That is intentional - "hero" stats are based on the current session to show your current table image. This cannot be changed – if you want to see your full stats you can see them in the main tracker.


You can't build stats to count actions by a specific other player – so you can't count 4-bets when "hero" open-raised.
If you mean that you want to know the percentage of times that a player 4-bets having made the initial raise then stevi3p's stat looks like what you want.
Both of his stats look fine.

i did'n meant other player action. 'hero' and 'player for we are measuring percentage' are the same..

so, stevi3p's stat is 100% accurate and i can use it, right?

his second stat 'Fold to 4Bet after 3betting' is identical with your built in '3Bet Preflop And Fold To 4Bet' ?

These stats are all pretty straightforward and look fine. Those two stats are effectively the same, yes.

Can you explain (or link me to an explanation) of how it's calculated when people are AI on different streets or I go AI on the flop multiway and then one of the other players bets out the others and we're HU on the turn?

If the cards of any players involved at the time of the all-in [being called] are not known because they later fold then EV is not calculated for the hand.

Took a quick look around how HEM calculates it and it seems they do it based on the equities when the last person is AI or there can't be any additional action. Seems it has it's flaws too, but they're of the evening themselves out over time variety afaict. I'd think it would be more accurate than ignoring it. Maybe help me understand why PT3 thinks otherwise?

Edit: Also, option for bb and not BB as base unit, please.

Is there any way to pay for pokertracker through site money?

Consider the following scenario:

UTG limps for less than 1 big blind and is all in. All other players limp behind. The flop comes 777. The SB checks, BB bets out, and all others fold. BB shows 72o. UTG flips up AK.

What would you want to see here for UTG's equity? There seem to be trouble with most of the major options:

(1) If it's 0% because he was drawing dead on the flop, that seems to be ignoring the fact that he got it in way ahead - which is defeating the purpose of all-in EV.

(2) We don't want to show 67.565% (AKo vs 72o) for UTG because there were a bunch of others in the pot preflop too, so UTG wasn't really 67.565% to win.

(3) We could start adding in random hands for everyone else, except their range isn't truly random, so it's not really going to be that accurate here either.

The end result is that none of the options seem that good, so if there are hands we cannot account for precisely due to later folding, we don't calculate EV.

And that option is planned, but I'll remind the development team about it for you.

Not at this point in time, I'm afraid. The poker sites do not allow it.

But doesn't 1) 0% mirror what you show by not calculating AIEV for the hand in these situations? The AIEV line goes down just like the green line. But in the case where we do still have equity on the flop and it's a less extreme example, (say we're 25%), wouldn't the AIEV line showing us with 25% equity be more accurate than it just recording the result of the hand? It's not as good as being able to do it accurately based on PF and perfect information, but I've gotta think it's better than the current way.

I mean, say we get it in all of these situations with exactly 50% equity, our equity is going to change when others get it all in/there's no more action, but it's certainly not going to be the 0% or 100% that's shown based on the results currently (i.e. whether we win or lose the hand). I'd rather have the range of say 25% to 75% be reflected than the all or nothing approach PT3 is taking currently by ignoring it.

Is it possible to pay for PT3 microstakes version with a gift card? I have an AMEX gift card that I would like to use but wasn't sure if gift cards are accepted. Thanks.

I understand your point, but I still respectfully disagree. Another way to use the all-in EV line is to see if you are getting your money in good - if you're making good push/fold decisions the line should continue to trend upward.

However, if we calculate things your way, we'll end up looking like we generally get it in worse than we do. It's kind of a variant on the Monty Hall problem - if we're up against three unknown hands preflop, and one hand bets the other two unknowns out on the flop leaving us heads up and all-in with them, we're going to see them turn up a hand that flopped very well very often. So our equity on the flop is going to generally be pretty low in these cases and that's going to drag our gold line down - making it look like we generally get it in badly.

If these things happen a lot, then your gold line will be significantly lower than it would be with perfect information which might cause a player to change what might be an otherwise good strategy.

