On Imperfect Imperfection: A Theory Post (Mad Long Yo)

[Prodigy54321]

Before I start, I want to make it clear that although I will be writing in a "matter of fact" tone, I am not at all confident that everything I write will be true (and considering how often I revise my opinions on things, odds are I will make mistakes). Feel free to tell me I am an idiot (and preferrably why) [img]/images/graemlins/tongue.gif[/img]

I've alluded to this idea a few times in various threads in this forum, but haven't taken the time to really explain myself...so here we go

When I was an STT player, I wrote Tips for Using SNGPT. if you read the comments you will see that, looking back at this post, I'm not too proud of it, except for the 3rd part, which I wrote before other SNGPT-like programs started making EV graphs available and people could more easily see ceilings and floors. Even though Part 2 sucked, it led me to what I am going to talk about now*

*This whole thing may actually be much more useful as a STT concept since considering pushing and calling ranges is so important in STTs, but I think it still holds well as a NL cash concept and that is what I am currently playing

On Imperfection: we obviously all know that when we make educated guesses about what a person's particular range will be at a given point during a hand (I will refer to this as a "read" from now on, and when I say "guess" I of course mean, "educated guess"), we aren't always spot on. If we were, we'd all be rolling in money. But that is not what this post is about, this post is about Imperfect Imperfection

*a quick note: I believe that what I talk about here will hold for all points during a hand, that is, whether we guess a person's range that they are raising a flop bet with or pushing the river with, it is subject to what I talk about here. (I think it will be much easier for you all to be thinking about a simple situation though such as deciding whether or not to call in allin at some point during a hand, since when there will be further decisions down the line, it gets more complicated)

On Imperfect Imperfection: This is the idea that our imperfect reads are not equal in all directions, that is, if we guess what a person's range is, the fact that this read is imperfect DOES NOT mean that, when we are wrong, this person will be tighter (in terms of being ahead of us (including magnitude)) and looser (in terms of being behind us (including magnitude)) with the same frequency or magnitude.

you may be confused at this point, so I'll try to make things clearer (I don't consider myself to be very good at communicating my thoughts, so feel free to ask me to clarify anything)

consider the situation in This recent thread. (it may be easier to consider that our decisions are to push or fold at this point and when we push, CO will always call...I don't think these are a stretch, and even if they are, I think the point I am trying to make will still hold)

consider if we guess that villain has the follwing range

*this range has some weird stuff in it like AA-KK, and you might think that another range is better, but forget about that for a moment because it doesn't matter...to make my point, I actually adjusted what I thought was a better range because I wanted our equity to be slightly less than what we would probably need to push in this situation...

Board: Qs 8c 3h
Dead:

equity win tie pots won pots tied
Hand 0: 28.573% 22.34% 06.23% 12165 3393.00 { QhJd }
Hand 1: 71.427% 65.20% 06.23% 35499 3393.00 { QQ+, 88, 33, AQs, KQs, QTs+, Q8s, AQo, KQo, QTo+, Q8o }

*we need about 2:1 to make the push here, (which, as I said, I think we actually have but to make my point pretend for a moment that the above range is what we guess the range to be) as you can see, we are a bit behind that with this range

now I want to show you why Imperfect Imperfection may make it profitable to push anyway

we've considered many things in choosing a range to put villain on...his vpip, pfr, AF and what not...but I am sure you will all agree that this does not mean that he will never make a play that we would find to be unlikely given his stats...

so what happens when we are wrong?

Will his range be tighter sometimes and looser sometimes? Certainly.

But will it be tighter and looser with the same frequency? Will it be tighter and looser with the same Magnitude? No

how much tigher could villain possibly get? (let's just consider which hands he makes his flop raise with, not which hands in his range he will actually have at this point in the hand, that is, the hands he will call our PFR with (even though this concept still applies)...it'll just be easier this way and will get my point across...)

maybe he won't make this raise with QT or QJ or even KQ (which in this case means we are further behind more often)...

but this tighter range is not much tigher than the range we first guessed he would be on and it doesn't include very many hands.

