Last night while lying in bed I had a thought about this. No idea if its an actual theory or if its even a credible idea but thought Id ask about others thoughts on this.

We try and define our opponents range as they do with us. Using our HUD after a large enough sample size we can get a close approximate percent of hands played by position. This allows us to formulate a range and are able to adjust our play.

Say we end up on the river and opponent bets. We can call and usually get a hand that corroborates our thinking. Given we assume a certain percent is bluffs, certain percent is value, etc. The hand usually makes sense in some way.

My idea of a range distortion is either opening or even flatting with a hand that is nowhere near what would be considered in our range, doing so IP usually. This would be done at a certain frequency, say like 5 percent of the time from CO.

The goal of this is to distort the opponents view of your range. It would cause the opponent to adjust somewhat. Thus changing their play to be further away from the optimal manner for what your range actually is.

This is obviously not a fully fleshed out idea of course. Does this already exist and what are your thoughts on the feasibility of this?

http://forumserver.twoplustwo.com/54...guide-1352539/

See also SHANIA theory: http://archives2.twoplustwo.com/show...fpart=all&vc=1

Although I think there's something to be said for building an "image", as some villains will over-adjust because of one hand played weirdly in a small sample, I think the theory is basically nonsense for long-term profits against regs. They will probably think you just misclicked. By adding -EV hands to your range for metagame/image purposes, you're just putting a -EV hand in to your range, and burning 3bb in the process (unless you somehow bink a surprising monster). If you want to balance, it's better to put a +EV hand into your range obviously.

In addition, a well balanced range already has some "surprises" in it for unobservant players. You might 3-bet 54s at some point, and the victim might incorrectly assume you 3-bet ALL suited connectors in that spot, or that you always 3-bet 54s in every spot, when you're actually only 3-betting a few combos with mixed frequencies.

More of a fan of adding mixed frequency low equity/0ev hands than something just bad.

Say you have 100 combos of hands that can defend vs. a 3-bet. you have 1000 combos of other hands to choose from that can potentially be +EV opens (albeit very small +EV), but not all can really defend vs. 3-bets. Therefore to remain unexploitable you can only open a certain percentage of these.

Say you're UTG for example and you're opening all AXs, 54s+ 86s+ or something. You will get owned vs 3-bets, but a lot of these low equity hands can be profitable and/or breakeven EV, but your range only allows you to have so many of them. Rather than only opening 87s+ or something I personally think opening other hands at mixed freq to maximize board coverage and make you harder to play against is arguably better.

Could be wrong, but just my opinion.

+1

basically you're talking about the concept of 'advertising'

This was exactly what I was getting at. Not sure if we differentiate anywhere in our thinking but I meant taking a grouping of hands that are not in your standard range but are essentially BE if villain is interpreting you only playing your standard range. If you played them consistently and it ballooned your VPIP or PFR you would get exploited based on your new range being too wide. If you only play them at a small frequency Villain will be put in much tougher spots this way seeing a discrepancy between HUD numbers and hands showing up here and there.

Then when you have compiled that grouping of hands, only opening at a certain frequency. Not every time you are dealt one of these hands.

Correct?

Yes. GTO play includes a lot of mixing. See below.
Yes, seems correct.

Example from Snowie (which is probably a decent approximation for optimal pre-flop play). This is a chart for the button's play when faced by an UTG open of 2.5bb at 100NL:



The green squares indicate hands that are 3-bet at 100% frequency. The ones with numbers utilize mixing. e.g. JTs is 3-bet 80% of the time, and called the other 20%. It's slightly better than breakeven as part of the overall strategy. 53s is only 3-bet 6% of the time, and folded 94% and it breaks even by doing so. Having that hand in your range means villain can never entirely rule out you flopping a boat on 533 (or a straight on A42) in a 3-bet pot, but it also means he should 4-bet bluff occasionally, because he knows you have some "junk" in your range. Since you have some very weak hands in your "occasional" 3-bet range, and these weak hands induce villain to 4-bet bluff, the EV of your value hands is slightly increased by balancing them with the weaker stuff at low frequencies.

If Snowie 3-bet every single hand on that chart at 100% frequency, it would be playing like a maniac, since over 12.2% of hands are 3-bet at some freq, but by mixing the frequencies of some combos, it's only 3-betting 8.5% overall.

It's kinda like polarizing your opening/3b range no? What broken stars said

Is Snowie's graph for an utg open at fr or 6m? I wonder how much it'd even make a difference for that spot otb vs an utg 2.5x for fr vs 6m?

snowie's always loved those suited Ace wheel combos

I do this occasionally but it's a psychology thing not really a poker theory thing. I'll show down 94o and the OMCs will lose their minds at which point I'll say "it was off-suit though."

I assume those EV stats are vs nemesis and not an average vs the population?

