Technical question for the smart people

Bear with me please, this will probably be very confusing.

This is my graph of hands where I call out of the blinds to a UTG, MP, CO, BTN raise with a pocket pair <JJ preflop and take any action on the flop:

http://i80.photobucket.com/albums/j1...to/callpre.jpg

I can deduce from this graph that my overall strategy for calling pocket pairs from the blinds is profitable (over this sample size).

This is the same graph except here I only include hands where I c/c the flop after the same preflop action:

http://i80.photobucket.com/albums/j1...prec-cflop.jpg

Ignoring shania effects for a minute, is it fair to say that if I just c/f the flop every time I don't hit a set then I would be up more money in the first, overall, graph?

I'm concerned that if I want to analyse the profitability of only actions after the flop c/c rather than the hand as a whole which includes preflop, then I should be adjusting the bb/100 lose rate in the second graph upwards by a rate of 3bb/hand. The rate of 3bb is calculated because I assume that every preflop open raise is to 3bb (I know that's not realistic).

Does the first graph include c/f? When you say 'includes any action' I would assume that only includes calling/betting/raising, not folding. If so, you're missing all the hands you call preflop and then fold flop from the first graph.

Its late, maybe I'm not thinking clearly.

Yeah it includes c/f.

I guess I don't need to add anything to the second graph then. lol I feel like this is really obvious and I'm just being retarded because I can't get it clear in my head.

Ahh, I get where you're getting at now. I don't think so though, but having troubles putting into words why.

If you're trying to analyze those spots, I'd just use an overall graph like the 2nd one as your baseline, and then compare more finely grained graphs to it, rather than to 0.

Like now you filter where you c/c flop without a set. Then you filter for ones where you raise turn, then when you c/c turn, etc.

I generally find it easier just to use the reports screen for this kind of thing though. What do you get out of the graph, just the showdown/non-showdown lines? The reports screen lets you see bb/100 winrates which is pretty handy for this comparison stuff.

I think youre grouping way too many hands/ situations here. For example Set hunting Vs an UTG raiser is going to be vastly more profitable than set hunting Vs a button raiser.

Also given that most opening ranges include low-mid PPs from most positions, calling a cbet with 77 is going to be a lot more profitable than calling with 22 just for the times you get to showdown.

You probably need to reexamine this, seperating early position from CO/BU opens, and also splitting pairs into low and mid. You might need a lot more hands for that to be meaningful though.

edit: Does the second graph include hands where you flopped a set? I think you do need to adjust for the amount you called preflop, since you are concerned with the EV of calling at the flop. If you tend to raise the flop with a set then your graph just shows that calling a preflop raise and missing your set is unprofitable. Well duh.

Quote:

---

Originally Posted by Pelion

I think youre grouping way too many hands/ situations here. For example Set hunting Vs an UTG raiser is going to be vastly more profitable than set hunting Vs a button raiser.

Also given that most opening ranges include low-mid PPs from most positions, calling a cbet with 77 is going to be a lot more profitable than calling with 22 just for the times you get to showdown.

You probably need to reexamine this, seperating early position from CO/BU opens, and also splitting pairs into low and mid. You might need a lot more hands for that to be meaningful though.

---

Yeah well that's the problem I'd need more hands to separate. I'm just trying to get a rough idea of something.

Quote:

---

Originally Posted by Pelion

edit: Does the second graph include hands where you flopped a set? I think you do need to adjust for the amount you called preflop, since you are concerned with the EV of calling at the flop. If you tend to raise the flop with a set then your graph just shows that calling a preflop raise and missing your set is unprofitable. Well duh.

---

Well yeah as long as I c/c with my set and check raise.

Nah it shows that missing my set and c/c-ing is unprofitable I think. It doesn't include hands where I c/f the flop after calling preflop.

Quote:

---

Originally Posted by bjsaust

Ahh, I get where you're getting at now. I don't think so though, but having troubles putting into words why.

If you're trying to analyze those spots, I'd just use an overall graph like the 2nd one as your baseline, and then compare more finely grained graphs to it, rather than to 0.

Like now you filter where you c/c flop without a set. Then you filter for ones where you raise turn, then when you c/c turn, etc.

