PDA

View Full Version : EV in not all in situations



Sky
01-07-2010, 09:18 AM
I've read in the faq http://208.109.95.123/faq/afmmain.aspx

"2) If you get 95% of your stack in on the turn and the final 5% in on the river, then it takes your equity on the street you got all in on. We will be introducing another stat which will try to take into consideration your equity based across all streets which should give a better representation."


How are the plans on implementing this method of calculating ev, i would be really glad to see it.

Patvs
01-07-2010, 09:29 AM
Currently, the main focus is on the multi currency / currency convertor.

Sky
10-24-2010, 06:24 AM
Any progress in that matter?

Patvs
10-24-2010, 01:55 PM
it's not going to be implemented. "EV by street" is biased.



EV by street may appear to be great, but also isn't, read: http://forums.holdemmanager.com/3rd-party-programs-compatible-holdem-manager/16665-enter-sect-7.html


Summary:

The EV "problem" what some people don't understand (HoldemManager uses 'All-in' EV):
A: If a player has 0 outs, or
B: the allin situation takes place on the river there is NO EV DIFFERENCE.
C: If a shortstack goes allin, and is called by two bigstacks who can continue to bet on a later street, the all-in by the shortstack is ignored, and his EV will be calculated on a later street.**
D: If you commit 99% of your stack with the best hand, but your last 1% goes allin with the worst hand, the $EV Difference will be calculated by your entire stack when you commit that last 1%.

**(C) Doing the calculation on the street when the shortstack was actually all-in is unfair, since the situation will often occur where one bigstack gets the other bigstack to fold on a later street. Maybe that bigstack who folded had the best hand on the street the shortstack was all-in. It would lead to biased results if you only calculate it when you can (when all 3 hands go to showdown), and don't when there is a 2 way showdown. For this reason HoldemManager will ignore the shortstacks EV when he goes all-in.

Fozzy: "You can't calculate all-in equity if you don't always know the hands you are up against."

To expand on this: So how will this even out in the long run? Actually it will converge, by winning the hand when a similar hand occurs. For example, if you go allin as a shortstack with AK.... vs KK vs QQ. If you lose the hand, you feel robbed of EV for having 25% preflop equity against these hands.
But if you WIN the hand (board comes Axxxx) the opposite occurs, as the hand is treated as if you always were a 88% favorite (AK on a Axx flop versus KK and QQ)

So as such a hand may seem to screw up your EV (when you lose the hand), it will balance out in the long run.



Why "EV by street" (which people who often see situation D want) is a bad thing:

Example:
- you have AA, you raise to 80% of your stack, donkey calls, flop comes K83 rainbow.
- you then go all-in, no matter the flop, because you're committed.
- out of 100 times, 88 times donkey folds.
- 12 times donkeys calls with a set (33/88/KK).
What shall EV by street wrongly do? It shall do no computations for the 88 times where donkey folded--> "no more calculation".
What shall EV by street do the 12 times where donkey calls with a set? "Show that donkey sucked out and that you got unlucky".

So although you ran really good by having donkey folding 88 times out of 100. EV by street focuses on the 12 times where donkey hit his set and tells that you're running below EV.
This is a well-known gambler fallacy. And this is why "EV by street" is biased.

Note: Tristanblue writes "it's precisely because EV by street does nothing to your adjusted-graph on these cases where the opponent folded that it is wrong". (therefore it is irrelevant it is more accurate than All-In EV when the hand does go to showdown - HoldemManager is only interested in more accurate long term EV results)

But what if your opponent never folds? Suppose there are two players A and B.
Player A has AA, B has KK. (both have $100 stacks). They commit half their stack preflop and the flop comes AK6. Player B (KK has 1 "out") to win the hand.
If I would play this hand I would always make sure I'm allin on the turn.
However Player X always commits the rest of his stack on the flop and turn *except for one dollar*. And he commits on the river.
Of course, 2% of the time, the rivercard is the case King. Player X's EV Diff is always 0. (because he goes allin on the river)
My EV Diff (I go allin on the turn) is -$2 (98 out of 100 times) and +$98 (2 out of 100 times)
So our EV graph actually looks exactly the same after 100 of these hands.
(-2 * 98 + 98 * 2 = also equals 0!) So the EV outcome (in the long-term) is the same no matter how you play the hand. And no matter which type of EV calculation you use.