I agree with most here that playing odd hands to try and mislead any players about our ranges isn't going to work vs an online HUD. You can't play a hand that's marginal for a 35% range and expect it to sell to our opponents when they have our stats at 12% range. They'll just see a one-off attempt to look like a maniac which is exactly what it is.

However, a small number of marginally -EV Variance hands that were specifically targeted to patch a range up a bit could be well worth the small price to play.

Let's assume that we are stuck with a fairly transparent calling range that's looks like the Snowie flatting range given above. By adding just a few variance hands until we showdown one example in the lower SC SG range then they we may pick off a bluff on a board they thought missed our range completely, and make it much tougher for them to steam roll low boards in the future.

They'd also have to start giving full credit for us playing those ranges since it's only marginally outside our normal range and a 1% change could be just variance or a minor adjustment to our range. Then, if your just not that comfortable and afraid of making a mistake with such a hand! you can discount their play to 50% or 25%. Until you become more comfortable playing them.

So for the price of equity that adds up to maybe -0.5BBs EV on average 5%-10% of the time, we get much more value from our play for the other 90-95% of our range whether it turns our set/air range to a set/draws range for getting extra value bet or two or preventing getting exploited on low boards by polarized ranges.

Just because we're willing to play with -EV hands doesn't mean we need to overplay them.

This^

I think this is more or less the same concept as "board coverage".
We just had a really good discussion on it w lots of good insights from these 2+2 regs.

We do miss out on something if we open super tight UTG and go to a flop like AQ9 or KQJ or AKT because we have so little air. That means two things; 1) we rarely have an air ball that benifits from our range advantage and 2) we will have a tough time value betting many of our top pairs for value because we don't have enough bluffs for villains to catch .

We also miss out when we open a range like 99+,KQ,AJ+,ATs+ from up front and catch a flop like; 765 or JT8 or even 554 because we now don't have enough big hands to bet multiple streets for value and so we can't bluff the abundance of air we do get here with.

We talked about splashing some A2s-A5s and 76s, maybe 98s into our UTG opens for better board coverage.

Basically, we can include/add hands that are ~ 0EV opens to our ranges and, sometimes, they will add EV to other parts of our range

Think how great it is to have 76s when you open UTG and go to a flop of T98 against a good LAG who flatted from the button.

Also imagine opening 7s6s UTG and catching a flop of AsKs6c. We just have so much leverage w combination of range advantage and equity here and finally a hand we can leverage our range advantage w.

This was basically the point I was trying to make only I somehow accidentally pasted something around the third paragraph when editing and turned it into nonsense.

Not that it was a quality post to begin with. I was tired and rushed to finish it before going to bed.

The chart was for 6-max. Vs an UTG open of 2.5x at FR 100NL, it recommends the BUT has a lower 3-bet frequency of 5.98%. It's mostly comprised of the same combos, but none of the suited gappers, and there are generally lower frequencies for the light 3-bets overall. One perhaps interesting note is that for full ring, AK becomes a flat on the button 8% of the time, and is a 3b/fold vs a pot-sized 4-bet. It doesn't fold AK pre at 6-max in that spot.
I believe they are the numbers just from Snowie's trials against "versions of itself" and "exploitative agents" that were used in the training period, so it's akin to a nemesis.
As an aside, I would love to know if the hands being imported via the new HEM Snowieapp are being added to the database (or being used for a new one), because I think a training tool that effectively datamined real hands from human play at specific stakes would be even more powerful/useful. (Snowie's "GTO-ish" ranges might look good in theory, but they don't necessarily work superbly well at microstakes, for instance, where a villain will rarely 4-bet bluff, and will often flat with hands that "should" be 4-bet according to Snowie).

I haven't been following PS stuff, is that an app for Snowie playing real games? Online? Live? Any win rates available? Sounds controversial.

Edit to add: I just realized I hijacked this thread, sorry, never mind.

There is a new SnowieApp now built into HEM, for people that subscribe to it, allowing users to get post-game analysis of their play. I have no idea what the makers of it are doing with all the hand histories that are uploaded to the cloud, but there is clearly a datamining angle that could be exploited for profit.
Players can't, however, use the SnowieApp to give them in-game advice (it doesn't have that capability) as that would clearly be cheating.
Thread in NVG: http://forumserver.twoplustwo.com/29...e-app-1586140/
/derail

While I've done some bad, -EV (though not hugely -EV, just a little bit) calls against regs I'd much rather make a bad raise than a bad call, for marketing purposes. The intention is usually to get more value in future hands and you you get more value by raising. Typical example would be 4-bet/folding a weak hand, with the idea that the next time you have a monster and the opponent 5-bet shoves AJ or something like that. If you call the 3bet with crap, the marketing usually doesn't even work, because you'll probably fold it before showdown, so the opponent will never know what you had.

abstraction/exploitability number? i dont know what probably decent means.

i know you're smart enough on the topic to keep a reflected and critical perspective, but honestly you just shouldn't be putting statements like these out here or anywhere.