If these things don't happen a lot, then the discussion is really fairly moot - if it's that rare and your EV line would differ only very slightly from the green line for that hand (25% equity in a pot < 1BB is not exactly a big shift after all).

We do take American Express - so that should work just fine.

Meh... recording the AIEV as either/or based on the result of the hand skews it more than the above. How can it possibly look like we get it in worse than it does mirroring the cash line if we lose the hand? Or if we suck out, we skew the line the other way since it's going to look like we got it in awesome.

Definitely puts a big dent in PT3 IMO. I've used you since I started playing and I hate to potentially be this whatever about something as relatively trivial as AIEV, but I'm going to have a look at HEM for my March hands and see how it compares. If I'm suddenly 12 BI under EV instead of 5, well that's going to have a fairly big impact on my psyche (even if it shouldn't) and might just be worth making the switch. I'll keep you posted.

March with PT3:



March with HEM:



~4 BI difference in AIEV seems somewhat substantial over 5k hands, IMO.

The EV-Adjusted statistic makes many assumptions, at least one of which is, in my humble opinion, fatal to the statistic's accuracy. In cases where Hero is all-in on the flop and called by two opponents, and on the turn one opponent bets and the other folds, HEM's EV-Adjusted statistic calculated EV for only the known cards on the flop. This leads to a statistical skew in which Hero will always be below EV in the long run.

This skew is most easily shown in an example. Consider the following beginning of a hand:

Any Site - $.50 NL - Holdem - 3 players
Hand converted by PokerTracker 3

Hero (BTN): $33.33
SB: $50.00
BB: $50.00

SB posts SB $0.25, BB posts BB $0.50

Pre Flop: ($0.75) Hero has A K

Hero raises to $1.50, SB raises to $4.50, BB calls $4.50, Hero raises to $14, SB calls, BB calls

Flop: ($42.00, 2 players) 7 5 2
SB checks, BB checks, Hero bets $19.33 and is all-in, SB calls $19.33, BB calls $19.33



Hero has AKo, Villain 1 has QJo and Villain 2 has T9o. The flop is 752, rainbow, and the pot is size $100 (let's ignore rake and the penny for math's sake). We know, therefore, via mathematics that Hero has 58% equity in the pot, Villain 1 has 21% equity in the pot, and Villain 2 has 21% equity in the pot (again, please excuse the small amount of rounding, it will not make any difference). If the pot were instead heads up against Villain 1 alone on the flop, Hero would have 78.5% equity. If the pot were instead heads up against Villain 2 alone on the flop, Hero would have 75.25% equity. However, Hero does in fact have 58% equity in this pot against both players, and Hero's true EV-Adjusted amount won should be $58.

Let us assume that the villains only bet if they improve to a pair, and do not bet otherwise. Due to this particular board structure, the villains need to pair to improve and can only improve by pairing. On the turn the board will pair either villain about 13% of the time (there are 6 pairing cards each and 46 unknown cards), and (due to rounding), if they do not pair the board, on the river they each again have a 13% chance of pairing the board. If neither villain pairs the turn or the river, the hand will be checked down. We can thus state if the hand is checked down, Hero's EV-Adjusted will be accurate. In all other cases, there will be betting, which as we will see, will skew the results considerably.

On the turn 13% of the time, Villain 1 bets and Villain 2 folds. When this happens, HEM then reports Hero's equity as 78.5% on the flop, so his EV-Adjusted shows $78.5! On the turn, Hero is far behind – he must catch one of his 6 outs to catch up, which happens only 13% of the time. If he does catch up, he will be shown as $22 above EV (when he is in fact $42 above EV) and if he does not he will be shown as $78.50 below EV. Thus he will show, in this case, in the long run, to be below EV by $65. In other words, when Villain 1 improves and bets and Villain 2 folds, the EV adjusted statistic's inability to account for Villain 2's cards on the flop causes Hero in the long run to appear almost 2/3 of the pot below expected value.

