holdemluck vs pt3 allin ev graph

[bagge]

I know the pt3 graph didnt use to be correct, but wasnt this fixed?

They show way different results.

[kraada]

The PT3 graph only has the EV change for all-in hands, so if you've got the normal amount won line and the EV line you're also seeing the line go up (or down) by hands which never got to showdown. The Holdem Luck graph doesn't show those hands at all I believe. So that might explain some of the discrepancy.

If you can find any particular hands that show different EV results in their software and in ours though please let me know so that we can investigate further.

[bagge]

On the graph's I'm currently comparing my non SD winnings line is on a rather consistent decline, I realize this would include hands that werent all in as well that wouldnt count towards all in EV, but since there really arent any peaks I thought it'd be a somewhat correct assumption that my all in non SD winnings was at best about break even but may very well be on the losing side. That I thought would result in PT3 EV line being adjusted negatively (or not at all, if it was break even) compared to the winnings line, meaning the EV line should be lower in PT3 rather than holdemluck.

The PT3 graph shows the EV line pretty much equal to the winnings line. Holdemluck shows me running 3636 BB below EV. This is on a sample of about 195k hands, where according to holdemluck about 2k were all in(went to SD). Is it safe to assume that when me and opponents have gotten our $ (went to SD) in I've been getting a bit unlucky and losing more than I should, but when me or opponents have moved all in without a call (no SD) I'm been profitable, and in the PT3 graph that weights up for the negative SD EV results of holdemluck? If so, is there a report I can run to verify this?

[kraada]

Run a custom report in the Holdem Cash Hand section that includes Amount Won and EV Allin Expected that includes the Hand # and perhaps your hole cards and the board. If you then add a filter for times you raised all in (in the blue Filters link click Simple Filters, Actions -> Preflop -> Raise All In), hopefully we can narrow down some hands that PT3 and HL aren't agreeing here. You can further filter by dates and you can filter by dates in HL which might help. We'd need to see some actual discrepancy before we can figure out where the problem is.

[bagge]

I'm not sure there is a problem, just trying to figure out why the difference and interpret how it affects the end result of my game. Does it sound to you like there is a problem from what I've said?

I did add all in preflop or flop or turn or river filters, but would that include call all in hands or just the ones where I shove first? PT3 came up with lesser hands than holdemluck 1995 vs 2004, so I guess I can assume that call all ins are included in this as the difference is so small. But PT3 also came up with higher won ammount value while having slightly fewer hands, roughly 40% more, could this be due to the new multi currency addition? It would seem like a reasonable ammount considering my currency is USD atm whereas I've been playing almost exclusively on EURO tables. But the EV result now surprisingly show I've been running slightly above EV. So, 9 fewer hands, difference in won ammount and the EV stat moving in the wrong direction vs holdemluck is whats having me confused now.

I do also have most of this sample in HEM (missing about 10% of the hands), but its EV stat results are very close to holdemluck's.I should mention that I think a part of this sample is from before the EV stat in PT3 was fixed, could some of those faulty calculations not still be in cache or something and needs to be updated somehow?

[kraada]

The EV stats in PT3 are currently generated as the reports run, so the older hands shouldn't make that much of a difference.

Hands where you call an all-in should also be counted, I was just trying to give you a way to narrow down the discrepancy some so that we could investigate further. These stats are new and it is entirely possible there are some things which need fixing.

[bagge]

ok thank you for your help. Going to trust holdemluck for now until I'm certain as its always nice to be able to expect some more winnings due to EV stats. :D

I might check and compare stats over different periods of time and let you know if I find something out about the difference, comparing hand to hand seems like a lot of work given the size of my sample.

[zubs1aa]

The PT3 graph only has the EV change for all-in hands, so if you've got the normal amount won line and the EV line you're also seeing the line go up (or down) by hands which never got to showdown. The Holdem Luck graph doesn't show those hands at all I believe. So that might explain some of the discrepancy.

If you can find any particular hands that show different EV results in their software and in ours though please let me know so that we can investigate further.

OK I'm baffled on this one- was hoping this thread had an answer.

I looked at one session only. first 37 hands the line moves up and down as one line- which makes sense since I didn't get allin.

