PDA

View Full Version : EV$ calculations in multiway pots: when will we have a solution?



Mon
09-12-2009, 12:41 PM
It's been a long time now since I use HM and I find it time to give a solution to the EV$ calculation problem on multiway pots.

Under my point of view, EV$ calcs are a valuable tool that we need to evaluate our real play, if we are not results oriented but play oriented. HM lacking a right calculation of EV$ is a major leak in my opinion.

I think that us, HM customer, have been patient enough with this issue, and we've been hoping it to be solved realease after release with no success. We've seen countless post in this forum stating it will be solved, but the solution you came upon (using the EV$ when the lasts to act players put money inv the pot) is far from correct.

It is not normal to have an EV$ calculation that tells you that, if you go allin preflop with AA against two players that have more stack than you and one of them catches a set in the flop, bet, and the third player folds, your equity in that hand was a 10%. Not acceptable when there are 3rd party tools out there capable of calculating EV street by street, as it should be.

Please, HM developers, address this once and for ever. If not the product, in my opinion, doesn't meet the goals it promised when I bought it, and I'll tranform from a patient to a dissapointed customer.

fozzy71
09-12-2009, 06:17 PM
I have forwarded this thread to the developer for a reply.

Rvg72
09-14-2009, 06:40 PM
It is not normal to have an EV$ calculation that tells you that, if you go allin preflop with AA against two players that have more stack than you and one of them catches a set in the flop, bet, and the third player folds, your equity in that hand was a 10%. Not acceptable when there are 3rd party tools out there capable of calculating EV street by street, as it should be.


This creates invalid results. If the first player who hit a set had 22 and the second player who folded to the flop shove had AA then should your EV really be 83% on the preflop shove?

Our EV is all-in EV and to remove bias it is calculated from the point where there are no more actions. Other tools do provide this info and we may possibly add it at some point but just because other tools calculate it that doesn't mean it is correct.

Roy

Mon
09-15-2009, 02:21 AM
This creates invalid results. If the first player who hit a set had 22 and the second player who folded to the flop shove had AA then should your EV really be 83% on the preflop shove?

Sorry, I don't understand your example. How could I fold to the flop shove if I already went allin preflop?

Please, correct me if Im wrong with the calcs in this example:

Preflop:

Hero has AA with 50BB
Villain1 has 22 with 100BB
Villain3 has 66 with 100BB

Hero goes allin preflop for 100

Equities: Hero 66% (0.66*150BB=99BB), Villain1 14% (0.14*150BB=21BB), Villain2 20% (0.20*150BB=30BB)

Flop: x2x

Villain1 goes allin, Villain2 calls

Equities: Villain1 87% (0.87*100BB=87BB), Villain2 13%(0.13*100BB=13BB)


Our EV is all-in EV and to remove bias it is calculated from the point where there are no more actions. Other tools do provide this info and we may possibly add it at some point but just because other tools calculate it that doesn't mean it is correct.

I understand that what you are calculating is Allin EV and not Sklansky Bucks, but given the missinformation and invalid results that Allin EV gives in multiway pots I strongly suggest that you implement Sklansky Bucks street by street as a better meassure of our real play.

Thank you.

stinkypete
09-26-2009, 03:23 AM
This creates invalid results. If the first player who hit a set had 22 and the second player who folded to the flop shove had AA then should your EV really be 83% on the preflop shove?

Our EV is all-in EV and to remove bias it is calculated from the point where there are no more actions. Other tools do provide this info and we may possibly add it at some point but just because other tools calculate it that doesn't mean it is correct.

Roy


just wanted to confirm here... when you say it is "calculated from the point where there are no more actions", you do in fact mean "no more actions" ie. everyone is all in?

ie. if i go all in and two other players (not all in) check it down, my EV would be whatever my EV is at showdown, correct?

fozzy71
09-26-2009, 09:42 AM
just wanted to confirm here... when you say it is "calculated from the point where there are no more actions", you do in fact mean "no more actions" ie. everyone is all in?
........

Correct

HM only calculates 'All-In Expected Value'. It can only do that if all players still in the hand, are all-in and their hole cards known.

stinkypete
09-26-2009, 12:23 PM
Correct

HM only calculates 'All-In Expected Value'. It can only do that if all players still in the hand, are all-in and their hole cards known.

ok good... thanks for doing it right! trying to get that pokerEV dude to do it right was such a pain.... lol

kingofthebeats
12-18-2009, 10:52 AM
This creates invalid results. If the first player who hit a set had 22 and the second player who folded to the flop shove had AA then should your EV really be 83% on the preflop shove?

Our EV is all-in EV and to remove bias it is calculated from the point where there are no more actions. Other tools do provide this info and we may possibly add it at some point but just because other tools calculate it that doesn't mean it is correct.

