Alternating Between Loose & Tight...Seamlessly

[ckboddic]

In some of the Cardrunners videos, Taylor Caby talks about alternating between loose and tight styles. He says what separates a great player from an okay player is his ability to play and understand both games, and also his ability to change between them. In another really good video Gaucho2121 says that one of the main reasons durrrr is so good is due to his ability to change from a fairly loose style to a tight style very seamlessly.

Obviously, playing a hard to predict style of poker is good for keeping your opponents on their toes, but I'm wondering why all the emphasis on changing between them seamlessly? Gaucho2121 made it sound like durrrr is really great because he's able to change gears without his opponents knowing. So does it all boil down to your opponents not knowing you switched styles so they'll make errors against you because of this? Is it that simple or am I missing something? If it really is that simple, how do you go about implementing seamless changing of gears into your game? What makes style alternating deceptive and hard for others to catch on to?

[Speel Posher]

i think that this thread has a lot of potential.

I am not sure if anybody knows if they switched gears seamlessly because it is in the eyes of his/her opponents.

I am working on switching gears myself as well. From my limited experience i would say that a good time to change gears is when you realize people are starting to call you down with less than they normally would. this is really vague, but i am definately no expert.

[mykey1961]

Lets say you play two different strategies.

L represents your strategy when you are playing loose, and T is when you're playing tight.

p would be the frequency you play loose, and 1-p would be the frequency you play tight.

A would an average strategy A = Lp + T(1-p)

Your opponents would attempt to play a stategy X that exploits A as much as possible.

The question is, how does L and T do vs X

It may be possible that L vs X gives L a +EV, and T vs X gives T a +EV

[ev_slave]

Good start to the theory mykey. But although your logic starts strongly, I feel that your analysis falls apart since your opponent is not playing rationally.

In Game Theoretic terms, you're playing a mixed strategy with the probabilities p and (1-p) of playing the two styles. A fundamental goal of mixed strategies is to reduce your opponent to being indifferent between the (in this example) two styles he could play (due to not knowing which strategy you're currently implementing). In your setup, you play a mixed strategy but require your opponent to play a pure strategy (X with probability 1). If he were rational, he would not develop a strategy X, but would choose his own probabilities, e and (1-e) to make you indifferent between playing L and T.

Now, this is all reliant on the ASSUMPTION of a rational opponent, against whom you can only implement a purely random strategy against to maximize EV instead of playing specifically to exploit his tendencies. In my experience (at low levels, admittedly) game theoretic considerations are a waste since most players have a/many fundamental leak(s).

BUT, I do agree that this could be a real strong thread, so let me contribute this. The concept of seamlessly changing gears IS important if you consider <font color="blue">the following model:

p(h) is the probability that you're playing T on a particular hand (indexed by h to show that it can change from hand to hand)

P(h) is your opponent's SUBJECTIVE PROBABILITY that you're playing tight on a particular hand (again, changes with each hand based on how you've been playing).

If you're opponent is dead-on, he can alter his strategy to match your play. Therefore, you're advantage (call it A) is:

A = a + b{(p(h) - P(h))^2}, where b&gt;0

if p(h) = P(h) you're at some minimum advantage "a" based mostly on skill (note: "a" doesn't have to be positive), but you achieve a higher advantage as the difference between p(h) and P(h) becomes greater. Thus, changing p(h) seamlessly (in a way that your opponent's P(h) doesn't change) will work to create a bigger theoretical advantage "A". Thus, your objective on each hand is to max A over p(h) subject to how P(h) alters with p(h). Seamless play lets you do this.
</font>
Sorry... long post. Although limited, I feel that the math is instructive, but please don't let the fact that the last few posts have been mathematical prevent you from sharing your experiences with changing gears. I feel this could be a helpful post to many.

[Perestroika]

Ev_Slave I thank you for sharing this post I have a few questions. What does a+b denote in your equation? Also as you stated -P(h) is a subjective quantity, what will the composition of this estimate look like? I think the closer we get to estimating this number the better we can analyze this concept.

[pzhon]

Switching gears is overrated.

