overadjusting

[Imthenewfish]

i was going to make a thread like a week ago on this because i wanted some feedback on a concept i hadn't really put much thought into. What do you consider an overadjustment? What would be an optimal adjustment?

i remember i was sweating someone awhile ago who decided he was going to 4b pre some redic 80+% in one spot where i agreed we should adjust by adding 4b bluffs, but how much do you adjust?

[Penneywize]

Well, if the spot you're talking about is 4bing pre as a response to a 3b with a presumably wide range from the blinds after we raise in LP, I am of the school that, in a vacuum - and say at a reasonably low / soft stake like 25 or 50 nl - we could 4b a very very large percentage and come out on top in the long run, barring readjustments.

In response to the concept of overadjustments in general, I really just consider this an inevitable mistake that players make in making adjustments. Incrementally, our 'adjustment' takes place in a single hand, and we may over-shoot or under-shoot over time; this becomes sort of blurry given the uniqueness of every hand and every situation, obviously, but I believe it is fair to say that you are either overadjusting or underadjusting as you begin to make any sort of adjustment.

To make a more concrete example of this: say we are in LP facing a raise from MP1, and have now accumulated enough hands on this particular opponent to decide that his range for raising in this position is rather wide. So our options to adjust are pretty obvious, and include 3betting lighter for value or calling a wider range in position. Assume we opt for one of those two decisions and follow through. Have we adjusted? Yes. Is it the absolute correct adjustment? Maybe, but the likelihood of this action exactly coinciding with the long-run tolerance of adjustments versus this player is incredibly low.

There are several reasons for this. First, we will almost always lack a sufficient sample size -- ignoring player style correlation -- to make a perfect adjustment versus a given player at a given point in time. Second, the player will not simply continue to play as such as he observes our adjustments to his playstyle, and will usually adjust himself.

So it's fairly plain to see that there is no such thing as a perfect adjustment; all of our adjustments, while certainly grounded in mathematics and poker logic, are either undershooting or overshooting an equilibrium. Our goal then should be to tailor our adjustments as we see to most closely coincide with this long run equilibrium.

A simplified example could be how we generally play against random players in nanostakes; our plan is to generally value bet, value bet, value shove our premium holdings and generally giving way when our opponents display unsolicited aggression against our non-nut hands. Assume our opponents do not properly adjust to this, calling us down light and not folding their more marginal hands, nor raising their weakest hands as a means of balancing their value raises. Our playstyle then reflects a "perfect" exploitation of our opponents. In reality, however, our opponents do adjust, and we never have perfect information on the types of opponents we are facing. We generalize, and hope our adjustments land within some tolerance that lands us maximum profit in the long run.

This is fine, of course. Perfect information and static opponent classification are fairy tales. In the meantime, however, we make adjustments with which we intend to maximally exploit our opponent's tendencies, and continue to adjust as our information on these opponents increases, and our history with them continues to evolve.

[Penneywize]

I should probably point out that my preceding post is almost completely conceptual / philosophical and not very useful for actual help in poker play. I need to preface my posts more

[Imthenewfish]

what do you mean by underadjust and perfect adjustment? other than that, i'm very familiar with the 'concept' of 'adjusting', and i'm pretty sure those examples aren't really related to what i was trying to get at.
i always thought an overadjustment was just taking an adjustment that was +EV in a vacuum, and taking it further. This should be +++EV, but i thought we don't usually do it because doing it is very exploitable, and will cause villains to adjust very quickly.
cliffs: Do you go batshit insane with adjustments and just rely on readjusting to how your opponents adjust, or do you gradually adjust and just focus on getting some eevees out of it without your opponent adjusting?

[Penneywize]

Newf you should know I wasn't trying to lecture you, I was really just posting up my own thoughts on the concept. Sorry if it came across otherwise!