RoyRoy, I agree that calculating the EV preflop when one or more player is not all in would be inaccurate but calculating EV on the street where the last players are all in even when some of the players are all in on a previous street is not the ideal solution either. It would be better if there was an option not to include hands where players weren't all in on the same street in EV calculations since the equity on the flop and turn is not very relevant to the player who was all in preflop.

fozzy71
12-18-2009, 07:05 PM
We will be expanding our EV functionality next year.

strannikspb
12-19-2009, 11:42 AM
http://s40.radikal.ru/i089/0912/54/adda154abfe2t.jpg (http://radikal.ru/F/s40.radikal.ru/i089/0912/54/adda154abfe2.png.html)

Why there is no EV? I went all-in on the preflop, and one opponent showed his cards: AJ. The other opponent folded after the flop but before showdown.
So due to this, my equity of AK vs AJ on the distance is 0-7.75= - 7.75 dollars))

JPFisher55
12-19-2009, 11:48 AM
We will be expanding our EV functionality next year.

Will this include EV and win% stats for all showdown hands like those currently available for all in hands?

fozzy71
12-19-2009, 12:51 PM
Will this include EV and win% stats for all showdown hands like those currently available for all in hands?

I really don't know at this point. That is a question for the developers, but I doubt they will be very forthcoming at this point either. It will include some sort of EV-by-street calculations, is about as detailed as I can be at this point.

strannikspb
12-21-2009, 02:02 PM
Could you please help with my issue? It's the last post on the previous page.
I also noticed other hands with no equity. All these hands are hands when I'm preflop all-in vs more than one player.

netsrak
12-21-2009, 03:00 PM
To which post are you referring?

All-in is calculated when 2 or more players are all-in or nobody is able to make any more actions. So in your case all-in ev is calculated on later streets.

strannikspb
12-21-2009, 04:55 PM
I refer to this post - http://forums.holdemmanager.com/showpost.php?p=112898&postcount=10

Sorry, but I did't understand. If all hands in hand are known, we can calculate equity for each hand, can't we. If we are all-in preflop, it's not important for us what they will do on the flop or turn. But if we see showdown, we can calculate our equity vs their hands. But I noticed that in that hand one player folded on the river. So we can't calculate our EV. But in the other hand 3 players limped, I pushed, 2 called and than went to SD. But 1 limper folded preflop. In this situation (when we have one or more opponent folded preflop) we also couldn't calculate our EV?

I also couldn't understand why ATs have 0.0% EV agaist KK & Q6s? It's strange, because in Hand Replayer this hand has 29% EV on the preflop. But in Equity row I can see 0.0.

netsrak
12-22-2009, 06:01 AM
I'm not going to start a new all-in ev discussion here. It is as it is at the moment but there will be some additions to this in future releases.

strannikspb
12-22-2009, 11:57 AM
Ok. But I didn't want to discuss ev. I want to inform you about errors in programm. As with ATs. It's not about what EV this hand really has. But it's about the thing, that EV for one hand is displayed different in your programm. I think that it's not supposed to be like this.

fozzy71
12-23-2009, 11:30 AM
http://s40.radikal.ru/i089/0912/54/adda154abfe2t.jpg (http://radikal.ru/F/s40.radikal.ru/i089/0912/54/adda154abfe2.png.html)

Why there is no EV? I went all-in on the preflop, and one opponent showed his cards: AJ. The other opponent folded after the flop but before showdown.
So due to this, my equity of AK vs AJ on the distance is 0-7.75= - 7.75 dollars))

Post the complete/original hand history if you want to discuss possible errors.

strannikspb
12-24-2009, 01:23 PM
***** Betfair Poker Hand History for Game 621584408 *****
NL $0.25/$0.50 Texas Hold'em - Friday, December 18, 11:20:27 GMT 2009
Table Cygnus 06 9-max (Real Money)
Seat 5 is the button
Total number of active players : 9
Seat 1: bulletbroof ( $48.50 )
Seat 2: ian5464 ( $75.75 )
Seat 3: svejko ( $19.27 )
Seat 4: ruben29 ( $49.36 )
Seat 5: BaracusT ( $65.30 )
Seat 6: shrek77 ( $55.36 )
Seat 7: milliarder1 ( $7.75 )
Seat 8: Tirlimbom ( $139 )
Seat 9: teamdobbs ( $115.49 )
shrek77 posts small blind [$0.25]
milliarder1 posts big blind [$0.50]
** Dealing down cards **
Dealt to milliarder1 [ Ah, Kd ]
Tirlimbom folds
teamdobbs folds
bulletbroof folds
ian5464 raises to [$1]
svejko folds
ruben29 calls [$1]
BaracusT folds
shrek77 calls [$0.75]
milliarder1 goes all-in
milliarder1 raises to [$7.75]
ian5464 calls [$6.75]
ruben29 folds
shrek77 calls [$6.75]
** Dealing Flop ** [ Th, Jh, 7d ]
shrek77 checks
ian5464 checks
** Dealing Turn ** [ 8h ]
shrek77 checks
ian5464 checks
** Dealing River ** [ 7s ]
shrek77 checks
ian5464 checks
** Showdown **
shrek77 shows [ Ad, Jd ] two pair, Jacks and Sevens
milliarder1 mucks [ Ah, Kd ]
ian5464 mucks [ 4s, 4h ]
** Hand Conclusion **
shrek77 wins $23.05 from main pot with two pair, Jacks and Sevens
************ Game 621584408 ends ************