When you switch between loose and tight, you end up playing weaker hands on average than you do when you maintain a constant style. If half of the time you call all-in with only AA, and the other half you call with QQ+, then relative to calling with KK+ you are sometimes folding KK instead of calling while you sometimes call with QQ instead of folding.

Switching gears may be important if each of your styles is bad, and your opponents will exploit this weakness, or if your opponents will react badly trying to adapt to apparently different styles. However, if you adopt a good constant style, you will sometimes appear to play tighter, and sometimes looser, just based on luck. In one orbit, I raised 5 times without lowering my standards, and my KK UTG was paid off well when someone played back at me with top pair. It would have been good to avoid raising with 88 UTG, since this had much less folding equity than normal after I raised 4 times.

[ev_slave]

[ QUOTE ]
Ev_Slave I thank you for sharing this post I have a few questions. What does a+b denote in your equation? Also as you stated -P(h) is a subjective quantity, what will the composition of this estimate look like? I think the closer we get to estimating this number the better we can analyze this concept.

[/ QUOTE ]

P(h) can't really be estimated unfortunately. But as Phzon described below, if you raise 4 times in a row, he'll probably think you're playing loosely, which will cause P(h) to decrease (he thinks you're less likely to be playing tight). In his example, this caused the magnitude of [p(h) - P(h)] to increase, which caused his opponent to make a bigger mistake calling his KK.

The actual important part of the equation that you can have some control over is [p(h) - P(h)]^2. As long as your p(h) and the opponent's P(h) is as far apart as possible, you are gaining advantage.


"a" and "b" aren't estimable either, unfortunately. "b" controls how fast your advantage grows as your opponent becomes further confused with your playing style, and "a" is how you and the opponent matchup if he knows exactly how tight (or loose) you're playing. If you're better than your opponent, a&gt;0 and if you're outmatched, a&lt;0.

But notice that the formula gives the possibility of you having an advantage over a "better" player (so your a&lt;0) as long as you can deceive him with how you're playing.

Since nothing can be estimated, the model only helps in trying to figure out why you would want to switch gears.

[Perestroika]

As a disclaimer I am a live player mainly so i cannot rely on hand tracking software for player analysis.

Now heres the concept this post made me think about...

We have two HU cash players, Player Z and Player 1 both who c/r bluff flops OOP 10% of the time.

Player Z- check raises 3 of the first 10 hands and then only 7 of the remaining 90.

Player 1- check raises 1 hand per every 10 hands.

How will their average opponent's react to subsequent raises?

[dfan]

Well the average opponent holding something like KJs might, after Z's 3rd check raise in a row, decide he must be full of it and re-raise, not something Z wants.

The same opponent is more likely to fold the same hand to Player 1's 3rd raise in 30 hands.

So if you are only talking about bluffing, you should space your bluffs out as far as possible.

But that is looking at bluffing in isolation. More often when you check raise you have the goods and are hoping for a call or re-raise. So a complete analysis of this would have to include both parts - the effect of a certain bluffing strategy on both opponent calling vs folding frequency when you are bluffing and opponent calling vs folding frequency when you have a hand. That makes my head hurt, so I'll leave it as "an exercise for the reader."

[mykey1961]

[ QUOTE ]
However, if you adopt a good constant style, you will sometimes appear to play tighter, and sometimes looser, just based on luck. In one orbit, I raised 5 times without lowering my standards, and my KK UTG was paid off well when someone played back at me with top pair. It would have been good to avoid raising with 88 UTG, since this had much less folding equity than normal after I raised 4 times.

[/ QUOTE ]

Or you could just get beat down like regular.

I have a 5 hand run of AA, AA, junk, AA, and KK.

I bet the 1st AA 5xBB preflop and got 1 caller from the SB, flop came all clubs and the SB pushed. No club for me so I folded.

Next AA I was in the BB, nobody even limped preflop.

Folded the junk in the SB

UTG bet AA 5xBB and got raised, pushed, and got called by AKs.. AA held up so I doubled up.

UTG+1 bet KK 5xBB got raised, I pushed, and got called by the big stack at the table. KK lost to AA.

4 excellent hands, and I was down to the felt.