However, here is my first allin hand: KK v 89dd on 595 board no diamonds.
I'm 90% favorite when I called allin for my last $5.17 to win 11.31 total. 10.56 after rake.

What is wacky is that the line diverges in the graph -
-at the hand before I was at ($.87) on the session.
-Graph has net expected $3.01, total $4.10. ---that's $1.09 different.
-Also, my $ won at showdown says $4.99 - it had been $0 for the session up until this hand. WAT? 5.57 to start the hand and I won 10.56 total after rake, for a net win of 4.99, not 4.90.

tried 6 ways till sunday, but I can't figure out how any of those numbers compute.

- since I win the hand shouldn't the lines be the same? (if the answer is no, then i think the graph oughta be fixed so that AIEV only moves away from zero/up or down as a result of all in hands...it's still confusing the $%^& ought of people as implemented.)

- ignoring the above, how is there a $1.09 difference? even if you are somehow adjusting this for the value at the hand before, that was (.87) - where's the other .22 coming from?

Be interested in hearing the explanation on this. I'm still in the camp for scrapping these graphs until they are crystal clear as to what they are saying. Better no feature then one with bad data or misleading data or clear data that has no applicability to poker in teh way most players would look at the game. (said it 1000x I know...)

HH text below.

Full Tilt Poker Game #20544150411: Table Flash - $0.02/$0.05 - No Limit Hold'em - 5:58:04 ET - 2010/05/03
Seat 1: yozey_walez ($5.22)
Seat 2: zubs1aa ($5.57)
Seat 3: Ekaterinochka ($1.60)
Seat 4: patty688 ($7.77)
Seat 5: Zreticoli ($5.12)
Seat 6: EmKayFour ($4.27)
Seat 7: idofeo ($0.41)
Seat 8: summ3r ($3.18)
Seat 9: aVar1zZzle ($2.09)
idofeo posts the small blind of $0.02
summ3r posts the big blind of $0.05
The button is in seat #6
*** HOLE CARDS ***
Dealt to zubs1aa [Kd Ks]
aVar1zZzle folds
yozey_walez folds
zubs1aa raises to $0.15
Ekaterinochka folds
patty688 calls $0.15
Zreticoli folds
EmKayFour folds
idofeo folds
summ3r calls $0.10
*** FLOP *** [5s 9h 5h]
summ3r checks
zubs1aa bets $0.25
patty688 raises to $7.62, and is all in
summ3r folds
zubs1aa calls $5.17, and is all in
patty688 shows [6d 9d]
zubs1aa shows [Kd Ks]
Uncalled bet of $2.20 returned to patty688
*** TURN *** [5s 9h 5h] [3d]
*** RIVER *** [5s 9h 5h 3d] [8s]
patty688 shows two pair, Nines and Fives
zubs1aa shows two pair, Kings and Fives
zubs1aa wins the pot ($10.56) with two pair, Kings and Fives
*** SUMMARY ***
Total pot $11.31 | Rake $0.75
Board: [5s 9h 5h 3d 8s]
Seat 1: yozey_walez didn't bet (folded)
Seat 2: zubs1aa showed [Kd Ks] and won ($10.56) with two pair, Kings and Fives
Seat 3: Ekaterinochka didn't bet (folded)
Seat 4: patty688 showed [6d 9d] and lost with two pair, Nines and Fives
Seat 5: Zreticoli didn't bet (folded)
Seat 6: EmKayFour (button) didn't bet (folded)
Seat 7: idofeo (small blind) folded before the Flop
Seat 8: summ3r (big blind) folded on the Flop
Seat 9: aVar1zZzle didn't bet (folded)

[kraada]

zubs,

Could you attach that HH to a ticket so I can get it to the developers for further investigation? It's much easier for us if we have these things in a ticket.

[WhiteRider]

- since I win the hand shouldn't the lines be the same? (if the answer is no, then i think the graph oughta be fixed so that AIEV only moves away from zero/up or down as a result of all in hands...it's still confusing the $%^& ought of people as implemented.)

No.
When you are all in the EV graph will NEVER match the amount you won (apart from the river when you're either 100% or 0% to win).
If you're 90% to win a $10 pot (at the time of the all-in) then the EV graph shows you winning $9, but you can never actually win $9 - you either win $10 or $0, so the graphs diverge.