I generally find it easier just to use the reports screen for this kind of thing though. What do you get out of the graph, just the showdown/non-showdown lines? The reports screen lets you see bb/100 winrates which is pretty handy for this comparison stuff.

---

Yeah i'd need a lot more hands to get any finer. Really I'd imagine that my flop action can be divided into 3 categories

hit a set and c/r
miss my set and c/c
miss my set and c/f

I wont have near enough hands if I want to go finer than that.

idk why I used a graph, it's actually harder to visualise I guess because one gets confused by the rates of change. I just used a graph to look at my 3bets once and was blown away by how fast the redline went up and it made me think better about where our profit comes from in 3betting. So I use them for this type of stuff a lot.

So you're down 1100 big blinds when you c/c flops with low pairs.

You're up 500 big blinds total with any action for low pairs. So if we assume that we can neglect donk leading (you prob don't do it that often), then all your actions are c/c, c/r and c/f.

so (c/c) + (c/r) + (c/f) = 500
(c/r)+(c/f) = 500+1100
so your winnings for either c/r or c/f are 1600 big blinds

So it could just be that you're winning a ton c/r and should be c/r more and not necessarily c/f more, if I'm thinking about this right.

It would be interesting to see you filter for the times you c/r with a set, and how much of those winnings come from showdown versus non-showdown. If you're winning a ton without showdown in that spot then it might be profitable to start c/r more non-sets even and using fold equity.

it probably means you call too much with 77 on QT2 vs opponents that will own u postflop.

C/f more, or mix in calling with sets a lot more often vs the guys that are good

Quote:

---

Originally Posted by griffey24

So you're down 1100 big blinds when you c/c flops with low pairs.

You're up 500 big blinds total with any action for low pairs. So if we assume that we can neglect donk leading (you prob don't do it that often), then all your actions are c/c, c/r and c/f.

so (c/c) + (c/r) + (c/f) = 500
(c/r)+(c/f) = 500+1100
so your winnings for either c/r or c/f are 1600 big blinds

---

How do you have any winnings when you c/f? Are you saying you lose 1100bb by c/c so c/f would gain 1100bb? I don't think this is the case, since Im fairly sure the bb gain/loss amounts are from the beginning of the hand, and in any case c/f would still gain (or lose) 0bb.

Quote:

---

Originally Posted by Pelion

Quote:

---

Originally Posted by griffey24

So you're down 1100 big blinds when you c/c flops with low pairs.

You're up 500 big blinds total with any action for low pairs. So if we assume that we can neglect donk leading (you prob don't do it that often), then all your actions are c/c, c/r and c/f.

so (c/c) + (c/r) + (c/f) = 500
(c/r)+(c/f) = 500+1100
so your winnings for either c/r or c/f are 1600 big blinds

---

How do you have any winnings when you c/f? Are you saying you lose 1100bb by c/c so c/f would gain 1100bb? I don't think this is the case, since Im fairly sure the bb gain/loss amounts are from the beginning of the hand, and in any case c/f would still gain (or lose) 0bb.

---

Yah I was thinking this too. That the "gains" of c/f, are mostly just not losing more money. So yah i suppose that means that c/r gains hero 1600 big blinds on its own.

This is my graph of hands where I call out of the blinds to a UTG, MP, CO, BTN raise with a pocket pair <JJ preflop and check/raise the flop. It includes flopped sets and bluffs with a bare pair (it's gonna be like 98% sets).

Keep in minds I fold pp's < 66 a fair bit preflop to this action. Looks like I probably shouldn't though.

http://i80.photobucket.com/albums/j1.../calprec-r.jpg

So now we have

c/r = +2034
c/c = -1060
c/f = 0

c/r+c/f+c/f = 974

So where'd I lose like 654bb? There's no way I lost that much by donking.

Maybe the rest is in spots where the flop goes ck ck, and then you lead and get called or then you c/c c/c or something?

All the money you put in preflop then lose when you c/f should be the 645 shouldn't it? Or does this filter only count the money you put in postflop?

Quote:

---

Originally Posted by griffey24

Maybe the rest is in spots where the flop goes ck ck, and then you lead and get called or then you c/c c/c or something?