--



TOURNEY EV
In addition: the $EV Won can be confusing when it's negative and a larger number than the actual buyin, or positive and a larger number than the actual first price money.
You have some extreme situations (often in Double or Nothing and in SUPER TURBO SNGs)
Extreme examples:
$169 buyin SUPER TURBO, $Won = $720 (1th place)---> $EV Won = -$196 (a negative value, that is bigger than the actual buyin)
$5.20 buyin DoN, $Won = $10---> $EV Won = $11.35 (more than you can actually win)
But allowing these values -outside of this range- will result in more accurate long term results.

So EV becomes quite a meaningless number if you focus on one game , but the more you play, the more accurate it gives a representation of your overall "luck".

http://forums.holdemmanager.com/manager-bugs/26732-pokerstars-double-nothing-ev-calculation-wrong.html

TryingHard
09-29-2012, 11:46 AM
Example:
- you have AA, you raise to 80% of your stack, donkey calls, flop comes K83 rainbow.
- you then go all-in, no matter the flop, because you're committed.
- out of 100 times, 88 times donkey folds.
- 12 times donkeys calls with a set (33/88/KK).
What shall EV by street wrongly do? It shall do no computations for the 88 times where donkey folded--> "no more calculation".
What shall EV by street do the 12 times where donkey calls with a set? "Show that donkey sucked out and that you got unlucky".


But shouldn't it balance out in the long run? We are also going to hold that set sometime, and in that situation it will be 88 times fold, and 12 times it will show that we sucked out on him and our EV will go down.

I believe that EV by street is biased, because a lot of good arguments were made. I just have this idea that it will balance out in the long run. Do you have any arguments against this?

Patvs
09-29-2012, 01:04 PM
Both are flawed.
Both will balance out in the long run.

The big downside of All-In EV is, it can screw up your Luck Adjusted Winnings if you focus on an individual hand, or a single game.

TryingHard
09-29-2012, 02:40 PM
If the EV by street will balance out in the long run, why doesn't HM use it instead of All-in EV? It is more accurate in the short run, isn't it.

Patvs
09-30-2012, 02:36 AM
If the EV by street will balance out in the long run, why doesn't HM use it instead of All-in EV? It is more accurate in the short run, isn't it.

Yes, but EV-by-street can only be calculated if you know the holecards of all players of ALL hands.
Otherwise it will result in biased results if you only calculate it when you can. It should be balanced out with the other hands that don't result in a showdown. ... but HoldemManager needs to know all the holecards of these hands, which is not possible.

homof
05-11-2013, 11:42 AM
Many (possibly most of) advanced HM2 customers consider EV-Hero-allin-street is more accurate for Hero in multiway pots in comparison with existing EV-method. And they show examples of hands, where a shortstack-hero goes allin preflop and its called by two bigstacks, which then play postflop...

So please explain - why not to add (in parallel with existing EV) - one more also (EV-Hero-allin-street)?
Then any customer will have a choice: which one to use in reports and graphs.

homof
05-12-2013, 10:12 AM
please answer there: http://forums.holdemmanager.com/general-support/407791-hero-ev.html (for HM2)

Patvs
05-12-2013, 10:13 AM
Many (possibly most of) advanced HM2 customers consider EV-Hero-allin-street is more accurate for Hero in multiway pots in comparison with existing EV-method. And they show examples of hands, where a shortstack-hero goes allin preflop and its called by two bigstacks, which then play postflop...

So please explain - why not to add (in parallel with existing EV) - one more also (EV-Hero-allin-street)?
Then any customer will have a choice: which one to use in reports and graphs.





You're absolutely right the current method is flawed:



When the short stack wins at least a piece of the main pot, the main and side pots are reported separately.
When short stack loses, the main and side pots are combined into a single pot and ALL three players are considered as going AI on the same street - and therein lies the problem.

Whenever players are AI on different streets, always calculate main pot and side pot(s) separately - don't combine them.



