Tuesday, March 27, 2007

High Variance vs. Low Variance Playing Styles in NL Holdem Cash Games

I don't think too much has been written on this topic, though it has been touched upon by various authors to some extent. I will do my best to explain my thoughts on the matter, and through comments and follow-up posts, I may be able to make some sense out of what I am trying to explain here.

First of all, I play a pretty low variance form of NL Holdem cash. It was not always this way, but my game has evolved to where variance has been pretty much minimized. This is a good thing in some ways, and bad in others. Playing a low variance style minimizes the swings in bankroll, and keeps you from having lengthy losing streaks that can mess with your confidence. This is a huge benefit. The negative in playing this way, is that it is not the way to fully maximize profit at the tables. I could make more per hour by allowing more variance into my game, but as a result I would have to deal with the bankroll swings that come along for the ride. My game has been criticized many a time, because of my ubertightness. I usually shake this off, because I know that the way I play works and is solidly profitable. The one criticism that has stayed with me, though came from F-train a while back when reviewing some statistics that I had posted. I will summarize what he said (and try to put his exact words here later). He basically said that my post flop aggression was off the chart, and that because of this I was leaving money on the table. I would need to learn how to go deeper into the board (past the flop), and because of this learn how to make some huge laydowns if I wanted to further improve my game. He has a very valid point here, but would it really be worth it?

So before I start lets just talk about the basics of variance. Some charts would be helpful here, but I will try to explain in a way the mathematically inclined will understand. Variance in it's simplest form could be represented by a Sine wave. It goes up and down around break-even at some amplitude, but over time the area under the curve goes to zero (overtime variance washes away). In the real world variance is random and chaotic, but still sums to zero overtime like a simple Sine wave does. A high variance game would have high amplitude in this signal. A low variance game would have low amplitude. The variance signal is then overlaid on your bankroll growth (Profit/Loss) signal. Imagine this as a straight line that slopes up if you are a winning player, or slopes down if you are a losing player. The rate at which you win or lose determines the slope of the line. So your overall bankroll growth is a combination of the variance signal and profit/loss signal. In general, the variance signal has a much, much higher amplitude and tends to drown out the profit/loss signal over short periods of time. Only over long periods of time (when variance washes away), can you see the underlying profit/loss trend.

For a winning player, variance is what makes it possible to lose. Imagine if the pot went to the person mathematically ahead when the money went in (no suckouts allowed), and when hand ranges came in to play the winner got their mathematical equity in the pot. Variance would be pretty much eliminated and winning players would win all sessions that they played in. Losing players would lose all sessions that they played in. Variance is what allows losing players to have winning sessions. Because of the way the human mind works, the losing player selectively remembers the winning sessions, and feels he is better that he really is. Variance keeps the fish coming back so it is a good thing. But, if you are a winning player, does a high level of variance really help. Possibly not. If you can minimize the variance, you can maximize the percentage of sessions that you win, and as a result potentially maximize your enjoyment and confidence in the game. For a losing player the opposite is true. Maximizing variance maximizes their chances that a session will be a winner. Wild play is actually better for a losing player given that the other lower variance choice is neutral on profit/loss. A losing player has no reason to play a low variance style. A winning player does in a way. They can use their profitability to outweigh a low variance signal and obtain much more consistent results.

So what makes one style high variance and one style low variance? Before I get into that, I will talk about how to recognize what type of style you or your opponents are playing. I have played cash games for a long time. I typically do not run my stack north of 2x my buy-in. I rarely "rack-out" (get stacked) in cash games. So I am minimizing my losses and minimizing my wins at the same time. I am attenuating the variance signal in some way. When my good friend smokkee started dabbling in cash games (when he had only been playing for several months), I would see him routinely sitting with 3x, 4x, 5x the buy-in. Levels that I have never got to in a cash game, he was regularly achieving. Of course he was getting stacked left and right as well. He was playing a very high variance form of cash game. I would put lucko, wes, Fuel55 and doubleas in the same category. They play a high variance (and profitable) cash game style. So the signs in general, are how easy and often to you get up to 3x or 4x your buy-in in a cash game, and how often are you reloading your chips. Most people play a high variance style. Nearly all new NL cash game players play a high variance style. NL is much higher variance than limit (obviously).

So now I will try to explain the aspects of a low variance playing style below:

1) Play very tight preflop. The TAG style is much lower variance than the LAG style. TAGs play fewer hands and get into fewer unique situations as a result. LAGs play more hands and push slimmer edges post flop as a result, increasing the variance.

2) Multitable lower limits vs. single table higher limits . If you play 3 or 4 tables at a time, your variance is reduced because you are getting much more hands in to wash it away. If you get stacked on one table, your winnings from the others will offset this. If you play a single table at a higher level for the same profit potential, variance is magnified by 3x to 4x vs. multitabling.