The mathematics for the remaining cases are left as an exercise to the reader, and the below table will show the results:



We can thus see clearly that the EV-Adjusted winrate is very far from the actual results – in 45.24% of cases, Hero's EV is vastly overrestimated. Further, given that we know the likelihood of each case occurring, we can calculate what the long term reported EV-Adjusted Win will be: $66.54. In the long run, due to HEM's calculation methods, Hero will appear to be $8.54 – or almost 15% of the pot - below EV.

While this case is a simplification, situations like these are quite common. The more multiway pots that a player enters while all-in under similar conditions, the more below EV he will be reported in the long run. I admit that these situations are complex – there is a feeling that in cases like these, there should be some answer as to "what is my EV" when one of the villains folds. The fact of the matter is, that you cannot remove their cards entirely – otherwise Hero is always below EV. However, since you don't know what cards to put in, you cannot be accurate about it. If instead of AK vs QJ and T9 it was AK vs KQ and QJ and the J fell and QJ bets and KQ folds we have the same end result as in one of our scenarios above, but the numbers are quite different.

This turned out to be much longer than originally anticipated – but the end upshot (or the tl;dr version) is: ignoring hands that are later folded after the all-in causes Hero to end up being shown as under EV in the long run – and I think that's a bad thing. Therefore, PT3 will not make these assumptions, and not show EV for hands in these situations.

2nd point first: I think this is wrong wrt how/when HEM calculates it. If Villain 1 bets and Villain 2 folds on the turn, HEM AIEV is calculated as if the players went AI on the turn and there were only Hero and Villain 1. (If any HEM people are lurking this thread, perhaps they can confirm?) So if Villain pairs up and bets, Hero's going to show as worse than he should, but will still show as having some EV > 0. In PT3, AIEV isn't calculated and the AIEV line moves in step with the cash line. So if true EV is $58 how is showing EV as $0 (we lose) or $100 (we win) more accurate? Seems like a number that's going to be between $0 and $58 is going to be more accurate.

I mean, say we always have an EV of $50 in these situations and because of turn action, it changes to a range from $25-$75 (generally with 0 and 100 as possible, but outliers), then isn't that narrower range better than the all or nothing approach PT3 takes?

I get that it's not perfect or necessarily accurate, but isn't it more accurate than ignoring it?

All my players have seemed to disappeared from csh games only and my graphs have reset themselves. However I still have all my hands appearing in the "general" tab.
Have you guys got any ideas of whats happened or how I can change this back. Thanks

I have replied to your thread on the PT3 forum – let's keep everything in one place.

does this work fully for Rush NL and PLO? I currently have HEM and it is intermittant.

It does for Windows machines. Rush support is not currently available on the Mac version.

Is there a way to minimize HUD stats or make someone's HUD stats invisible??

You can't make individual players' stats invisible, but you could create a different Profile with many less stats and switch to that at the table to effectively "minimise" the Hud at a certain table.
Tutorial: Managing HUD Profiles

I am using a mac running OSX 10.5....
Hopefully I found the right place to post this... Having problems with my default hud settings.... this is what shows up... Uploaded with ImageShack.us

I tried to import my settings from my other profile (which I would like to make my default) to the default profile but it did not fix the problem. Also, sometimes the stats cover the "fold button" or get in the way. How do I move the stats so they dont interfere with gameplay ?

That looks as if you have accidentally changed one of the pop-up groups to be "show on table for everyone".

See the Tutorial: Managing HUD Groups for how to correct this manually, or the Tutorial: Managing HUD Profiles for how to use your new default profile in the Hud.

I hope I'm posting this in the right place.

I have Kraada's HUD and am very used to using it now, but I feel like I'm missing out because there isn't AFq or AF on the actual hud, and playing 8-12 tables makes it harder to bring up the popups all the time.

So I know how to technically add this to the HUD but my question is what ranges of % would I use for AF if I decided to make a 5th stat on the second line AFQ% colored by AF%?

Would this be useful?
Is this redundant?

Personally I don't like AF or AFq, as I don't think they really tell you very much you can't get more easily from looking at the postflop stats in the popup I've included in my HUD. As such, I couldn't really tell you the best way to configure those color ranges.