for some of them, so yes, they would easily limp JJ or AK. They will ultimately light their own money on fire, but it's not uncommon for others to be caught in the conflagration, hence the reason for this unusual thread.

In fact, let me give you a CLASSIC example of what I'm talking about. At a big 1/2 game I play in (it plays more like 2/5) where the stacks may range from a couple of $150 short stacks to $2K or more, there was a hand that I will never forget but is not uncommon for this group. Button straddles to $5. SB folds, BB opens to $20, UTG +1 calls, UTG +2 raises to $40, LJ calls, HJ raises to $100, CO calls and button calls with 3 8 off. I'll say it again...button calls the $100 with 3 8 off. Flop comes 338 and he ultimately wins a huge pot as the rest of the hand plays out. Everyone is incredulous looking at what happened and I ask the player, "Hey 'V'...can you tell me what possessed you to play those cards?" His response, "With all of the betting, I figure they've all got the big cards which means there's nothing left except little cards." OMG...stuff like that is just so painful! Thankfully I wasn't in the hand and V is known to almost always lose by the end of the night, BUT...it's not uncommon for him (and some others) to absolutely defy all logic and cause a lot of table destruction because they're impossible to hand range.