---

checked that and I'm up 400bb from spots when the flop checks through.

o wait hold on I think I might know whats going on

Quote:

---

Originally Posted by meeloche

All the money you put in preflop then lose when you c/f should be the 645 shouldn't it? Or does this filter only count the money you put in postflop?

---

this

unless im massively misunderstanding

looks like the the first graph isn't what I thought it was.

You're trying to ignore the money up until the flop, but you really cant (and I don't think should) do that.

One more coffee and my brain should start working for the day.

Quote:

---

Originally Posted by Pelion

I don't think this is the case, since Im fairly sure the bb gain/loss amounts are from the beginning of the hand, and in any case c/f would still gain (or lose) 0bb.

---

No, c/f loses 3bb. But thats better than losing say 8bb if you call a 5bb cbet then fold on the turn. Just losing less with losing hands will increase overall winrate. The question Irisheyes should be asking, is if its best to take a guaranteed small loss when he misses, or should he try to make a win by c/r'ing sometimes instead but risk bigger losses instead. One way to help determine this is how often his c/r's are seeing showdown atm.

Quote:

---

Originally Posted by meeloche

All the money you put in preflop then lose when you c/f should be the 645 shouldn't it? Or does this filter only count the money you put in postflop?

---

HEM filter only filters out hands. Results will always be full results for the entire hand, no matter what street you filter on. Might be possible to get that kind of info with some direct SQL on the DB, I still dont think you'd want to. You want to measure profitability on the entire hand, not just a specific street surely?

if the first graph was what you thought it originally was then yes, if you ignore shania, if you c/f every time w/o a set it would increase your winnings in the first graph. Of cours, the first graph isn't what you thought it was, and ignoring shania in how you play sets is silly.

So I got some sleep and came back to this fresh

All included hands follow the following formula:
Open raise at a table with more than 2 players from UTG, MP, CO or BTN,

a. I fold in the SB or BB.
627 hands, -70bb/100

I call in the SB or BB with a pocket pair < TT. We go HU to the flop and one of the following actions occur:

b. c/r only
94 hands, +2695bb/100

c. c/c only
451 hands, -235bb/100

d. c/f only
702 hands, -336.55bb/100

e. c/check only
584 hands, +62.3bb/100

Totalled:
c/check + c/f + c/c + c/r. No donking.
1831 hands, -28.67bb/100

The individual conditions add up to the total so this is correct.

So Conclusions:
1) My current strategy is better than folding every time preflop.
2) My check/calling flop strategy loses me money, but less than I would lose if I started check/folding the hands that I currently check/call.

Check/raising is awesome but doesn't win without showdown enough to make me think that check/raising sometimes when I haven't hit a set (i.e. with 2 outs) is a good idea.

Your c/c and c/r situations are biased. The c/r situation is biased by the fact that you have a huge hand/set everytime, and are probably winning huge pots. The c/c situation is biased by the fact that you rarely have a set and are likely giving up to heat.

What you probably should compare is the c/r situation with a set vs the c/c situation with a set, or the c/c and c/r situation without a set.

Can you subdivide each of the three categories into 22-55 and 66-99 please?

Quote:

---

Originally Posted by griffey24

Your c/c and c/r situations are biased. The c/r situation is biased by the fact that you have a huge hand/set everytime, and are probably winning huge pots. The c/c situation is biased by the fact that you rarely have a set and are likely giving up to heat.

What you probably should compare is the c/r situation with a set vs the c/c situation with a set, or the c/c and c/r situation without a set.

---

Yeah I know I'm not really trying to compare c/c to c/r. Originally what I was looking for was a leak in c/c-ing. I suspected I was losing lots of money there but I guess if my sometimes c/c, sometimes c/f strategy is better than a pure c/f strategy it cant be too bad. Obv can always play better though.

Quote:

---

Originally Posted by mcatdog

Can you subdivide each of the three categories into 22-55 and 66-99 please?

---

Yeah sure. Keep in mind I fold preflop a lot more with 22 than with 99.

