Hi,
played around 100k hands on PLO50/100

Ran massive under EV (around 50 Stacks) and after half way through now I am running massive over EV (also around 50 Stacks).

When I was below EV, I thought I am just unlucky, I can beat this limit.

Now I am way above and thinking "Do I beat this game or am I just lucky?"

Can someone please help me how to understand the adjusted All-In? Is it even accurate as measurement in PLO?

Thanks!

I wouldn't bother with the EV line. It doesn't tell the full story of how you've been running and is of no practical use other than as self-pity porn.

Much better to spend half an hour at the end of the session (or better at the start of the next) to review your big hands and see how well you are ranging villains and how your EV is looking in these big spots.

It is far more productive to focus exclusively on what things you can improve and spend 0 time squirming over variance.

Sent from my SM-G930F using Tapatalk

^it IS a better indicator of real winrate than your real results are, though. So if you want to know if you legit beat a limit or not, looking at EV line is better. If you're down 50 stacks over sample X but EV line is positive you should probably keep playing. If you're up over sample X but EV line is -50 stacks then you should probably reconsider what you're up to...

But All-In Adjusted only counts when the money goes in right?

Suppose in a 100bb game, I can 4bet pre AAJTds to 99bb, V calls, flop is KKx, money goes in against KK72r leaving me with almost 0% equity, but I get running quads Aces and win the pot. EV Graph stays at 0bb, but winnings up a 100bb. So yes I got lucky, but the most money went in, when V was a massive dog. That cannot be seen in the EV Graph, which makes it useless, I guess?

tl;dr



EV graph would be 0.5bb or something, showing what your results will be longterm on that flop. Real win rate will mostly be 0bb, and sometimes 200bb ... both of which are much more inaccurate.

The fact you got unlucky on the flop doesn't mean you "deserve" to suckout or that AI EV is wrong when you do.

Yes but the EV graph is at 0.5bb and the winnings at 100bb, meaning I got very lucky. But the KKxx made the bad call and got lucky in the first place when the majority of the money went in.

By not knowing this hand, everybody would say I am a lucky guy and should stop playing, just by look at the EV graph

in a sample of 100k hands or more these situations will probably be occurring not often enough to eliminate the still siginificant meaning of all in adj. EV. although not every -EV session will mean you played like a donky. and there will be some hands where the equity shift after you got most of the money in will be in your favor too which might further even out these effects

Who says the KK made the bad call? Just because you had AA that hand?
And, to repeat, the fact KK "got lucky" doesn't mean you deserve to get lucky back ... the best representation of your equity on the flop is AIEV and not the 200bb winning re-suckout.

You do understand that your preflop equity isn't 100%, right?

If this hand is representative of the hands you play, they'd be correct.

I haven't given aiev too much thought because it's really not worthy of it but I wholeheartedly disagree. Playing the loose/awful PokerMaster games that I do, I often find myself in 4b pots with an spr of ~.5 and sometimes even .05. aiev measures luck when all-in, it doesn't measure your edge over the competition nor do I think it brings us closer to knowing that figure, either. Since aiev doesn't measure edge over competition, I think looking at our real wr is more accurate.

and this is spot on.

I could be convinced otherwise since I've only spent maybe 30 minutes of my life thinking about this.

I added the bolded part for clarity.

And it's a ridiculous assertion. Absolutely ridiculous. I'm not even going to bother arguing logically against it. You're bat**** insane for claiming it, or just lack absolutely fundamental knowledge of statistics

edit: I'm drunk at the moment and missed this part of your post though, which I do give you great credit for tbh. Sorry for the rest of my reply being so brazen

AIEV takes into account your equity after getting all in. Some people believe this is 'more accurate' than the regular winnings line, but that's simply not true. AIEV is also inaccurate because it fails to take into account which portion of your opponent's range you run into.

