Two Plus Two Poker Forums Why is EV so difficult to explain?
 Register FAQ Search Today's Posts Mark Forums Read Video Directory TwoPlusTwo.com

 Notices 2013 Two Plus Two Party at South Point July 6th Party starting at 5PM Food served at 6PM

 Probability Discussions of probability theory

06-20-2012, 09:27 AM   #31
veteran

Join Date: Mar 2009
Location: In the wires
Posts: 2,334
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by Sherman The other calculation (the one PT currently uses) is useful if you want to see if your all-in decisions are +EV or -EV (i.e., if you are playing well).
That sums up my point nicely. You can't include hands like the example above in such a calculation.

06-21-2012, 04:12 AM   #32
journeyman

Join Date: Aug 2007
Location: Sydney, Australia
Posts: 306
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by NewOldGuy The bottom line is that if there is any remaining option for any player in the hand to make a subsequent decision before showdown, then you can't calculate "equity" for any player because obviously decisions are not random. EV on the other hand, can be calculated using whatever assumptions you want to make about the distribution of the open decisions. But I don't think that is the question here at all.
Unless I'm missing something I believe that the current calculations do take into account actions after the players all-in decision.

6 max all even stacks. Player utg shoves.
a)What is utg players all in EV if he's called by button and all else fold.
b)What is utg players all in EV if he's called by button and big blind.

Surely there are instances where a) and b) are different all-in EV result for UTG player. Therefore calculated all in EV in PT is affected by subsequent decisions before showdown.

Unless PT only calculates all-in EV where player is closing all action (i.e last to call) which I didn't believe it did. Am I wrong?

BTW I still think that calculated all-in EV where action closes on same street (if that is how it happens) is a more useful calc than what OP suggests. But only calculating all-in EV when actually closing all action in the hand is the only true measure of the actual value of a decision (and even that is debatable as it discounts card replacement in any non-heads up hand).

For mine it also highlights why we need be careful with simply assuming all-in EV is de-facto a measure of how well we "actually" played.

06-21-2012, 08:59 AM   #33
Carpal \'Tunnel

Join Date: Jun 2005
Location: Psychology Department
Posts: 7,450
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by mickb70 Unless I'm missing something I believe that the current calculations do take into account actions after the players all-in decision. 6 max all even stacks. Player utg shoves. a)What is utg players all in EV if he's called by button and all else fold. b)What is utg players all in EV if he's called by button and big blind. Surely there are instances where a) and b) are different all-in EV result for UTG player. Therefore calculated all in EV in PT is affected by subsequent decisions before showdown. Unless PT only calculates all-in EV where player is closing all action (i.e last to call) which I didn't believe it did. Am I wrong? BTW I still think that calculated all-in EV where action closes on same street (if that is how it happens) is a more useful calc than what OP suggests. But only calculating all-in EV when actually closing all action in the hand is the only true measure of the actual value of a decision (and even that is debatable as it discounts card replacement in any non-heads up hand). For mine it also highlights why we need be careful with simply assuming all-in EV is de-facto a measure of how well we "actually" played.
You are partially right. PT will still calculate your EV if you shove and another player calls. However, it does not count (or even include) your EV on hands where more than one person calls. That is what this whole thread is about.

You are certainly right that all in EV is not a great measure of how a person is playing. All-in EV is just easy to calculate, so unfortunately many players use it to determine if they are playing well or poorly. It reminds me of RBIs in baseball...they are easy to calculate so we count them up. But as a performance statistic it is one of the worst.

 06-21-2012, 09:04 AM #34 veteran   Join Date: Mar 2009 Location: In the wires Posts: 2,334 Re: Why is EV so difficult to explain? The +/-EV statistic isn't even intended for interpreting how "well" you are playing, it is specifically a measure of how "lucky" you are playing. The total value of the all-in wins over time could be related to quality of play, but not the EV delta. But that's beside the point of this discussion.
06-21-2012, 10:04 AM   #35
Carpal \'Tunnel

Join Date: Jun 2005
Location: Psychology Department
Posts: 7,450
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by NewOldGuy The +/-EV statistic isn't even intended for interpreting how "well" you are playing, it is specifically a measure of how "lucky" you are playing. The total value of the all-in wins over time could be related to quality of play, but not the EV delta. But that's beside the point of this discussion.
Actually, if you think that is its intention, then you should agree with Karganeth's method for calculating EV. That is a better measure of "lucky vs. unlucky" than the one PT currently uses.