3) Don't limp from early with hands that you normally will not call a raise with. Limp early with small to medium pairs, but not with medium suited connectors. You can usually call the raise with the pair, but not with the suited connectors. Sometimes you will limp with suited connectors, get raised, get some callers, and feel you are priced in for a flop. This is a very high variance play, as it is pretty unlikely you will hit the flop hard. You might also lose a huge pot to a higher flush when you do hit. Suited connectors in general are a high variance starting hand.

4) Don't call big raises with middle pocket pairs preflop. It is ok to call a reasonable raise with a small pair preflop (3x-4x). Be careful when it is raised more than this. Use the rule of 10 (or better yet rule of 15) to see if your opponent's stack, and yours are big enough to "set mine". Avoid the borderline situations by folding, and look for situations that have great implied odds. Playing a pocket pair for a big raise with implied odds barely there is high variance.

5) Don't raise small to medium pocket pairs from early or middle position (full ring). It is best to limp/call these hands to minimize variance. Also if you don't hit your set, don't get too crazy post flop. Folding to pressure when you miss your set minimizes variance. Trying to play on with overs on board when you have a middle pair is high variance. You will remember the massive pots that you won when raising pairs and hitting a set, or calling a big raise and hitting, but the most profit is made when you limp multiway and hit over the long term.

6) Be aggressive on the flop. Take your stabs at the pot early, and find out where you are at early before the pot gets too big. Checking, betting, raising, and folding on the flop are low variance. Calling the flop is high variance (in general).

7) Don't be overly aggressive without a made hand on later streets. Double and triple barreled bluffs are high variance. Try more single barreled bluffs, and less of the double and triple barreled types to minimize variance.

8) Don't make heroic calls on the river. When you call the river in NL you should be pretty sure you are ahead. Ace high calls on the river are high variance. Look for a better spot. Call when you pretty much know you have it, or are at least priced in.

9) Don't get cute with big pairs preflop. If you look down at AA or KK preflop, and there are a lot of limpers, raise it up pretty big and take it down right there, or at least get heads-up. If someone raises before you act, reraise them enough that you take away there set mining odds, and push the flop if they call. Don't limp ever with these hands preflop if you want to minimize variance. Make a big raise, and eliminate the competition.

10) Learn how to fold TT-QQ and AK preflop. There are times when these hands are obviously beat. Fold them in these situations and look for a better spot. Going to war preflop with these hands is a very high variance play.

11) Don't slow play or float (too much). Slow playing and floating are both high variance plays. Betting your good hands, and folding your junk is low variance. There are times when these plays are correct (for profit reasons), but they will always increase your variance.

12) Semibluff your draws (especially when in position). Draws in general are high variance. If you can control the betting on draws you can lower the variance. Semibluff your draws, and try to take free cards in position, and take pots down uncontested before your draw hits. Manage the pot size when drawing so that most of it goes in after you hit to minimize variance. Calling with draws all the way down is high variance.

13) Learn to make huge laydowns. Probably the most difficult situation you can get in is when you have a good hand, and are facing a ton of pressure. You may need to fold a set, or a straight, or a flush, when it is pretty clear that you are beat. If you can make these folds your variance will be reduced.

14) Choose tables with short stacks present. If you are playing at a table with all full buy-ins, variance is maximized. As the mix goes towards shorter stacks variance (proportional to potential pot sizes) is reduced.

15) Play full ring vs. short handed. Full ring cash games allow you to exploit bigger edges with bigger hands. Short handed the blinds will eat you up if apply this strategy. You have to push smaller edges with more marginal hands short handed, so your variance will be higher accordingly.

16) Play in position. When you are making moves/playing hands in position you can keep your variance lower by having better control of the pot size and betting. When out of position, the variance goes way up because others can exploit your positional disadvantage, and manipulate the pot size beyond your control.

This is what I have so far and they are not in any particular order. I am sure there are many other ways that I have missed. Let me know if you can think of anything, or if you think I am way off in what I am saying. I would say that most NL cash game players should go through stages. Initially they will play high variance. Then their variance will come down a bit as they become better. The better players will allow more variance back in their games if it means more profit. I am not sure if I am ready to do this yet, but I will be a little stuck in my win rates until I am.

Labels: , , ,

9 Comments:

At 1:23 PM, Blogger Fuel55 said...

nmjNice article and I generally agree with you with the following comments that are most applicable as the limits get higher:

Point 1: Play very tight preflop. The TAG style is much lower variance than the LAG style. TAGs play few hands and get into fewer unique situations as a result. LAGs play more hands and push slimmer edges post flop as a result, increasing the variance.

While in principle this sounds good in bigger games people figure out that you are playing 88+, AT+, KQ and little else they will mine all sort of trash on your ass. Also with the big pairs you win small pots often and fairly regularly lose monster pots to sets/overpairs etc.

