Winnings/equity with money already in pot; Defining hands

[Stally]

Hello,

I'm going to refer to all-in adjusted equity as EV here for simplicity, I realise you guys consider those as subtly different.

I'm very interested in making stats that show how profitable certain actions are based on EV/winnings, but I'd need that to be modified to not count chips already added to the pot. I have some experience making stats already.

For example,
If I looked at my own hands, I would like to calculate the EV of calling flop with bottom pair, so I will be adding money already called/bet on previous actions to the EV of this hand. e.g. if I call a min-raise from BB, then I put 2bb into the pot and will add 2bb to the hand's EV. Another example would be a decision preflop from the BB - 1bb already in the pot. (A bit more simple of course.)

Is it possible to know how many chips have been played already? Or do I need to deconstruct previous streets betsizes to calculate this? If it matters, this will be for CAP play, so is it possible that there is something such as effective remaining stack to help, if nothing else is viable?

What do I reference for EV/hand?

I also am not sure how to classify made hands, and/or board and my actual holding. Is there a link that shows this info?

This above is just an example, as I'm hoping to create a variety of stats in this manner.

Also, I would also like to calculate the EV as a proportion of total EV. e.g. if 0.1% of my hands are me calling flop with bottom pair, this play adds 0.1% of the stat's EV to my overall EV (If my EV/hand of the decision was +1bb, then I gain +0.001bb p/hand overall, or 0.1bb/100). For this I assume all I would want to know is the stat for total hands of the player in db - then divide the "cnt_[...]" by this total hands?

I know this is not a simple request, I appreciate any response.
Thank you

[kraada]

We store all of the bits of individual data that you would need to analyze in this general way - bet sizes are stored and we have effective stack sizes by street which should give you enough information there. What I'm not sure you can do is go from that data to actually figuring out how much should have been won with a hand - there are very fine nuances to boards which I suspect may make this sort of analysis "blow up" into an intractable problem.

Let's try to construct a simple example and I think you will start to get a feel for what I mean. Let's look at top pair top kicker. First of all, let's eliminate paired boards entirely and boards with flush draws and boards with possible straights - each of those kinds of textures would need to be in its own category as things like ATT we could be behind against trips, monotone boards we could be against a made flush and AT9 two tone we could be against QJ with the flush draw which has extremely good equity against us. That's already a whole lot of categories and we're still on the flop!

Even with our rainbow, unpaired, non-straighty board when the turn falls a variety of things could happen - a backdoor flush draw can come up or not, a board can pair or not . . . each of these double the analysis space. And that's for a simple made hand like "top pair top kicker" . . . it only gets worse for things like middle pair good kicker imo.

I think you'd be better off picking a specific kind of made hand and drilling into it (because that kind of analysis is incredibly useful!) rather than trying to knit all of it together to figure out how good "ATs" is in some sort of theoretical sense.

[Stally]

I appreciate your concern for me focusing on the wrong information, or creating too detailed an analysis. Actually I wasn't 100% sure what you meant in your response - my initial idea was to look at a specific kind of made hand, the same as you suggested at the end - my initial example was calling bottom pair on the flop.

I do understand that different board textures are going to give different results but that is something I can look into if I decide to focus more specifically on certain areas after my initial stats work. I've done a lot of work already using other programs on equities, ranges, different boards, decision trees, etc, but I think that looking at my actual data is going to be very useful, rather than modelling situations. I should also point out that I play CAP, so this simplifies the problems you're talking about - for example when you flop tpgk and raise, it's a near-nut hand, and it's either all-in on the flop if you get 3bet, or the rest of your chips are all-in to one more bet on the turn.

So the goal is to start of with fairly broad stats, and then any that look interesting I can dig deeper with. I should be hopefully able to tinker up various stats on my own once I know how to play with holdings, boards, and made hand types.

When you say "figuring out how much you should have won", all I am looking for is the "all-in adjusted winnings" from my results of performing these actions in the past. If it isn't possible to find this for the hands that will pass the filters into the stat, then actual winnings will suffice, but it will be a bit of a hindrance for my results.

Thanks for your response,
Roy

[kraada]

CAP definitely does simplify things considerably, especially as you will likely only be looking at texture on the flop or, at most, the turn.

I'm still not sure how you'd do this necessarily in a single report but we do have a wide array of texture filters in the More Filters -> Board Texture area. If you created a simple Player report (or Hand report even so you could replay hands) you could add Net Expected Won and then use the More Filters area to see how things change in various ways when these filters are applied.

[Stally]

I want to create different stats, each for specific situations, that will show my EV in those situations, adding on the bets already put into the pot at that point.

So for the example of bottom pair on the flop, i just create a stat that shows the EV of calling bottom pair on the flop, but adds 2bb to that (if i called a minraise preflop).

The idea is to have all these stats available to create a report, that I can look at all in one go:

to either see leaks, or which areas i lost in during particular periods of time, or comparing play from different periods with each other if I generated a couple of reports from different time periods.

[Stally]

The point is that I do limited but similar things right now, it's just too time consuming to set up all the different filters, look at each group individually, and either take notes/screenshots/save reports to show an individual situation - I just want to all everything all at once.

[kraada]

Real EV is such a complex concept I really don't see how to put this together in a meaningful way. I wish you the best of luck in your project and if I can help with specific pieces I'd be glad to.

[Stally]

I don't want EV of imaginary ranges, just my all-in adjusted winnings for the hand when I make a call/raise in a certain spot. Then if I add on the amount of chips already put into the pot previously, it gives me a good idea of how profitable said call/raise is.

[Stally]

What I need is to know how I find out how many chips are put into the pot me, what hand strength I have, and the all-in adjusted winnings of that particular hand. I would also like to know how to determine board cards/types and my holecards, for some of my the stats I will make.

[Stally]

I think the best thing is for to exactly define what I need to get me started. So I'd like to create a stat to show how well I do when i call a flop cbet with bottom pair (not pocket pair, no straight or flush draws).


(1) a) So I need a column to calculate the sum of (all-in adjusted winnings for every hand that has the following criteria: call a 2bet preflop, call a flop bet with bottom pair [and no gutshot, no open-ender, no flush draw, not pocket pair], add to this money I put into the pot preflop*).

b) Count how many hands with another column (which I should be fine to do).

c) Then I can create a stat dividing top column by bottom column (this again I have no trouble with).

* This is the key part - if i called 2bb pre (or was big blind and called 1 bb), and my hand lost 1bb total, my hand gained 1bb from the flop onwards (that it happens on the flop or later is of no consequence)



(2) I'd also like to do the same thing for preflop. Here what I'll need to know is how to choose specific holecards (using the above information and my own knowledge of the stat creation, I will be fine with the rest of the column/stat creation).