06-21-2012, 10:15 AM   #36
veteran

Join Date: Mar 2009
Location: In the wires
Posts: 2,334
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by Sherman Actually, if you think that is its intention, then you should agree with Karganeth's method for calculating EV. That is a better measure of "lucky vs. unlucky" than the one PT currently uses.
Nope, that method is still wrong because we are retroactively applying an equity to a decision point that occurred before we knew the equity. Once you decide to go all-in, if no decisions are left then the delta between actual average outcome and predicted average outcome (equity) is entirely a matter of luck.

06-21-2012, 10:30 AM   #37
Pooh-Bah

Join Date: Jun 2009
Posts: 5,862
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by heehaww Because then it's not an EV, it's just the real dollar amount won/lost! At the time of showdown, you either have a 100% chance of winning and win 100% of the pot, or you won 0% of the pot...
The quoted text is correct (although the poster seems to think there is something wrong with it).

To clear up the confusion, remember that we are talking about the expected value of a variable. EV(X) where X is a specific situation. Always. If someone just says "expected value" then it is still referring to a specific situation, just that they figured it was implied what situation was being talked about.

Here are some examples where I haven't left out implied details:

EV(wager on a coin flip, before the coin is flipped) = 0.5 * wager_size
EV(wager on a coin flip, after the coin is flipped and shows Tails) = 1 * wager_size
EV(2+2) = 4
EV(earnings on a hand, 22 v AK, before community cards are dealt) = 0.52 * potsize
EV(earnings on a hand, 22 v AK, flop J55) = 0.75 * potsize
EV(earnings on a hand, 22 v AK, board J5599) = 0

In other words, adding new information changes how much you expect to earn (not rocket science)

 06-21-2012, 10:38 AM #38 Pooh-Bah   Join Date: Jun 2009 Posts: 5,862 Re: Why is EV so difficult to explain? Sorry, after that post I realised my post was poor and deleted the text you quoted. Better post follows soon
06-21-2012, 10:40 AM   #39
Carpal \'Tunnel

Join Date: Jun 2005
Location: Psychology Department
Posts: 7,450
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by Kittens Sorry, after that post I realised my post was poor and deleted the text you quoted. Better post follows soon
I deleted mine then. :-)

06-21-2012, 10:57 AM   #40
journeyman

Join Date: Aug 2007
Location: Sydney, Australia
Posts: 306
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by NewOldGuy Nope, that method is still wrong because we are retroactively applying an equity to a decision point that occurred before we knew the equity.
If this is true then PT should only calculate EV when you call all in to close action.

Modify my example.

UTG has TT decides to go all in. That is the decision point. Button has JJ, Big Blind has 99.

Action subsequent to decision point:

a) button calls, BB folds
b) button folds, BB calls.

Did the EV of the decision change? Based on what I'm hearing PTs calculation will. Or did the EV of the resultant situation based on future events change? I'm started to feel OP has some point. Seems arbitrary that we allow further action on same street but not subsequent streets.

06-21-2012, 11:05 AM   #41
Pooh-Bah

Join Date: Jun 2009
Posts: 5,862
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by Sherman This entire discussion revolves around two different ways of getting a number (call it EV or whatever you want). The numbers are not the same thing. Therefore they cannot be interpreted the same way. What does each one mean?
OK. You say that people want to know if they are playing well or not.
However the EV line cannot show that exactly. Here's why.

Let's say your opponent shoves preflop and you 100% accurately put him on a range of TT+, AK+. You call with KK, and he shows AA. Whoops?. Your EV gets recorded as 20% * potsize , when in fact you played well.

