I'm going to post what Matt Marinelli posted on his twitter (#1 online cash game player in America).

He simulated 6 bots at 0 rake to see the discrepancies in winrates over certain hand periods. Here are the results.



This means you could play 500k hands and break even over that time and you wouldn't know if you were a 6.5bb winner or -6.5bb loser

This is why Thought Process>Winrate.

Discuss.

I think your understanding statistical theory needs to improve. In this example, everyone has a win rate of zero, so the difference will be determined by luck (and no information is given as to what the size of the standard deviation is). But if you were in a game where some players played poorly and others played well. the expert players would not need such a long length of time to show a positive win rate. This length of time again will also depend on the standard deviation. See my book Gambling Theory and Other Topics - Expanded Edition for more discussion.

Mason

I understand how win rates work but thank you for shilling your book.

I'm saying your results vs regs are mostly determined by luck (i.e. fast fold pools)

If you are playing vs a fish that has a -30bb/100 winrate than this bot exercise isn't applicable.

The point of the chart is to show how much game selection matters.

Happy Easter!

Well, yeah, to a degree your results are determined by luck, and in some cases maybe entirely. In heads-up someone could be the better player but because of card distribution keep running into spots where they have a good hand but the other player has a stronger hand, this can also happen in other formats, but I played mostly heads-up so I understand it better there. You can get completely destroyed by the deck, because without physical reads how can you possibly put your opponent on a hand that has you pipped well enough to fold a made hand, unless there's action that is absolutely defining? A lot of the time it is just raise from button, call from bb, check/call, check/call, check/call. Even in action heavy hands one person's hand is invariably better, you have to somehow exercise precise discernment when villain can have a range of hands against your made hand. If you think about it, there's no winning if the deck just favors the other person, which is literally what happens every time. You can still win if your opponent makes enough mistakes though, or you somehow play so well to beat a good strategy.

It's true, reg battling is mostly gambling/a fun hobby. You can have big edges over regs but generally only when they're orders of magnitude worse than your average opponent.
Comparing winrates of winning players, say 3bb 100 vs 6bb 100, even over big samples is almost completely useless

You're showing that in games where you're a break even player they're not really worth playing. I don't think you need a simulation to see that.

Also in my book, one of the chapters is titled "Win Rate Accuracy." That chapter will do you much more good than this silly simulation, and it'll apply if there are some big losers in the game.

I also find it interesting when people who are charging a lot of money for their coaching seem to have little understanding of this stuff, especially since it was written up over 35 years ago.

Mason

Good point, except that being a "reg" is a matter of showing up on a regular basis. That someone is a "reg" likely has a surprisingly low correlation to whether or not they are a cash game winner against specific opponents. There are plenty of "regs" who play badly or seemingly play in a robotic manner against all opponents.and may be open to being exploited.

Okay. While what you're saying can be accurate in the short-run because the standard deviation will dominate your expectation, over time (or a large number of hands) this will reverse, and statistical theory tells us this is the case. So when you've played for a long time and are still losing, it means that the standard deviation is either exceptionally large or your expectation is not positive (or perhaps both).

Mason

I think in most cases this will be correct.

Mason

When you finish playing, you can count your money. That is the best way to know if you were a winner.

THAT is the "Process". "Everything else is just commentary" - Hillel.

Repeat every time you play so as to obtain a large enough sample to know if you are a winner over time - Cf. Kenny Rogers ?

If two opponents are in a close range of skill level I don't know if the level of skill advantage is even enough to overcome card distribution as it might occur commonly heads-up. I don't understand the probabilities and different nuances well enough. It just seems like one person is being given the game/session because of how the hands are dealt regardless of the skill advantage if both players are playing decent strategies. I only played low stakes though. I suppose over extremely large samples this could all be negated though.

In the short-run, you're certainly correct. But as the sample size gets large, the skill difference between the players will almost always show. However, if the skill difference between players is very small, it can take a very long time to see (for sure) the skill difference.

Here's an example. Suppose Player A has been winning at $30 an hour and Player B has been winning at $25 an hour. Can we conclude that Player A is better than Player B. In almost all cases, you can't make this conclusion

Mason

All in all you need one million hands for confidence, not happening me. Just want to place myself favorably if luck hits. And being able to pay if not reaching the money in my MTTs, like, ever.

No. There's no magic number needed for confidence. What does happen is that the more hands you play the smaller the confidence interval (that you choose) surrounding your estimate will become.

Mason

Yes, but look at the simulation.

I did look at the simulation and it shows exactly what you would expect (which means that the simulation was probably done correctly).

Mason

Anyhow, agree with OP that thought process is more important than win rate. Winning is just a bit more probable when playing well. Playing badly though shows much faster, if it's true "fish" can have -50bb/100.

Yeah, but with heads-up, with the streakiness of big hands and all-in pots (where one person is basically just pipping the other person), is it possible to get a $ won edge over your opponent if you only have a small edge? I don't understand things well enough to say for certain. The reason this matters is because, if cards were distributed evenly over a large enough sample someone would invariably win more if they had an edge somewhere with bet sizing or whatever, but with the streakiness of how large pots and all-in pots work, maybe that's negated? I don't know.

"Winning is just a bit more probable when playing well" providing we're only thinking in terms of short sessions (perhaps 4 hours or less). But if we're thinking in terms of a longer length of time (perhaps 100 hours of live play) and you play well compared to the other players then being ahead becomes highly probable.

Mason

Then what the hell is the point of this thread?

OP: "Your poker results are determined by luck"

Mason: "Yes, if 2 opponents have identical WRs, but if one is a fish at -30bb/100 and the other is a reg, the reg will win, and you don't need a massive sample to see this"

OP: "Yes, this is why game selection is important, which is my point"

Wow, great thread, we've restated the obvious here.

View: If you go outside when it's raining (without an umbrella) you will get wet

OP you're a good poster around here and a solid poker player, you know better than to start vague NVG chicken vs egg threads

And the obvious has very little to do with the thread title.


In before someone brings up pergolas, palapas, canopies, gazebos...

Imagine being a poker coach and making a thread like this as if youve made some profound game breaking discovery. Lol @ your students.

It's a valid thread. Some people might be blown away by the winrates some people have (especially at nosebleed stakes) and not understand variance might play a huge role in it.

He didn’t say he made a profound game breaking discovery. In fact, he said he is posting this based on a Twitter post somebody else made. It’s not even his original content.

Furthermore, he said he wants it to be discussed, which it is. So his thread is successful with his intended goal.

Imagine making a snarky response to try to embarrass someone only to demonstrate your own poor reading comprehension. Lol @ kebab.

If you suck yes, if not and are able to beat games at 10bb+ which is totally doable at lower stake MTTs and cash, then no.

Any serious player will always supplement their higher buyins with lower buyins.