EV calculation wrong at HM sng hu omaha?

[Mejx]

Hey,

Quick question.
I play the hu sng Omaha games. Lets say I loose one hu sng. During the sng I got 3 times all in with lets say 45% equity and I loose everytime, than they show me in the GRAPH (not reports) that im still OVERRR EV??? How can it be?

Here a quick hands file from one of the tourneys.
Here I got it all in 2 times with less than 50 % and lost both of the times. So actually I should be under ev. But hm sums up the ev difference and sums it to the tourney entry fee and show that im over EV??

[Mejx]

here it is

[Patvs]

Example HU SNG:
1500 chip start stacks.
$10 buyin.

After some hands
Player A: 750 chips
Player B: 2250 chips
They go allin exactly 50% coinflip
So player A has a 50% to double up and end up back with his $10 equity startstack, and 50% to be eliminated
Player B has a 50% chance to win and 50% to drop down to a $1500 stack.

EV Diff is: equity - the actual outcome of the hand
Player A: Equity = 50% * $10 + 50% * $0 = +$5
Player B: Equity = 50% * $20 (winning!) + 50% * $10 = +$15


So if player A loses the hand (winning $0) his EV Diff is 5-0 = $5
Then player B's EV Diff is 15-20 = -$5.



--


Hey,
During the sng I got 3 times all in with lets say 45% equity and I loose everytime, than they show me in the GRAPH (not reports) that im still OVERRR EV??? How can it be?



Every hand you lose, you had the chance to WIN the SNG. It's irrelevant you had 20% equity or 80% equity: all those hand will result in a positive EV Diff for you. The Luck Adjusted winnings is the SUM of the EV Diffs.
That can result in much stranger results, where the EV Diff is more than you can actually win, or lower than what you can actually lose.


Read the last part ("tourney EV") of this thread: http://forums.holdemmanager.com/manager-general/24059-ev-not-all-situations.html

[Mejx]

ok, here is an example if my hand:

its a 7$ hu omaha hyper sng.

http://i44.tinypic.com/24pkunm.jpg

So here im "givemeyrhand" and i get it in with 42,1 % on the flop.

Holdem manager shows that i actually should lost 549 chips what makes an ev difference of 5,63$ since the entry fee is 7$. ok i get this.

http://i43.tinypic.com/14l1xdt.jpg

So far im lucky. So i should be over EV right?

last all in hand, villain has only 40 in chips and i , :

http://i40.tinypic.com/ip4l7c.jpg

here the ev difference is -0,82 $. this i understand too.

So now comes the crucial point:

Since i have been really lucky in the first all in i should be over ev. But holdem manager adds the 5,63$ in ev difference to the -0,82$ of ev difference so that the total ev difference is 4,81$.

And now it adds this 4,81$ to the 13,70$ which i won, so that my total ev from this torunement is 18,51$.

http://i40.tinypic.com/5l5nr6.jpg

GRaphical it adds the 4,81 of ev difference to my winnigs in the torunementso that it looks like this:

http://i41.tinypic.com/5x7qbp.jpg

But since i won LUCKILY the first hand, where almost all the chips got in, i should be over ev right?

[Patvs]

You shouldn't have lost 549 chips. (cEV Diff = -549)
You should have won 549 less chips than you won.
Pot was 960 chips. (and you paid 10 chips in the small blind)
Your EV was 42,1% (42,1% of 950 = +400 chips)
You actually won the 950 chips pot (960-10). So the cEV Diff = 400-950 = -550

"Diff" always is: what you should have won minus what you actually won.



In your first hand the cEV Diff is accurate.
The $EV Diff is not. It should be negative.

520 (pyrra)
480 (givemeyourhand) (480 effective stack)

500 chips = $7 equity
(42.1% * $13.44) + (57.9% * 0) = +$5.66
You win the hand and end up with $13.44 in equity.
$5.66-$13.44= the real $EV Diff should be -$7.78

Can you attach the original handhistory of this hand?
/2550

[Mejx]

there are so many hands with a wrong ev calculation that i uploaded all my hands from this month. you should see by the ev difference that im right.

to find the hand discussed above, just look for the date.

another hand in this history which is wrong:

12/13/11 03:02:44 I get it in with 8% on the turn and loose with at98 and still the ev difference is 1051 after i lost 1106 chips.

thx for your help.

had to upload the hands on an onther sever since its limited here to 1 mb.

File-Upload.net - Hands.txt (http://www.file-upload.net/download-3955384/Hands.txt.html)

[Patvs]

I've written up a report on this.
HM2 makes the same mistake.

Surprisingly you sent hands where the cEV is accurate and the EV Diff is not.
We received hands before where the opposite occurred (EV Diff accurate with the wrong cEV).



Focussing just on your first hand, I think I know what the mistake is HoldemManager makes:
520 (pyrra)
480 (givemeyourhand) (480 effective stack)
They go all-in on the flop.
500 chips = $7 equity
(42.1% * $13.44) + (57.9% * 0) = +$5.63
You win the hand and end up with $13.44 in equity.

HoldemManager shows your $EV Diff = +$5.63. But that's your equity! "Diff" always is: what you should have won minus what you actually won. So HoldemManager is simply forgetting to subtract the "outcome of the hand" from that.

[Mejx]

HoldemManager shows your $EV Diff = +$5.63. But that's your equity! "Diff" always is: what you should have won minus what you actually won. So HoldemManager is simply forgetting to subtract the "outcome of the hand" from that.

yeah i know now.


Surprisingly you sent hands where the cEV is accurate and the EV Diff is not.

take a look on the hand from 12/13/11 11:49:03.
I get it in with 28% on the turn and win. the ev difference is -2852 what means that i should have max 1-3% win chance on the turn and the $ev diff is Zero ??!!! i think the difference in chips of-2852 chips doesnt lead to a $ev diff of zero.



just take a look on the all in hands in the hands i send to you and you will see that a lot of hands are wrong.

how long is it goint to take to fix the bug?


I've written up a report on this.

ok thx for your help :)

[Patvs]

011/12/13 11:49:03 hand

It shows you have 27.5 equity (correct), but it's wrongly using 0% for the cEV calculation.--> (thus also resulting in no EV Diff)

This type of bug was already written up. It should be fixed as soon as possible.

[Mejx]

another hand which is wrong:

when is the update comming?

greetz

[Mejx]

another hand

[Mejx]

another one

[Patvs]

We're already aware to this bug. Bugticket is: 2550 and is still on the to-fix list.

[Mejx]

any updates?

its so annoying to play with a wrong ev calculation...

if the bugticket is 2550, how long will it approximetly take to fix it??? i mean almost 4 months have passed since i opened this thread.

thanks

[RDB]

Sorry for the delay. This is still scheduled for a future update but I do not have an ETA. However I have forwarded your concerns and I have bumped this issue with the development team.

There is a new reference for this ticket. The internal ticket # on this issue is now HM-1545. We will often release builds on a limited basis on only the forums for a few days first so please check back on this thread http://forums.holdemmanager.com/hm2-releases/ for that ticket #.

If you would like a specific email back when a new build that addresses HM-1545 is officially released please email hm2support@holdemmanager.com with "Notify Me When HM-1545 Is Resolved" as the subject line. We have assigned someone to search for these subject lines after each release, so this should ensure you get notified after this fix has been released.

[Mejx]

any updates??? 5 months past...

[Patvs]

Still an open ticket.
Omaha EV issues have been completely overlooked in the last update.
I already bumped the ticket july 8th for the programmers.