fozzy71
12-24-2009, 02:52 PM
You can't calculate the preflop equity because 2 other people were in the hand to the river, and didnt show their hands preflop.

That is explained in this thread as well as others.


http://www.holdemmanager.net/faq/afmviewfaq.aspx?faqid=68

http://forums.holdemmanager.com/showthread.php?t=20994

http://forums.holdemmanager.com/showthread.php?t=17621

http://forums.holdemmanager.com/showthread.php?t=20000

ImSpecial
02-15-2010, 12:39 AM
Was going to post this as a bug but it seems from this thread and others it's a design problem.

I get all-in pre-flop with Aces in PLO... clearly the more correct figure to use for my equity here is 30% (picture 1), rather than the actual figure used of 13.2% (picture 2)... although I agree that figure will be affected if there are folds post-flop.

In some ways it's not really important what their actual hands are I guess - really we want equity against the ranges of 4 people willing to get it in at the point where we got it in. Tricky problem to implement.

yngwiescruggs
02-23-2010, 03:15 AM
I refer to this post - Holdem Manager Forums - View Single Post - EV$ calculations in multiway pots: when will we have a solution? (http://forums.holdemmanager.com/showpost.php?p=112898&postcount=10)

Sorry, but I did't understand. If all hands in hand are known, we can calculate equity for each hand, can't we. If we are all-in preflop, it's not important for us what they will do on the flop or turn. But if we see showdown, we can calculate our equity vs their hands. But I noticed that in that hand one player folded on the river. So we can't calculate our EV. But in the other hand 3 players limped, I pushed, 2 called and than went to SD. But 1 limper folded preflop. In this situation (when we have one or more opponent folded preflop) we also couldn't calculate our EV?

I also couldn't understand why ATs have 0.0% EV agaist KK & Q6s? It's strange, because in Hand Replayer this hand has 29% EV on the preflop. But in Equity row I can see 0.0.

bump. This exactly is really bothering me. I play a lot of double or nothing games and these sort of situations happen all the time so it affects my EV a lot. It would really help me if these situations were calculated correctly. In fact, if the other players do not fold, then you can and should absolutely calculate EV correctly as of the time you shoved.

Even if another player does fold postflop, estimating some non-zero value seems better to me than saying zero. Perhaps there could be an option to turn this on and off depending on if the user thinks it's misleading? For example, AA vs. QQ and A8. I am AIPF but covered by both. Board comes down J34 and QQ bets, A8 folds. Just calculating the EV of AA vs QQ, while not perfect, is much much better than saying your EV is zero!

If a player goes all in vs. 2+ others who have him covered and then all calling players see showdown, all-in EV should definitely be calculated as non-zero.
If a player goes all in vs. 2+ others who have him covered and then a player folds between the AI and SD, then players should have the option to have their all-in EV calculated as either zero or as if they were only against the hand that made it to showdown.

Here's an example of a hand where it says I have 0 EV. This happens to be one of the more obviously wrong case where all hole cards from all players in the pot are known.