22-55, 675 instances altogether
-71bb/100

66-99, 941 instances altogether
-12.78bb/100


c/c flop, 66-99
267 hands, -192bb/100

c/c flop, 22-55
82 hands, -399bb/100

c/r flop, 66-99
48 hands, +2707bb/100

c/r flop, 22-55
38 hands, 2443bb/100

check/check flop, 66-99
296 hands, +77bb/100

check/check flop, 22-55
222 hands, +5.6bb/100

c/f flop, 66-99
330 instances, -343bb/100

c/f flop, 22-55
333 instances, -329bb/100

Results are pretty in line with what I expected.
22-55 < 66-99 across the board.
playing 22-55 from the binds costs pretty much the same as folding them for my sample.

I brainstormed a bit and came up with some waffle about improving my play after a flop c/c. It's all disjointed and probably shit but I'll throw it out anyway in case anyone wants to read my brain dump:

1) c/ring the river if he checks back the turn and bets the river.

This is usually a good move since the villain's range is weak since he would most likely bet the turn if he had a good hand rather than check the turn and only bet the river.

We need to be careful that we rep a range that can have strong hands in it when we go for the river check raise however. This is often a problem because if the flop is anyway wet then every villain assumes that you will c/r his continuation bet with any strong hands you could have. Also if draws blank off on the turn and river and you check/raise the river then the villain will often assume that you have a missed draw and could make a big call.

Perhaps dry paired boards are decent for this because we would often c/c the flop with trips on a dry flop and take this same turn and river line when the villain obliges.

Also this might be ok if an obvious draw like a flush comes in on the river. TAG villains will usually double barrel their flush draws on the turn so when the turn checks through we can assume that they couldn't have hit the flush on the river very often. On the other hand however we might often play a flush draw by c/cing the flop, checking the turn and then going for a check/raise on the river when the flush hits. Will the villain assume that we would lead the river if we made a flush? Perhaps a problem with this is that lots of villains wont VB the river thin here because they are afraid of getting check/raised.

I think in general people are not very apt to do a bet/check/bluff bet line IP so perhaps we need a board where a villain would be likely to pot control TP on the turn and go for a bet on the river. Villains who pot control a lot are not apt to make big hero calls?

2) Improve my calling down skills on various turn river cards? Scare cards?

3) c/r the turn sometimes

This is something I don't really like because
a. Villains often have a relatively strong range when they bet the flop and turn
b. It's hard to have more than 2 outs
c. We rep a weak range because usually we would have check/raised the flop with our strong hands

Two situations where it might be ok
1) The turn card is something which looks to have hit our range well, say the board is Tc 8h 4s [9d]. This is a good card for our range because it hits 9T, TJ, 89, QJ, 99. Also it is conceivable that we would c/c this flop with 88 or TT. The major problem with this hand is that usually villains will have a decently strong range when they bet this turn. They might b/f AT/KT/QT but they would b/c most other hands and perhaps wouldn't even bet the hands from the b/f range in the first place. The villain can have as many nut hands in his range as we can
2) The other situation where this might be ok is if we had 8x8d on Jd 6h 3d [Td]. When the board is like this we can rep some flush hands when we c/r the turn, we have more outs since sometimes the villain will call and we will hit a diamond on the river and win. Keep and mind that we can pretty much never go broke on the diamond river if called on the turn and usually we will have to c/f the river on our best out, sometimes getting bluffed off the best hand in the process.

4) Alter my flop c/c ranges such that I have more 56s with BDFD type stuff than 88. Or more A-high and overcards than pocket pair.

5) c/f the flop? c/r the flop?

Dealt with this in the conclusions of the stats from HM:
My check/calling flop strategy loses me money, but less than I would lose if I started check/folding the hands that I currently check/call.
Check/raising is awesome but doesn't win without showdown enough to make me think that check/raising sometimes when I haven't hit a set (i.e. with 2 outs) is a good idea.

All seems good except number 1 for the reasons you indicated
You rarely have a good hand when you get to the river in that spot and you would rarely CR because you wouldn't be expecting villain to bet alot because he checked behind turn. Ergo it is a very standard bluff catching spot in villains shoes.

2. Is good - I would do a filter where you check called flop with an unimproved pocket pair and folded to a turn bet in comparison to when you called the turn bet. I think you will find calling the double barrel is much more profitable than folding to it assuming you thought you were ahead of his range on flop.