I hope it is clear to all that it is far too massive an undertaking for a poker tracker program to try and track what the ranges would of all the opponents at the time in question.

What the EV line does in fact record is how much you would have won if you had been neither lucky nor unlucky w.r.t. cards dealt after you made your final decision on the hand.

EDIT: it also does not track luck that happened before your final decision, for example, the case where you bet \$80 of your \$100 stack preflop with AA, then the opponent calls and flops a set. Undoubtedly you got very unlucky here but it will show up on your EV graph as you playing poorly (if you look at your EV line as showing how well you play -- which I am suggesting you should not treat it that way). I don't see any way the EV line can be reliably made to track the luck factor in this example.

Of course there is the issue that any particular one dealing of cards is either lucky or unlucky (unless your equity was exactly 50%!), however "expectation" means the average over an extremely long run of repeating the same trial -- or, to put it another way as OP did, how much money your seat would be worth if you were to sell it to someone else at that point in time.

Quote:
In this scenario you were all in on the flop. Therefore the EV should be recorded as your EV as it was on the flop , i.e. AA v JT v 98 on the KJ2 board. The fact that there was subsequent action is irrelevant as you were already all in, and the main pot is locked. We have made the conscious decision to not try to factor in things like "villain's range", "fold equity", "equity due to one player in the side-pot being forced to fold away his equity in the main pot", and so on, as discussed earlier.

OP seems to be saying that the calculation should be your EV as it was on the turn after another card had been dealt after you had already gone all-in. This figure would not help you judge whether you had been lucky or unlucky because it also includes the outcome of the turn card being dealt.

To make the example more extreme, let's say you went all in with AA on an ATT board, and villains had T9 and A4. The turn is the case T. The villain with T9 bets and the villain with A4 folds.

OP is claiming that our EV should be recorded here as \$0 (since we're drawing dead), however I would content that it should be recorded as about 96% * the size of the main pot -- where 96% is the equity of AA v T9 v A4 on the ATT flop.

Finally, I think we can turn to the real original problem. The above scenario quoted by OP is unrealistic as people don't fold face up online. I guess this was just a hypothetical scenario to clarify things before we moved onto the real scenario, so I'll do that now

My reading of OP's original post is that the real problem is when the AA v JT v 98 hand happens, but the 98 player folds face down.

We now have no way of calculating our EV as it was on the flop. Hence no calculation should be made.

As in my ATT example, I would not consider it correct if my EV were recorded in the database as \$0, i.e. I did not run bad when the T came on the turn. I'd want to see a gaping void between the EV line and the earnings like when the case T came.

That's my 2 cents

Last edited by Kittens; 06-21-2012 at 11:13 AM.

