Okay so im starting this thread in order to get to some form of a conclusive answer RE: Street-by-Street EV vs HEM All-in ev. Also to stop the clogging of the low content thread.

Now stinkypete points out a particular (and quite valid) example of a hand where he hero folds and SbyS ev gives an (apparent) incorrect picture of the ev whereas HEM gives the (apparent) correct ev.

Oh and by the way this is followed by an extremely arrogant quote which has me tilted severely as it is (imo) reflective of the attitude of the general population of the world today and why our society is so fked up. But thats another matter. Who is fk is Wangstein anyway, if google doesnt know him I cant imagine he is worth quoting.

On to the example in question.

So what if instead for,

Hand 2:
Recon20k has JJ
stinkypete has AA
board is 223K. $1000 goes in before the turn. stinkypete folds the turn.

How is this any different to the example you gave ?? We never went to showdown so how do you know ?? The answer is you don't.

Now how can we judge over a rather large sample (and yes we do need sample sizes to get accurate pictures of ev, 2 hands just doesn't cut it) that stinkypete or Recon20k are actually making money.

Well one thing we can look at is showdown vs non-showdown winnings. Since we know stinkypete likes to make hero folds we know we can get him off QQ on the turn. So we shove and voila he folds AA. It is quite likely he is losing big in his non-showdown winnings which is reflective of a weak-tight style.

I dont necessarily think this is strictly reflective of a weak tight style as showdown vs non-showdown winnings can mean many things dependent on who your opponents are, but it is certainly worth debate when addressing ev. Typically we would like to break even with our non-showdown winnings, which would indicate that we are making reasonable decisions and we would like to make a majority of our winnings from showdown.

There are some issues regarding bluffing which affect both and needs to be ironed out, but the first question to stinkypete is, is this new scenario any different to the one you posted ?

Now ill give you a quote,

"The greatest obstacle to discovery is not ignorance -- it is the illusion of knowledge."
-Daniel J. Boorstin

my street-by-street EV: -$500
my HEM EV: -$500
my actual result: -$500

no problems here. next example? i'd recommend something that goes to showdown.

i like this one. applies perfectly well. who is this Boorstin fella? a great philosopher no doubt, but does he compare to Wangstein?

So here is an extremely existential quote which has me severely existing and is (imo) reflective of reality in general and why our society is so present. But that's matter.

"I think, therefore I am."
-Rene Descartes

lol, I'm pretty sure Wangstein is just "someone's" private joke.

"I did not have sexual relations with that woman" -bill clinton

"Baluga has the game solved, thus he's able to explain everything in a very systematic, yet simple way" - DeucesCracked Member tdoomx

Hold up. Are you a politician or something? You completely dodged the question. You gave a specific example as to why street-by-street ev is "incorrect".

These are
1) A showdown hand, and
2) A non-showdown hand.

Which you equated to being the same thing.

"i'd recommend something that goes to showdown."
-Stinkypete

lol. Something that goes to showdown is the ONLY thing we can use. This is the whole point.

Am I to assume that you agree there is a vast difference between your example and my alternate example ??
i.e. a street-by-street ev on SHOWDOWN hands is accurate?
If so, case closed and street-by-street is a better measure. If not please enlighten me as to why they are different.

Apparently he was an American historian, professor, attorney, and writer.
http://en.wikipedia.org/wiki/Daniel_J._Boorstin

Don't really know much about the guy but I've always liked the quote. However Wangstein, who the fk is he?

ergh i think i made a mistake, gonna watch the game and comment tomorrow.

the hand you posted does not, by itself, do anything to suggest that either street-by-street EV or HEM EV is better than the other.

i deliberately chose 2 hands that were symmetric in terms of hole cards and the board but played out differently depending on who held the cards to illustrate the point that sbs EV fails while HEM EV is correct.

give me an example where your JJ hand plays out differently with hands reversed and i'll show you why HEM EV is superior to (or at least no worse than) sbs EV. or suggest two different ways to play the hand (one showdown, one non-showdown), and i'll explain to you why HEM EV is a better estimate of their EV difference.


technically it's accurate for the showdown hands on an individual basis. if you could see hole cards for all hands, including non-showdown hands, you could use a street-by-street analysis with no problems.

the problem arises when you try to combine sbs EV in showdown hands and realized results in non-showdown hands. you can't mix them without introducing some serious bias.

if you put in $400 of your stack with 80% equity on the turn and then get sucked out on and call off another $100 on the river, sbs EV will say your EV in the hand is 64% of the pot or a net of +$140.

if you were to fold the river instead, sbs EV would say that your net is -$400 (because you can't see hole cards).

can you not see why this is a problem? playing bad actually increases your sbs EV by $540!

the HEM EVs for these situations are -$500 and -$400, respectively.

so let me ask you, when you call off $100 on the river drawing dead, should your EV decrease by $100 or increase by $540? i personally think the answer is pretty obvious, but the low content thread suggests otherwise.