Ex: You have AAKKddss, and flop is 234dds. You stack off on flop. In one scenario your opponent has 4456dd, and in the other they have JJTTddss. In scenario 1 you are doing horribly, but in scenario 2 you are doing amazing. Without knowing what your opponent has, though, you are comfortable stacking off either way. You just happened to run into different portions of your opponent's range, and AIEV doesn't take that into account.

In the long run, both lines will even out and end up being virtually the same.

Because it is. If you have 55% equity of $100 then $55 is more accurate than picking $100 55% of the time and $0 45% of the time.

Winnings will also fail to take that into account (as will everything other than a time machine).

Yeh, good argument ... just play a few million hands and the less accurate line is likely to be close to the more accurate line.

Let me do math for you. $100*55% + 0*45% = $100*55%. Wow, it's the same thing! $55 = $55!

I'm not even sure if you are stupid or trolling at this point.

I tell you what, we know that 1,000*$55 is $55,000 and that's obviously the accurate answer ... so I'll happily bet you that 1,000 runs of a 55% shot at $100 isn't $55,000 (and is thus. less accurate).

Gonna guess trolling, but here's me holding out hope.

don't be results oriented and just continually work on your game , look at long term ev and not the swings that happened along the way to see how you are doing

It's so hard to tell good trolling apart from genuine ineptness. Ugh.

You're right that AIEV doesn't take ranges or bad/good flops with low SPR into consideration. But neither does real winnings. As illiterat points out.

In a nutshell: poker is full of variance, and of many kinds. AIEV takes one of those kinds of variance into the equation and removes it. That's the only thing that differs it from real winnings line - it has one element of chance removed from it, compared to real winnings.

If we're comparing two things:
A: real winnings. It has ALL the variance of poker built into it.
B: EV line. It has ALL the variance of poker built into it minus one part of variance which is the luck factor when being all in before the river

B is literally A minus one part of the variance which separates A from real theoretical winrate

edit: and no, in the long run both lines will not even out and end up being virtually the same. This is gambler's fallacy.
edit2: One more attempt I guess:

Imagine three lines in a graph:

x. Theoretical actual real winrate. A magical line which we can never really know, but it's the line which has our true real winrate. It's what we're trying to figure out.
y. Real winrate. The normal green line. This line is line x, but with all the variance of poker added into it.
z. EV line. This is also line x, with all the variance of poker added into it, but then one part of variance has been subtracted

I'm trying to find ways to explain that z will on average be closer to x than y will be. And it will. I'm right and you're wrong.

You write: Yes, EV line is more accurate than the regular winnings line.
This doesn't follow from the previous sentence you wrote. In the first sentence you're talking about wether real winnings or EV line is more accurate. Then you jump to wether AIEV is perfectly/actually accurate. THese are two different things. No one is saying AIEV is accurate. We're saying it's more accurate than real winnings is. You're making a strawman argument. Yes, AIEV is also inaccurate, but less inaccurate than real winnings.

Your posts are well written. Even when they are long there are very readable. I think you should do a well.

Ahem, back to internet keyboard warrior mode. The last statement is wrong. The longer the term, the further the regression to the mean. The gambler's fallacy is relevant to the short term.

Sent from my SM-G930F using Tapatalk

Thank you, it took me A LOT of edits (much more than 2) to make it as clearly readable as I wanted

Yea about the gambler's fallacy part I mean it as "Normal line and EV line do not have to end up equal even in the long run" (unless we mean *forever*, that is) - and that if you have a bad (or good) run, that doesnt mean that the future will even it out for you. You can run really bad for half a million hands, and your expected results over the next 500k hands are not going to be that the lines start converging - the likely thing is that they stay exactly as much apart as they have been since chance doesn't have a memory

if the green line is all the bb i won will it mean that if i have a positive red line my green line will be above my aiev anyway because aiev is some kind of show down winnings but red line is non show down?

No. Your red line changes both lines identically. To repeat:

Great got it, thank you! So im running super good not just good