Now consider just how much looser he could be.

his vpip certainly suggests that he could call our PFR with a wide range of hands...if we are off in our guess about what he will make his flop raise with we could be WAY off and even if we are not way off, the number of hands that he could be doing this with is much greater than the number of hands he might refrain from doing this with if he is actually tigher than we first guessed.

even just adding Q9 and Q9s is enough to raise our equity to a point where we should push...but that is not the point...

can you say you've never seen people make these kinds of moves in this situation with hands like Q9, Q7, JJ-99, or even 98 or 87? Not to mention to occasional random retarded plays that we see with hands like 44 or A3 here (for which there is no real counter in terms of tightening a range...unless this guy randomly folds 33 or AQ or something to our c-bet as often as he makes his raise with a terrible hand (which I do not think is the case))

these hands are not in our range because of how we guessed he would be playing because of his stats, but when we are wrong, there is plenty of room for his actual range to slide in the looser direction

I am not saying that it is very likely that we are way off about his range or even moderately off...what I am saying is that the chances that we are slightly off, moderately off, or way off need to be addressed...

I think most people do address it, but do not consider Imperfect Imperfection. That when they are wrong about a range, that range is not always free to be wrong equally in both directions.

We must consider how this will effect our true equity.

discuss

[CruS]

Very nice,
I'll think a little about what I think on the subject and give you an A for the effort.

[whyherro]

I don't want to be a party pooper, but this is implied every time someone puts together a range - specifically in that you always assign a certain % to air. This implicitly takes into account the fact that unaccounted for hands constituting his range are, taken in average, weaker than the specified range.

[Prodigy54321]

[ QUOTE ]
I don't want to be a party pooper, but this is implied every time someone puts together a range - specifically in that you always assign a certain % to air. This implicitly takes into account the fact that unaccounted for hands constituting his range are, taken in average, weaker than the specified range.

[/ QUOTE ]

I posted this because I often see people giving data from pokerstove with ranges that cannot get much tighter but can get much looser and offering thier advice based solely on the number that it spits out.

I am sure that there are plenty of people who consider the ways in which their input can be wrong and the ways in which the output may be affected (although I would argue that most do not give it enough weight, and I think it may be because they don't really understand why), but there are at least some who do not...and I guess this is for them

[Prodigy54321]

[ QUOTE ]
I don't want to be a party pooper, but this is implied every time someone puts together a range - specifically in that you always assign a certain % to air. This implicitly takes into account the fact that unaccounted for hands constituting his range are, taken in average, weaker than the specified range.

[/ QUOTE ]

also,

this whole thing is more in a hope that we can do a little better than just saying "consdier our equity to be a little higher because he might just be doing that with air"

I'm sure we won't be able to accurately quantify what I talked about, but if we consider that the effects of our imperfection are not just dependant on if the person might be "doing that with air" but also dependant on just how tight the person is expected to be in the situation considering a notion that they, for instance have a very low AF, I think it can help us change that '% that we assign to air' depending on the situation.

[EMc]

tl;dr


lolz


goot work for realzy

[DaycareInferno]

wut

[Genesis]

If I'm reading this right, at this late hour, this seems similar to the comment in Harrington on Hold'Em about figuring out your opponent's range and throwing another ~10% in there for the times he has air/is doing something unexpected.

[Profish2285]

First off lemme say good post and thank you for trying to explain this. I have a question though. You made this post mainly about the other post which you linked. In that post,the villain only had about 33 bb's I think. That of course left us as hero, in a push or fold situation. How much does all of this change if villain had 100 bb's or 200 bb's?

[TilTandWiN]

Thank you for the post. I am on my way to understanding the use of a range - however at the moment I have a real problem with it - how do I use this during real time play?