42 hours/month is fine for a rec player. It's actually about what I get in as a purely rec player (although I'll admit I'm likely on the low end of the scale of hours for rec players in my room, although it's probably a pretty decent life-balance target).

But agreed (and the point I was trying to make), at this rate it could easily take many years before you have a semi-handle on what is going on winrate-tracking wise.

Ggoodlucktohim,imoG

I’ve been playing about 60 hours of $1-$2 a month, and have averaged the following win rates:

July: 7 BBs/hr
August: 49 BBs/hr
September: 11 BBs/hr
October: 28 BBs/hr
November: 10 BBs/hr

So, uh, yeah, 60 hours tells you approximately nothing, lol.

On hell of a run in August though.

My fave stat on my PokerJournal app is my day-of-week breakdown for my 1/3 NL game, Tuesday versus Wednesday:

Tuesday: $29.14/hr over 484 hours
Wednesday: $12.16/hr over 411 hours

There is zero overall difference between a Tuesday night game versus a Wednesday night game over the last 12 years in terms of player pool / conditions / yada yada yada. And yet the difference between the two results-wise is a massive factor of 2.4x.

The conclusion is pretty obvious, imo: lol @ ~500 hour sample sizes.

You could probably find very similar lol results differences over these ~meaningless sample sizes by comparing even numbered days versus odd numbered days, days you wore a hat versus a toque, etc.

GcluelesssamplesizenoobG

ETA: So just for fun, I broke down my 1/3 NL stats by even day versus odd day. Nothing too startling in the results like Tuesday versus Wednesday, but still:

Even Days: $59,752 over 2,618 hours = $22.82/hour
Odd Days: $52,071 over 2,826.33 hours = $18.42/hour

So I've won $4.40/hour = 1.47 bb/hour more on even days than odd days (so even days are about 24% more profitable than odd days).

So here we all are trying to tweak our strategies to eke out that extra bb or so per hour and it turns out the only thing that we should be doing is playing on even days, who knew?

GunfortunatelydidnottrackballcapversustoquedaysG

I dropped $1600 this week, which WAS my bankroll—everything else had been deposited in the bank—so now I’m officially, officially done.

Just on the bad side of some truly vicious variance and cold decks. Ah well. Was going to end in a month anyway (once my new job started). I just wish I’d been able to make a little bit more money for my family before then.

It turns out I WAS just on a hot streak early on. I finish the year with a win rate of 19 Big Blinds an hour—a far cry from the 28 I was at after my first two months!

GL davo, have a nice holiday and enjoy the guaranteed money next year.

Leave us degenerates, or come back later for some fun rec. times

I’ve read that it takes a minimum of 200-1000 hours of live poker to confirm whether one is a winning player. Can someone please share how to make this calculation? I’m sure it would be different for 75% confidence rather than 95% confidence?


Alternatively, what about confirming if one is a losing player? Surely it doesn’t take losing a whole 20-30 buyin bankroll to be reasonably certain (75% confidence)? How many hours do you think it takes to confirm whether one is a losing player? Is the calculation the same as confirming a winning player?

It's hard to say in a vacuum, depends on your playstyle and the games you play in ofc. If you were on the nittier side you need less time, if you were laggy then more. Same with your average opponents' style. When you're generally playing a table full of omcs you don't have much variance, in fact it's very hard to even get stacked unless you are a station. When playing aggressive guys you have to bluffcatch/bluff a lot, and there's a lot of variance in that. You can have months where all your bluffs get looked up and they seem to always have it and vice versa.

I'd rather draw conclusions what plays I see around me. You might see a lot of limp/calling, stationing, not bluffing obvious spots, missing value, only 3betting premiums, or the other end of the spectrum, putting in a lot of money with trash hands, running hopeless bluffs, stacking off super light etc. In this case you probably are winning unless you make the same mistakes. When you are often put in uncomfortable spots, and have no clue what to do and what villains are doing, then you are probably losing in that game.

To conclude if you can realize what you do better than your opponents (you must not be deluded ofc), you are almost surely a winner. You can never be sure judging only by results, but you at least need 500 hours imo, especially being a small winner. I'm a bit better than breaking even in the last ~200 hours in games I'm 100% sure I'm beating. And it's not even that horrible, just losing flips most of the time, getting rivered in big pots, whales donate to others, missing in 3b/4b pots more often than expected etc. Not getting set over setted 3 times a week or losing 10 allins a row to a drunk guy shoving every hand, or some similar horror story that shitregs constantly whine about lol

I had to make a lot of assumptions, like saying it's always 8 handed, and came up with 2824 hands played would be 95% confidence that your ROI is your true ROI, assuming you play 30 hands per hour that would make it about 700 hours

I've merged this in to the thread in which it is discussed. Short answer: it depends on your observed variance and winrate. Longer answer, search for 95% confidence interval in this thread to see the calculation.