PokerStars Game #37619632521: Tournament #228418857, $10.00+$0.40 USD Hold'em No Limit - Level IX (200/400) - 2010/01/03 23:10:04 ET
Table '228418857 1' 10-max Seat #8 is the button
Seat 2: Smiley9999 (975 in chips)
Seat 3: Yngwie S (615 in chips)
Seat 5: sonyat01 (505 in chips)
Seat 7: Gax10 (6405 in chips)
Seat 8: Eugene'nin (1870 in chips)
Seat 10: gropekkk (4630 in chips)
Smiley9999: posts the ante 40
Yngwie S: posts the ante 40
sonyat01: posts the ante 40
Gax10: posts the ante 40
Eugene'nin: posts the ante 40
gropekkk: posts the ante 40
gropekkk: posts small blind 200
Smiley9999: posts big blind 400
*** HOLE CARDS ***
Dealt to Yngwie S [Js Jd]
Yngwie S: raises 175 to 575 and is all-in
sonyat01: folds
Gax10: folds
Eugene'nin: folds
gropekkk: calls 375
Smiley9999: calls 175
*** FLOP *** [Td 3s Ks]
gropekkk: checks
Smiley9999: checks
*** TURN *** [Td 3s Ks] [4c]
gropekkk: checks
Smiley9999: checks
*** RIVER *** [Td 3s Ks 4c] [Ad]
gropekkk: checks
Smiley9999: checks
*** SHOW DOWN ***
gropekkk: shows [Qh Ah] (a pair of Aces)
Smiley9999: mucks hand
Yngwie S: mucks hand
gropekkk collected 1965 from pot
*** SUMMARY ***
Total pot 1965 | Rake 0
Board [Td 3s Ks 4c Ad]
Seat 2: Smiley9999 (big blind) mucked [6h 2c]
Seat 3: Yngwie S mucked [Js Jd]
Seat 5: sonyat01 folded before Flop (didn't bet)
Seat 7: Gax10 folded before Flop (didn't bet)
Seat 8: Eugene'nin (button) folded before Flop (didn't bet)
Seat 10: gropekkk (small blind) showed [Qh Ah] and won (1965) with a pair of Aces

saygn
02-23-2010, 09:57 AM
hmm I did not know about this. I was under the assumption that all bugs for $EV were fixed. In fact, $EV was the ONLY reason I bought HEM, so this is a bit disturbing. If the developers are looking for a good template, I'd definitely suggest the Juk Luck analyzer on the 2+2 software forum. As far as I know, he had worked out all the bugs.

morny
02-23-2010, 01:56 PM
Mike recently announced he was looking into a better way to calculate this so it is something were working on but i dont have an ETA at the moment

yngwiescruggs
02-23-2010, 02:27 PM
Here's an example of a hand where it says I have 0 EV. This happens to be one of the more obviously wrong case where all hole cards from all players in the pot are known.

PokerStars Game #37619632521: Tournament #228418857, $10.00+$0.40 USD Hold'em No Limit - Level IX (200/400) - 2010/01/03 23:10:04 ET
Table '228418857 1' 10-max Seat #8 is the button
Seat 2: Smiley9999 (975 in chips)
Seat 3: Yngwie S (615 in chips)
Seat 5: sonyat01 (505 in chips)
Seat 7: Gax10 (6405 in chips)
Seat 8: Eugene'nin (1870 in chips)
Seat 10: gropekkk (4630 in chips)
Smiley9999: posts the ante 40
Yngwie S: posts the ante 40
sonyat01: posts the ante 40
Gax10: posts the ante 40
Eugene'nin: posts the ante 40
gropekkk: posts the ante 40
gropekkk: posts small blind 200
Smiley9999: posts big blind 400
*** HOLE CARDS ***
Dealt to Yngwie S [Js Jd]
Yngwie S: raises 175 to 575 and is all-in
sonyat01: folds
Gax10: folds
Eugene'nin: folds
gropekkk: calls 375
Smiley9999: calls 175
*** FLOP *** [Td 3s Ks]
gropekkk: checks
Smiley9999: checks
*** TURN *** [Td 3s Ks] [4c]
gropekkk: checks
Smiley9999: checks
*** RIVER *** [Td 3s Ks 4c] [Ad]
gropekkk: checks
Smiley9999: checks
*** SHOW DOWN ***
gropekkk: shows [Qh Ah] (a pair of Aces)
Smiley9999: mucks hand
Yngwie S: mucks hand
gropekkk collected 1965 from pot
*** SUMMARY ***
Total pot 1965 | Rake 0
Board [Td 3s Ks 4c Ad]
Seat 2: Smiley9999 (big blind) mucked [6h 2c]
Seat 3: Yngwie S mucked [Js Jd]
Seat 5: sonyat01 folded before Flop (didn't bet)
Seat 7: Gax10 folded before Flop (didn't bet)
Seat 8: Eugene'nin (button) folded before Flop (didn't bet)
Seat 10: gropekkk (small blind) showed [Qh Ah] and won (1965) with a pair of Aces

I'm not sure this adds much but I just wanted to clarify on the implications: If I HAD been lucky enough to win this hand, HEM's EV calculation would say I was a 100% lock to win the hand, EV wise. So while this calculation somewhat evens itself out over the long run, it ruins the value of the calculation to the player over a relatively small sample size. Basically, when you run good in these situations, you are misled into thinking you are really beating the game by more than you are, and when you run bad in these situations, your EV is calculated much lower. (Here, my EV calculation is off by about 1 full buy-in from what it ought to be regardless of whether I win or lose)

Over the short term this is a pretty big deal and also depending on your style of play (in the DoN games specifically) it may be a big deal over the long term as well, i.e. some people tend to get in more 3 way pots than others, some get in more as the shortstack, some people get in more as the big stacks.

Thanks HEM for looking into this.