 06-21-2012, 11:28 AM #42 Pooh-Bah   Join Date: Jun 2009 Posts: 5,862 Re: Why is EV so difficult to explain? Cliffs of my post:People want the EV line to show their play "without the luck factor" This is impossible As a compromise, the bolded definition in my previous post is used: the luck of the remaining deck cards OP seems to think that if we are all in on the flop, but EV cannot be calculated due to incomplete information, then we should use EV on the turn instead I think this is poor because the turn card is a luck factor that we're trying to remove, and I give an extreme example to demonstrate this
 06-21-2012, 01:46 PM #43 adept     Join Date: Jun 2008 Posts: 1,187 Re: Why is EV so difficult to explain? Let's take a step back. From the wikipedia article: In probability theory, the expected value (or expectation, or mathematical expectation, or mean, or the first moment) of a random variable is the weighted average of all possible values that this random variable can take on. In cases where we do not know the probabilities of events which affect either our potential winnings or the probability of winning a given amount, we cannot calculate our EV. For example, I made a bet yesterday for \$10. You cannot calculate my EV because you do you have complete information. There are two missing pieces of information you need. You need the potential winnings (i.e. the odds which were layed) I got and the probability I will win. If we know only the odds that were laid, you still cannot know what my EV is. This is because you do not know the probability of winning. If you had the probability of winning but not the odds I was given, you cannot know what my EV is because you do not know what my potential winnings are. In the example I gave with AA vs 98 vs JT, we calculate our EV as soon as we know both the probability of winning and our potential winnings. Lets go step by step in the hand to see when we can calculate EV. Let's see if we can calculate our EV as soon as we are dealt our cards AA. We do not know our potential winnings or the probability of winning therefore we cannot calculate our EV. The flop is dealt and we go all in with AA and are called by 98 and JT who have more money behind. We cannot know our EV at this point in time in this situation because although we know our potential winnings, \$300, we do not know the probability of winning. On the turn, the player with JT bets and 98 folds face up. Now we know both our potential winnings \$300 and our probability of winning 0.88095. We can now therefore calculate EV. I think that a lot of people here think that EV is specific to poker. The correct EV calculation doesn't care about the rules of poker or when a player went all in. The only thing it cares about is how much value you expect and the algorithm I proposed does exactly that.
 06-21-2012, 01:47 PM #44 adept     Join Date: Jun 2008 Posts: 1,187 Re: Why is EV so difficult to explain? Heads up pot SB \$100 55 BB \$100 KK SB pushes all in and is called. Flop: T 8 5 Turn: K River: 5 In this situation as soon as we are dealt our cards, our EV is unknown because we do not know either our potential winnings or our probability of winning, but as soon as he goes all in we do have this information. Our EV pre flop after the all in is \$200*0.81139=\$162.27800. If we sold out seat, it would sell for \$162.27800 because this is our EV. The flop is T 8 5 Our EV now is \$200*0.09192=\$18.38400. If we could stop the hand before the turn is dealt and sell our seat on the market, it would go for \$18.384 because it's our expected value. The turn is K and our EV is now \$200*0.97727=\$195.45400. If we could stop the hand on the turn before the river is dealt our seat would sell for this much on the market. The river is 5 and our EV is \$200*0=\$0. Our seat would sell for no more than \$0. So from this single hand we calculated 4 different expected values: \$162.27800, \$18.38400, \$195.45400, \$0. They are all correct calculations in certain points of time but they are NOT equally useful to us. When we use the term EV we often really mean "the most useful EV". In this case, which EV calculation is most useful? To us, it is the first value, \$162.27800. The other EV calculations take into account things which no player had influence over (the order of the cards in the deck). We can most accurately determine the skill of players by using the first EV calculation and that's what we're interested in.
06-21-2012, 05:32 PM   #45
Pooh-Bah

Join Date: Jun 2009
Posts: 5,862
Re: Why is EV so difficult to explain?

Quote:
 Originally Posted by Karganeth The correct EV calculation doesn't care about the rules of poker or when a player went all in. The only thing it cares about is how much value you expect and the algorithm I proposed does exactly that.
I don't think you actually read my post.. you just restated your original argument without reference to my objections.

You keep saying "our EV". You're ignoring the fact that EV(on the flop) and EV(on the turn) are different. The value you expect changes when the turn is dealt.

People want to know what their EV was at the point they went all in. If it can't be calculated at that point (for the reasons you describe), they don't want to substitute the result of some other calculation. They don't care what the EV was after the turn card, in this example.

It seems that the Poker Tracker people know how to calculate EV, but your disagreement with them is which information should be a part of that calculation. They don't want to include the turn card in the calculation (if we went all in on the flop), and I would agree with that decision.

Can you explain what use it is to you personally, to know what your EV was on the turn in the example hands we are discussing? It's sure no use to me.

 Thread Tools Display Modes Linear Mode

 Posting Rules You may not post new threads You may not post replies You may not post attachments You may not edit your posts BB code is On Smilies are On [IMG] code is On HTML code is OffTrackbacks are Off Pingbacks are Off Refbacks are Off Forum Rules

All times are GMT -4. The time now is 12:46 AM.

 Contact Us - Two Plus Two Publishing LLC - Privacy Statement - Top