Chip EV custom stat and graph

[selection989]

Hie guys,

I was quite impressed that pt3 now displays an ev graph for tournies but feel that it is missing something. To illustrate, consider the following:
Say I have 1500chips on the river, holding a royal flush in a $10 heads up sitngo and shove and get called by villain. In this case, my ev in chips goes up by 1500chips and if the blinds are 15/30, then my ev goes up by 100 bb on the river(excluding the money originally in the pot for simplicity). If right after this, I do this in a heads up sitngo, then my ev in chips goes up by another 100bb and in chips goes up by 1500. In total I am now up 3000chips ev wise and 200bb. Obviously it makes no sense to compare these. For this reason I propose that we create a stat specifically for heads up called where chip ev is simply:
i) EV(Ammount of of bbwon)/(starting stack in BB)*starting buyin.

this way, the EV's are scaled according to the buyin level(which they should be).

Is this possible??

[PJs Ronin]

Why not remove all scaling factors and just make it all-in EQ% success or fail?

[kraada]

selection989,

That's not currently possible as the starting stack size is not stored in the database directly, and neither is the equity values for each hand. The latter is planned, and once you have that you could probably work around it to build a stat which dealt with the former, but without equity being part of the database you won't be able to get off the ground here.

PJs Ronin,

That's been requested, and once we have equity in the database we can certainly do it. But in some sense it wouldn't really show you how lucky or unlucky you are. If you're playing a big MTT and win a whole bunch of all-in hands vs very short stacks around the bubble when you're a big stack, but then lose a bunch of big flips later to end up in 6th instead of making the top 2 my intuition tells me that the luck from the all-ins vs the shorties around the bubble doesn't make up for the bad luck of losing the big flips at the final table - losing those more directly impacts your bottom line therefore they should weigh more somehow. I don't know the right way to do it, but just a straight up count of all-ins doesn't seem to capture everything that's important.

[PJs Ronin]

If you're playing a big MTT and win a whole bunch of all-in hands vs very short stacks around the bubble when you're a big stack, but then lose a bunch of big flips later to end up in 6th instead of making the top 2 my intuition tells me that the luck from the all-ins vs the shorties around the bubble doesn't make up for the bad luck of losing the big flips at the final table - losing those more directly impacts your bottom line therefore they should weigh more somehow. I don't know the right way to do it, but just a straight up count of all-ins doesn't seem to capture everything that's important.

Luck has nothing to do with it. For tourney players, it's all about the math.

For MTTs, graphing chip EV introduces more confusion than it does provide clarity. In a graph, winning 1000 chips from 3 all-ins early in an MTT would get swamped by another all in when the blinds reached 10,000/20,000. I maintain that tourney players operate under 2 tenets. First, most times they make up their mind as soon as they see their holecards as to what they will do. Second, every tourney player is trying to get as many chips as possible into the pot, as soon as possible, with the best hand possible. Thus, tourney players ask two questions. One: Was I in front when the all in (from either hero or villain) was made? Two: Did it hold up? Graph me that, or wrap those stats into a report, and I can start making some qualitative decisions about tourney play.

And I'm not sure why this can't be done now, without any changes to the db schema. We are only looking at hands that went to showdown, therefore we know hero and villain holecards, the equity and the result. Provided PT3 can identify that someone is all in, the rest is just math. There are a bunch of posts ITF requesting this, let's stop poodlefaking and 'git er done'.

[/hijack]

[kraada]

We are only looking at hands that went to showdown, therefore we know hero and villain holecards, the equity and the result.

No, we don't, and that's exactly my point. Equity data is not currently in the database. Everywhere you see it currently it is being calculated on the fly, which is why we do not have any equity statistics in the Player Statistics section.

What you can do now, though, is create a custom report in the Hands which shows hands where you ended up putting in all of your chips in, and who the winner was. In fact, I've created that report for you and attached it to this post. It will show you your hole cards, the hand number of hands which were all in, your equity at the time, and who won the hand. If you want to replay any individual hand you can double click on it, and you can further drill down using the Filters -> Simple Filters section to look only at all in situations of any type that you prefer.

I hope the report helps you in your analysis.

Ninja Edit: I've also posted this in the repository, but it may take a little while to get approved and available there.

[PJs Ronin]

Appreciate your efforts kraada. This is so frustrating in that I can see the brass ring but it's just out of reach.

Is there anyway I can filter on M in the report you provided? Anything where hero or villain M < 10 (although I'd like to be able to adjust that number) needs to be removed.

If not, then I will admit defeat.

[WhiteRider]

You could add a custom filter to the report, like this:

(tourney_holdem_hand_player_detail.amt_before / ( tourney_holdem_blinds.amt_bb + tourney_holdem_blinds.amt_sb + (tourney_holdem_hand_summary.cnt_players * tourney_holdem_hand_player_detail.amt_ante) ) ) > 10

..to show only hands where the M in the hand is greater than 10.

[kraada]

Just to add: that will go by the M of whatever player is selected. You'd have to help me figure out how to define "Villain" in this case a little better - we don't want any player dealt into the hand because they could just open fold and then they don't matter . . .

[PJs Ronin]

Just to add: that will go by the M of whatever player is selected. You'd have to help me figure out how to define "Villain" in this case a little better - we don't want any player dealt into the hand because they could just open fold and then they don't matter . . .

Agreed. However, as we're only interested in hands that went to showdown this would exclude open folders. The 'showdown ' aspect should also help identify villains because anyone who gets to showdown is a villain... unless they are Hero.

Will try your spiffy custom filter later today... ty.

Ninja Edit: Hmmm... I think we're looking at this from two different perspectives and thus I may not have addressed your point. Is your focus at the beginning of the hand and mine is at the end?

[PJs Ronin]

Ummm... kraada, your report including hands where Hero did NOT go to showdown... not good.