A new approach to calling ranges?

[TT_fold]

I got to thinking during today's sets of SNGs that a different approach to calling ranges might serve as a more accurate model for push/fold poker.

I see the current model as somewhat flawed in that assigning a calling range of 44+, AT+, KJ+, for example, assumes that an opponent will ALWAYS call with 44 but NEVER with 33. This seems too great a disparity (100% of the time with 44 but 0% of the time with 33) between two hands of similar value.

Let us assume that a villain with 10x BB will call a button open-push with AA 100% of the time, and 72o 0% of the time. Couldn't all other hands be assigned a percentage in this manner? In this example, I might set the TT percentage as 95%, K9o percentage as 25%, 76s percentage as 5%, etc.

Obviously this is a more complicated model than eastbay's, but theoretically it could be very useful in computing a more exact $EV for a push/fold situation. Thoughts?

[citanul]

how useful is this if you have absolutely no way to even start to guestimate this for your opponents? i doubt you could even accurately do this for yourself.

to top this off: are there hands that you call in a given spot 95% of the time in short stacked big blinds situations, and fold 5% of the time? i really really really doubt it.

my assumption is that the amount of utility that could theoretically be added by such a complicated model is tiny in relation to that lost by attempting to make minute assumptions about your opponents with very little information. ie, you are more likely to lose information than gain it.

citanul

[viennagreen]

i think that the usefulness of eastbay's program is that after using it for a while, you get a general idea of what the right play is-- and then you are able to apply your knowledge while playing....

with your proposal--- you could indeed set up some very exact situations (if you had total knowledge of your opponent). however, i don't see the utility in it.

[Misfire]

I don't think he's saying any particular opponent will call with a hand a certain percentage of the time % and fold the same hand the rest of the time. I think he's getting more at the idea that we can assume a random opponent fitting this guy's description and playing style is x% likely to fold this hand.

Whether this idea makes for a better EV estimate, I don't know. Like you said, it sounds like more trouble than its worth. I'm sure eastbay gets us in the ballpark. It's not like we can whip out his program in the middle of a hand anyway.

[TT_fold]

[ QUOTE ]
I don't think he's saying any particular opponent will call with a hand a certain percentage of the time % and fold the same hand the rest of the time. I think he's getting more at the idea that we can assume a random opponent fitting this guy's description and playing style is x% likely to fold this hand.

Whether this idea makes for a better EV estimate, I don't know. Like you said, it sounds like more trouble than its worth. I'm sure eastbay gets us in the ballpark. It's not like we can whip out his program in the middle of a hand anyway.

[/ QUOTE ]

Yes, this is the point I was trying to convey.

Like I said, this approach works better in theory than in practice, rendering it pretty much useless for us.

[Freudian]

[ QUOTE ]

Whether this idea makes for a better EV estimate, I don't know. Like you said, it sounds like more trouble than its worth. I'm sure eastbay gets us in the ballpark. It's not like we can whip out his program in the middle of a hand anyway.

[/ QUOTE ]

I don't see the point of breaking down a guess into decimals. Most likely the guess is incorrect anyway. If it is in the correct neighborhood it is ok and useful for analysis.

[roma12]

i think alot of you missed the point of the OP. Hes trying to show a donk perspective-which i am assuming, non of you have. Why have 44+ instead of 22+. to you yes it might be different, but to our fish, it is not. As well as A6s vs A2off...they simply don't know the difference

[citanul]

[ QUOTE ]
i think alot of you missed the point of the OP. Hes trying to show a donk perspective-which i am assuming, non of you have. Why have 44+ instead of 22+. to you yes it might be different, but to our fish, it is not. As well as A6s vs A2off...they simply don't know the difference

[/ QUOTE ]

so? so you put his calling range in your mind instead of at 44+ at 22+, and instead of at A6s+ at A2o+. that doesn't seem to be what the op was talking about at all. i dunno, i'm tired and this didn't seem of any theoretical or practical value to me when i wasn't tired. so i'm just going ot leave it alone.

citanul

[roma12]

[ QUOTE ]
[ QUOTE ]
i think alot of you missed the point of the OP. Hes trying to show a donk perspective-which i am assuming, non of you have. Why have 44+ instead of 22+. to you yes it might be different, but to our fish, it is not. As well as A6s vs A2off...they simply don't know the difference

[/ QUOTE ]

so? so you put his calling range in your mind instead of at 44+ at 22+, and instead of at A6s+ at A2o+. that doesn't seem to be what the op was talking about at all. i dunno, i'm tired and this didn't seem of any theoretical or practical value to me when i wasn't tired. so i'm just going ot leave it alone.

citanul

[/ QUOTE ]

ok....so hes talking about how a donk will call 95 percent of the time with XX as opposed to 100 percent of the time, which is what eastbays programs implies. Does this change anything? How are you going to put someone on "calling w/ XX __% of the time"...the weather? time of day? So yes, i was wrong in my original reply in terms of understanding the OP point. but does he have one? wouldnt lowering the "amount of times" someone calls with XX, be the same as decreasing his calling ranges? As to my original drunk response: Alot of examples on this forum have ICM calling ranges of 55+ , A6s+, etc. what i was saying before is, to donks... A6s=A2...