What I meant by underadjusting was that, assuming we decide to adjust one way given an opponent's tendencies, say by widening a calling range, we actually do not widen it enough. So we've adjusted, but not quite enough as we should have. Say our opponent raises 50% in a given position in the long run but our stats on him indicate 30%, and we adjust by widening our continue range somewhat, while we should probably widen it considerably more.

By perfect adjustment I just mean the type of adjustment we could make had we had perfect information on the opponent, all else being equal (i.e. either ignoring or completely taking into account their own propensity to adjust). Again it's kind of a mythical thing and probably not very useful in actual analysis and play.

By the way I'm glad to see you're back on the forums!

[Penneywize]

cliffs: Do you go batshit insane with adjustments and just rely on readjusting to how your opponents adjust, or do you gradually adjust and just focus on getting some eevees out of it without your opponent adjusting?

I basically believe that there exists an optimal amount of adjustment for each and every opponent, i.e. an exploitation 'sweet spot' that enables us to get the maximum expectation out of them at a given point in time. So, to liken this to the two scenarios you mentioned 1) adjusting drastically, hoping to exploit the opponent maximally and in short order, and 2) adjusting more gradually, the trade off, it seems to me, is as follows:

The first method may allow us to make the maximum amount in the short term, however we are A: more likely to be adjusted against, and B: more likely to get the adjustment wrong.

In the second case, we are less likely to be immediately adjusted against, and less likely to make an overly erroneous adjustment, however we may give up some value in the short term.

Thoughts?

[rpm]

as best i understand it, there is the game-theory optimal (unexploitable strategy) and then there is the maximum exploitative strategy.

so for ie, let's say some 100bb stacked TAG opens to 3x in the CO with 53% of hands, and continues to a 3b only 12% of the time. we decide we want a healthy 3bet bluff range to exploit him.

the unexploitable strategy is to 3b something like 62% bluffs, 38% hands which we plan to 5b (CBF digging up the maths behind this as my external harddrive is being a bastard, but it's essentially bet/bet+pot)

so we would determine what hands we can profitably flat, what hands we are 3betting/looking to get it in with (A range hands), and then we would take the top of our fold range and add it into our 3b range until we are bluffing (ie folding to 4b) 62% of the time.

however, if we were seeking to play a maximum exploitative strategy, we would basically determine what hands we can flat profitably for our flatting range, and then 3b ALL the rest. this is because we have determined that, due to his fold %, 3-betting any two cards is +EV in a vacuum (ie we don't need any pot equity for 3betting to be +EV here), and so if we are looking to exploit him as much as we can, we want to be using those type of hands we just folded before (because they weren't profitable to flat, and we were concerned with balancing our 3b range for unexploitability) as 3bets, because they obviously have a higher EV as 3bets (because they have 0EV when folded, and we know that 3betting any two cards is +EV).

the latter would obviously be over-adjusting. any villain will notice when we are 3betting his opens something like 75-80% of the time. the way i've had it explained to me is to think of it as "defence" and "offence". seeking to play unexploitably is defensive, seeking to play an optimal exploitive strategy is offensive. there is no determining factor where between these two we should fall. however, there are some general situations we can consider:

- against the reggiest of regs in our games, we probably want to maintain something closer to an unexploitable strategy, because these guys are the most likely to be able to exploit us if we open ourselves up to it.

- against people who we rarely play against and probably won't play against for a while, we can be more ruthless in our exploitation (less time for them to cotton on/adjust etc)

- against people who we think are better players than us, we are best sticking to something close to the unexploitable strategy

anyhow, basically, it depends. some people are comfortable making very fluid and constant adjustments, and they can thus be more exploitative (and thus explotiable) in their strategy.

other people prefer to exploit their opponents to a smaller degree, but over a longer period of time. for this to happen we want to be closer to the unexploitable strategy (because this makes it less clear to our opponents that we are beating the shit out of them and they need to adjust)

i don't understand these ideas deeply, but this is it how it has been explained to me.