I’m actually most curious about how long it takes to confirm someone is a losing player.

20 hours should be enough if the losses are egregious and if one keeps getting it in bad.

But perhaps 100 is also enough if they’re losing at like 10BB/hour?


How long would it take to confirm a 2bb/hour loser?

20 minutes

hmmmm, i actually think it's independent of the players, regardless of how you play, loose passive loose aggressive whatever it all evens out in the end and the calculations should be totally independent of play style, the standard deviation doesn't care about that over the long term, but i'm willing to be wrong

should also be completely independent of win rate too, it's actually very simple calculations if you ignore those factors, the standard deviation is inherent to the game itself and has nothing to do with the players

Um, no. And no offense, but if you are not kidding, I suggest a statistics class, and if you are, I refer you to the no trolling rule for this forum.

Cliffs: SDev doesn't matter in the long run, but it matters a lot in determining what the long run is. The higher your SDev is, the less confidence you can have in your observed winrate and the more hours it will take to be able to build a 95% confidence interval that is all above (or below) a true breakeven "win"rate.

The actual formula to determine the 95% CI bounds in relation to your observed winrate is 2*(standard deviation per hour in BBs)/square root of hours played. It should be very obvious that as the SDev goes down, the hours will go into it less times, making a smaller CI band. It should also be obvious that as hours go up, they will go into the SDev fewer times, also tightening the band.

Additionally, as should be very obvious, if you are trying to know if someone is a winner or loser, than observed winrate matters a ton in how long it will take. Consider for example, someone with an observed winrate of 20bb/hr over a small number of hours who's 95% CI is +or- 10BB/hr. That means that his true winrate is almost certainly between 10 to 30BB/hr. He is a winning player. Now consider a player with all else being equal (same number of hours and CI band), but he has an observed winrate of 7BB/hr. That means his longterm winrate is almost certainly between -3 to 17BB/hr. We don't know yet if he is a winning player.

However, if we take the same two players and give them enough hours to shrink their CI band to +/- 5BB/hr, we would now know that they are both winning players.

You either know you are the biggest winning player at the table or you aren't it. Obsessing over winrates is a pointless effort and you should instead aim to constantly be improving and fixing leaks that will be continually getting smaller. You shouldn't care if you are a slight winner or loser when you aren't going to be a pro in that situation anyway so just keep playing in your spare time and keep putting the study in.

Year end results time:

893 hrs
189 days played
+ $37,910
$42.46/hr
65.1% Win %

20% games @ 1/2 5.5 BB/hr
40% @ 1/3 12.14 BB/hr
40% @ 2/5 12.9 BB/hr

Weighted avg 11.12 BB/hr

2/5 Game availability restricted by 8 table poker room open only 4 days/week

6 month totals returning to poker:

+12323 for 320 hours @ 12.8 bb/hr mostly at 1/3

Year end results after returning part time.

December the variance caught up with me a bit. Cold decks more than bad beats.

261 hrs @ 1/3 11.07 BB/hr 81% win rate.

I know small sample size, but it is what it is.

Year end after starting with $1500 and dipping into the fund multiple times for probably $2k



I donÂ’t do the records thing. I figure it is all variance to a degree.

Hey guys. Happy new year to everyone. I posted my results for this year playing mainly 1/3 at MGM national harbor. It only has April 1st-today tracked, I lost the results from the first 3 months of the year. Not sure what im exactly looking for as a response but any feedback is welcome. I dont think 600 hours is truly a big enough of a sample. I think ive heard people say 1,000-1,500 hours will give you a bit more accurate sample of what your true winrate is. I think I want a bit bigger of a sample and roll before I move up to regging 2/5 but encourage others inputs. Thanks so much!

ownin

How do you figure promotional dollars into your winrate? My casino drops $3/hand into the promotional fund. You get that back in the form of regular high hands. If I don't figure any high hands into my winrate, I'm essentially paying $8/rake. If I'm running hot with high hands, it gives a false impression of win rate.

The promotional fund is essentially high variance, neutral EV.

I include high hand promos to my winrate but do not include bad beat jackpots.

I do note that I won $x in a high hand when it does happen

I record all my promotion wins as notes on my phone, but I do not count them as part of my winrate. Seems to me that's the norm both online and live.

Either way, although promotions are becoming more stingy, if you have enough of a sample size, you can do the math to calculate your hourly expectation for them. Mine hovers at $5 an hour.