[TT_fold]

[ QUOTE ]
[ QUOTE ]
i think alot of you missed the point of the OP. Hes trying to show a donk perspective-which i am assuming, non of you have. Why have 44+ instead of 22+. to you yes it might be different, but to our fish, it is not. As well as A6s vs A2off...they simply don't know the difference

[/ QUOTE ]

so? so you put his calling range in your mind instead of at 44+ at 22+, and instead of at A6s+ at A2o+. that doesn't seem to be what the op was talking about at all. i dunno, i'm tired and this didn't seem of any theoretical or practical value to me when i wasn't tired. so i'm just going ot leave it alone.

citanul

[/ QUOTE ]

I guess what my formula accounts for is facing a random opponent. Let's be honest, when playing tons of tables at once it is tough to put players on specific calling ranges. Let us assume for a minute that half of the players at a specific level are total donks and half are tough pros. In that case, your push might get called by A2o 50% of the time.

Does that clarify my position at all?

[roma12]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
i think alot of you missed the point of the OP. Hes trying to show a donk perspective-which i am assuming, non of you have. Why have 44+ instead of 22+. to you yes it might be different, but to our fish, it is not. As well as A6s vs A2off...they simply don't know the difference

[/ QUOTE ]

so? so you put his calling range in your mind instead of at 44+ at 22+, and instead of at A6s+ at A2o+. that doesn't seem to be what the op was talking about at all. i dunno, i'm tired and this didn't seem of any theoretical or practical value to me when i wasn't tired. so i'm just going ot leave it alone.

citanul

[/ QUOTE ]

I guess what my formula accounts for is facing a random opponent. Let's be honest, when playing tons of tables at once it is tough to put players on specific calling ranges. Let us assume for a minute that half of the players at a specific level are total donks and half are tough pros. In that case, your push might get called by A2o 50% of the time.

Does that clarify my position at all?

[/ QUOTE ]

I think your thought process is so logically flawed-although i understand where you are coming from, that it is almost impossible to get a clear response to your original post-which i assume is what you are looking for. like i said in my previous response, lowering a villains percentage of the time he will call is the same exact thing as decreasing his calling ranges. When are you EVER going to get a read where you feel that villian is going to dump a certain percentage of the time, yet call w/ the same hand the remaining percentage of the time? The point is, i think you are asking for a more accurate mathematical-theoretical veiwpoint on push situations, and what i am saying is...theres nothing better than the ICM/eastbays program, analysis, in terms of pure mathematical examples.

[zoobird]

Not EXACTLY the same. If I assume my opponent's calling range includes ATo as 'shorthand' for the fact that I really think he would be only 50% likely to call with either ATo or A9o, its the same thing unless my hand is something like 99 or TT in which case it makes some difference. I don't know if its enough difference to matter though.

[pergesu]

I think I get what you're saying here.

Assume opponent will call 100% of the time when he holds AA, and 50% of the time he calls with KK. He will get dealt AA once every 220 hands, same with KK. Now he will only call you 3/440 of the time - 25% less. Then you figure out your EV after going vs AA, EV after going vs KK, and weight it by the probability before doing to over all ICM.

Wow. That would be some [censored] up math.

[viennagreen]

no--- that doesn't make sense.

it isn't like you have to set each player to have the same calling range in eastbay's program.... so...

i don't know--- maybe i'm not understanding your point... but you can set "donk1" to have a loose calling range and "pro1" to have a tight one, based on your observations, and doesn't that do the same thing?

[Isura]

[ QUOTE ]
I think I get what you're saying here.

Assume opponent will call 100% of the time when he holds AA, and 50% of the time he calls with KK. He will get dealt AA once every 220 hands, same with KK. Now he will only call you 3/440 of the time - 25% less. Then you figure out your EV after going vs AA, EV after going vs KK, and weight it by the probability before doing to over all ICM.

Wow. That would be some [censored] up math.

[/ QUOTE ]

Yup, that what he means. Although people are brushing aside the OP's idea since it isn't very practical, this type of calculations could be done very easily by a computer (ie bot). An intelligent system could do a pretty good job of assigning the probabilities based on player characteristics and the game situation.

[Freudian]

[ QUOTE ]
[ QUOTE ]
I think I get what you're saying here.

Assume opponent will call 100% of the time when he holds AA, and 50% of the time he calls with KK. He will get dealt AA once every 220 hands, same with KK. Now he will only call you 3/440 of the time - 25% less. Then you figure out your EV after going vs AA, EV after going vs KK, and weight it by the probability before doing to over all ICM.

Wow. That would be some [censored] up math.

[/ QUOTE ]

Yup, that what he means. Although people are brushing aside the OP's idea since it isn't very practical, this type of calculations could be done very easily by a computer (ie bot). An intelligent system could do a pretty good job of assigning the probabilities based on player characteristics and the game situation.

[/ QUOTE ]

Sure, with ten thousand hands on a player we could calculate calling range instead of guesing it. But we don't. So we make an educated guess. Sometimes not even that.

I think this is where experience comes in. After a couple of thousand SnGs we start to get better att guessing what a VP$IP 30% PFR 9% player with 1200 chips would do with AJs UTG in level 4. Not as good as fine combing a huge database would be, but good enough to give us an edge. I doubt we will ever be good enough to start thinking in terms of them dumping it 25% of times and pushing the rest.