4) Don't call big raises with middle pocket pairs preflop. It is ok to call a reasonable raise with a small pair preflop (3x-4x). Be careful when it is raised more than this. Use the rule of 10 (or better yet rule of 15) to see if your opponents stack, and yours are big enough to "set mine". Avoid the borderline situations by folding, and look for situations that have great implied odds. Playing a pocket pair for a big raise with implied odds barely there is high variance.

You have to call raises when both stacks are 10x the raise or bigger. Set mining is where most of your profit comes from. It also shows that you aren’t a folder to reraising.

5) Don't raise small to medium pocket pairs from early or middle position (full ring). It is best to limp/call these hands to minimize variance. Also if you don't hit your set, don't get too crazy post flop. Folding to pressure when you miss your set minimizes variance. Trying to play on with overs on board when you have a middle pair is high variance. You will remember the massive pots that you won when raising pairs and hitting a set, or calling a big raise and hitting, but the most profit is made when you limp multiway and hit over the long term.

Disagree. Limping is for pussies. If you are playing a hand – RAISE.

Otherwise I pretty much agree.

 
At 3:54 PM, Blogger Blinders said...

Fuel,

Lets just say that these suggestions are for minimizing variance, but not necessarily maximizing profit. I already said that I don't maximize profit playing this way. I would argue that many of these things maximize profit and minimize variance at the same time, while others will lower your profit rate to some extent if you make the lower variance adjustment I am suggesting. What's left are probably profit neutral choices.

 
At 12:37 AM, Blogger smokkee said...

very good post.

i agree your style is low variance. but, it's also lower profit long term.

i don't like limping with pocket pairs 77 or higher in EP. i'll raise, c-bet regardless of flop and hope to take it down. most times, that will do it. if you get called or reraised on a scary board, fold or proceed cautiously if you didn't flop a set. i really don't like limping with pocket pairs 77 or higher.

also, playing more hands gives you the image of a LAG. but, if you're playing most of your hands in LP with a bunch of limpers, it's an opportunity to take advantage of your position in the hand. most players will insta-fold to aggression to a player in LP. they're not limping with monsters in a cash game a majority of the time.

it's a good way to pick up small pots and build a LAG image. that's when you're monsters can get paid off.

 
At 9:02 AM, Blogger Miami Don said...

Lot's of good points.

Although as you go up in limits this style makes it much harder to get paid as its too transparent to very good players. So much of poker is about reads, if ones style is plainly predictible, it will get exploited over time.

 
At 9:35 AM, Blogger NewinNov said...

Nice post and good comments.

 
At 1:10 PM, Blogger WindBreak247 said...

(A Month Late)

This is a simply phenominal post for a developing TAG like me. While I already abide by a lot of these guidelines, seeing them on paper and in detail is a great refresher, and I will keep this post within a quick reach whenever I feel the cold grip of high variance play.

 
At 1:06 PM, Blogger ChrisXena said...

Two points. First I disagree witht he point in in #5 about not raise with medium pairs, and just limping (in effect treating them as a drawing hand to trips+). These have only 2 realistic outs and a a drawing hand is a lousy number to draw with, against likely straight and flush draws which also beat trips. So, also as Fuel said, the better option is to raise with them. This serves to help eliminate many of those superior and more likely to hit drawing hands. Instead you want to go against possible unmade pairs only. If you can eliminate half the number of players going to the flop, you have increased your odds substantially of winning when you fire yet another barrel at the pot, or if unmade trips going on to the River.

And surprisingl this same thinking goes to even made higher pairs, in order to reduce variance, and many times KK and AA are beaten on the flop by two pair or tips, or outdrawn further on. So aggression actually reduces variance in the shorter term, although for single bets you may have one-hand variance go up(very short-term variance) if someone happens to have a much better hand and calls your bet.

2. Secondly, all the talk of LAG vs TAG with the idea of table image so you don't get trashed when your playing style becomes so apparent as a super tight player. So a player can disguise his true self, and create a false image. I do this by playing like a LAG player and first and then sutley shift into a more TAG mode, then reinforce the fase image by playing a slightly less tight hand occassionally, and letting others know this if possible. One guy used to play 7-2s all the time and flashed them and complain about bad luck, and his "action" got higher on his usual TAG hands.

So control your table image, and disguise how tight of player you are, and thus increase your action with mostly low variance play.

Ms ChrisXena
Poker Princess Warrior

 
At 2:03 AM, Blogger vladb said...

Very good article, enjoyable read! Regarding point #5, I would disagree and argue that you should still never limp pre in mid position with a hand you intend on playing. For example, if you hold a hand like 77 in mid position, raising sets you up to win a bigger pot if you hit your set. It also eliminates trashy hands competing against you that might hit a lowly flush to defeat your set for example.

 
At 12:14 PM, Blogger Jim Philips said...

it is quite tricky to explain it but you did a great job with it. I will take your recommendation on online bookmaker

 

Post a Comment

<< Home