We have a bugticket on this issue is HM-6112. I've added your email address to the bug ticket, so you'll be automatically notified when it is fixed

albertovn
05-22-2013, 03:18 PM
Hi,

In long term both converge to the same value: the actual winnings (in long term there is no luck). So thinking only in long term none EV would be necessary. Nevertheless, in short terms I really think the EV by street would be much more accurate (despite not being perfect).

I too think that giving the user the possibility to use any of the EVs (the current one or by street) would be the best (he would know the meanning and limitations of each one, so let him decide which one to show).

I'm sorry for my grammar, I'm not english. Because of that I'm not sure about the current state of this. Will it be implemented? Is it under consideration?

Thank you very much and best regards

Patvs
05-22-2013, 06:02 PM
Read post #4.
It is more accurate if you look at 1 hand which goes to showdown.

But Street by street EV can't be implemented because it can only show you when you were unlucky.
It can't compensate for this (because the compensation-hands don't go to showdown).
"It's precisely because EV by street does nothing to your adjusted-graph on hands where the opponent just folded that it is wrong".

albertovn
05-23-2013, 03:37 AM
OK, I know, but there are several cases when the opponent wouldn't had folded. For example, in a real hand I had KK from pre-flop against A2o. The board where rainbow, low cards, non paired. In the river I had 4bb left on a pot of almost 100bb (SPR on the river less than 0.05!!). If the river would have been X, the opponent would have called the 4bb left with A high in a rainbow, non paired board for sure, but the river where the A and I lost with EV=-50.

I'm just thinking, could it be possible to compensate the hands folded with the folding equity? If the opponent folds the river the 60%, instead of an EV=winnings could be EV=winnings*0.6 (or something similar...). It's a tricky situation, I see the problem... Thank you very much nevertheless.

udbrky
05-25-2013, 11:05 AM
Fold equity can only ever be estimated, it is not an exact science.

Patvs
02-15-2014, 10:51 PM
OK, I know, but there are several cases when the opponent wouldn't had folded.




Read post #4 from this thread, the section on 'But what if your opponent never folds? '.
Also read post #7 and #8.

Patvs
02-18-2014, 08:17 PM
Showdown EV is still WAY more superior to all-in EV. All-in EV is by no means "most fair" as showdown EV takes into account a FAR LARGER sample and breaks it down to street-by-street EV.



It will still not be implemented as it only shows accurate EV for 1 hand.

It makes no sense to show a graph with only showdown hands with 'street by street EV'.

I'll give an extreme example:
-1a Good player raises with AK preflop.
Bad player calls with a crappy suited cards hand.
Flop comes Axx. (bad player flops a flush draw)
Good player bets POT--> bad player calls.
Turn is a blank.
Good player bets POT--> bad player calls.
River: the bad player hits his flush.
Good player bets---> bad player calls and wins the hand.

Street by street EV will show you the most accurate EV for this hand: true
In fact... this hand isn't used at all with All-in EV, as no 'all-in' takes place.
Street by street EV will show you how unlucky you were... since you committed most chips with the best hand
(preflop, on the flop and turn)


1b Opposite hand occurs.
Bad player has AK. Good player calls with TJs.
Flop comes Axx.
Bad player bets POT--> Good player calls. (or folds)
Turn is a blank.
Bad player bets POT--> Good player FOLDS.
(a good player won't call two pot sized bets with his draw, and won't see a river)

1c Another opposite hand is:
Good player raises with AK preflop.
Bad player calls with a crappy suited cards hand.
Flop comes Axx. (bad player flops a flush draw)
Good player bets POT--> bad player calls.
Turn is a blank.
Good player bets POT--> bad player calls.
River: the bad player DOES NOT HIT HIS FLUSH (!)
Good player bets---> bad player folds.


If you filter for 'showdown hands' with street-by-street EV, you'll get a ton of these 1a hands, all showing you how unlucky you are.
But you're not being unlucky. Hand 1c (bad player not hitting his flush, you winning the hand) are more common.
That bad player also deserved street bv street EV preflop with his suited cards, and on the flop and turn when he had his flush draw.

So only showing type 1a hands in a graph is a useless graph.
(whereas showing all All-in EV hands is not.... sure some hands will 'screw' up the graph, but that will even out in the long run)