EDIT: note that if you could always see hole cards, sbs EV would be +$240 for the case where you fold the river. and sbs EV would be fantastic. but as we know, we can't always see hole cards. and we can't selectively use sbs EV on showdown pots because it introduces a whole lotta bias.

RE: non-showdown hands.
By their very nature we will never see them and cannot judge for sure on any particular hand what decision we made was correct or incorrect. As far as i can see, all we can do is take a large sample of non-showdown hands and try to keep them at or above $0.

I think this is a fair assessment of the hand. We called off $100 with 0 equity. but got to that point holding 64% of the equity. Given that it was a small river bet this small loss is reflective of a small mistake.

However, if it had been a full pot river bet it would also be a much bigger mistake. This is reflective of our opponents "reverse implied odds" where he makes a small mistake on the turn for a large payoff on the river.

Yes this is correct. If we fold we lose the pot.

Well you have to appreciate that it was only another $100 into an $800 pot.
It was a small bet and reflected a small mistake. Whereas calling an $800 river bet would be a much bigger mistake.

So the street by street ev for these situations are +$140 and -$400.

Lets look at 2 scenarios with respect to SbS ev.

1) we call off $100 with 0 equity on the river SbS says our EV is +$140.
2) we call off $800 with 0 equity on the river SbS says our EV is -$560.

Say for instance our opponent had some dominated wrap&fd combo with 20% equity. If he misses we bank $400 to non-showdown.

For scenario 1) if he hits and we cannot fold we lose but retain $140 in ev which is reflective of us playing the hand reasonably well. HEM ev however tells us that we misplayed the hand to the tune of -$500. Not an accurate picture im sure you'll agree.

For scenario 2) if he hits and we cannot fold we lose but SbS ev tells us our bad call on the river cost us -$560 in ev. HEM tells us we misplayed the hand to the tune of -$800.

In these two pictures it seems to me that a SbS EV analysis is a better reflection of play that HEM All-In EV.

There certainly are some issues here regarding showdown and non-showdown winnings. For the moment though ill leave it at that.

Oh and btw who the fk is Wangstein?

i've already conceded (though i never disputed it) that sbs EV gives a more accurate picture of your EV for individual hands.

but as i've said again and again, the fact that bad payoffs increase your sbs EV is a HUGE problem. if you're making bad payoffs to increase your sbs EV and your opponents are making good laydowns in similar spots, your sbs EV is being massively inflated. this isn't hard to see from my examples.

the reality is that sbs EV is useless for any meaningful analysis. i can't think of a single useful application for it. HEM EV has tons of them.

The only purpose of any of these algorithms AFAIK is to reduce the variance in a sample of hands.

If SBS EV is more accurate over a large sample than HEM EV then it is superior. And it sounds like you're conceding that it is stinkypete?

I think he just conceded the opposite, saying SBS is more accurate for individual hands.

How about calculating your EV over a sample of hands more accurately that HEM ?

But less accurate for an aggregation of hands? WTF?

Pete, what percentage of the overall luck factor in poker do you think HEM EV corresponds to? Just out of curiosity.

correct.

no, if you'll read my posts, you'll see that i'm saying the exact opposite. (unless i reversed sbs/HEM somewhere... if i did, let me know)

I think he was saying that for SD hands on an individual basis, so an aggregation of hands would compound the deviations of all NSD hands, making results over a sample less accurate.

I think, at least that's what I understood.

It should be really easy to determine which method is more accurate, unless they coincidentally turn out to be super close.

Anyone have a SBS calculator that does graphs? Just compare substantial samples from a couple of different players in HEM and SBS and I imagine one of the two methods gives you a noticeably smoother line.

I mean I would think someone has already done this but I dunno.

Now i am confused.

Maybe you could fill in the blanks here as to your opinion pete,

For a large sample of showdown hands _____ is a better picture of expected value than _____.

it's only valid for showdown hands. so technically it's fine for individual showdown hands. if the rules of poker mandated that you must go to showdown, it would be a reasonable approach. or if you could see everyone's hole cards after the hand is played.

the problem is that it can't be combined with non-showdown hands, so while it might be more "precise" in the sense that it has lower variance, it's less accurate because it has a bias that can't be quantified. there's no way to know what that bias is.

Ah I see your point Pete. Line smoothness would be misleading then. You could still test it with a very very large sample to see which converges closer to actual results.

I can't see how either method would work for non-showdown hands.

don't have the numbers, but it probably wouldn't be hard to look up your HEM EV adjusted standard deviation and your actual standard deviation.