Tuesday, February 12, 2008

Why Do I Suck With AA?

So it is a long-burning question in my mind why I suck so badly with AA so I thought I'd throw some math at it. Big surprise, I know.

My winrate with AA is simply pathetic - I believe it is 1.6 BB / hand when most of the results I've seen from other players are slightly north of 3 BB / hand. Let's assume, then, that my deficiency is about 1.5 BB every time I've been dealt it. How significant is this? Well, I've been dealt AA 400 times, so I'm down 600 BB from where I "should" be if I earned what others earn. This is a non-trivial percentage of my entire earn for two years. Even if I "just" look at my Cake hands (softer site, and I should be better at poker than when I started 2 years ago), my AA suckage means a deficit of 200 BB which is 20% of my 1000 BB total - this alone would boost my winrate by 0.7 BB/100. Yikes!

There are many options for the source of this leak:

  • Not letting AA go when beaten.
  • Losing more than necessary with AA when beaten.
  • Not getting value from AA when ahead.
The second option is kind-of a subset of the first one, but the first option speaks to not folding AA enough, and the second one is about showing it down for more bets than I should have shown it down for. Make sense?

Well, the first option is the one I want to tackle in this post - how often do I lay down AA, and how often "should" I lay it down? It turns out I have only laid down AA 5 times out of the 400 times I've been dealt it (once on the flop, three times on the turn, once on the river). I'm not even convinced that the river fold was correct (I folded top set on the river to a single bet heads-up on a 4-flush board). So I fold 5 of 400 or 1.25% of the time. Is this good? Bad? What is the "reference"?

Well I checked in with my poker "coach" and had him look at his 150K+ hand database. Out of his 711 AA hands he folded 32 times (once on the flop, 15 times on the turn, 16 times on the river) for a 4.5% "fold rate". So, based on this one data point I am simply not laying them down enough.

But wait - we can determine how big an "error" this is in my case. If I "need" to fold an extra 3.25% of the time, that means I showed down 13 AA hands that I "shouldn't" have. However, figure for each of these hands I lost an extra 1-2 BB / hand (either a river bet I shouldn't have called, or a turn and river bet I shouldn't have called. So at most I lost 26 BB more than I should over those hands. But my overall "deficit" with AA is 600 BB !!!!

The conclusion I can draw from this is that although I am very obviously showing down AA more than I should, it is not even close to the biggest leak I must have with AA - it explains less than 5% of my deficit. This calculation at least jibes with some platitude that I remember (from a 2+2 or Ed Miller article or something) that said that if you showed down every AA hand you got at limit poker, it would be at most a small error.

So, it looks like it is off to the database to look for hands where I've either not gotten value (most likely, knowing my own tendencies), or paid too much to show down. It is only 400 hands - how hard could it be to go through them all?

I'll leave this post with the one AA hand I folded on the flop - within a month or two of me starting to play the game online:

** Dealing down cards **
Dealt to Hero [ Ad Ac ]
BigFreeze101 folds.
ukvixen calls [$0.15].
Hope06 calls [$0.15].
bwhit86 calls [$0.15].
Mcblunt111 folds.
rainbowpants folds.
KGssd1 calls [$0.15].
Hero raises [$0.30].
Pandaro39 folds.
Rennwurm folds.
ukvixen calls [$0.15].
Hope06 calls [$0.15].
bwhit86 calls [$0.15].
KGssd1 calls [$0.15].
** Dealing Flop ** [ Jd, Jh, 6s ]
ukvixen bets [$0.15].
Hope06 raises [$0.30].
bwhit86 calls [$0.30].
KGssd1 calls all-In.
Hero calls [$0.30].
ukvixen raises [$0.30].
Hope06 raises [$0.30].
bwhit86 calls [$0.30].
Hero folds.
ukvixen calls [$0.15].
** Dealing Turn ** [ 7c ]
ukvixen bets [$0.30].
Hope06 calls [$0.30].
bwhit86 folds.
** Dealing River ** [ Tc ]
ukvixen bets [$0.30].
Hope06 calls [$0.30].
ukvixen shows [ 2d, Jc ] three of a kind, jacks.
Hope06 shows [ Js, Ks ] three of a kind, jacks.
KGssd1 doesn't show [ Qc, Qh ] two pairs, queens and jacks.
Hope06 wins $2.91 from side pot #1 with three of a kind, jacks with king kicker.
Hope06 wins $1.92 from the main pot with three of a kind, jacks with king kicker.

Palm Pilot Update

15,326 hands, just short of 4,000 BB = 26 BB/100

Shortest post ever.

Monday, February 11, 2008

What Does -371 BB Look Like?

As a brief followup to my previous post, here is what a 371 BB downswing (at non-Cake sites) looks like:



Up about 744BB in the first 52K hands for 1.43 BB/100, then down 371 BB (almost exactly half of my BB winnings to date) in 6K hands for an amazing -6.2 BB/100. Since I was playing at a (much) higher level at the end, the $$$ results of these 58K hands hands is actually in the red by about $100. Thank God for bonuses, and thank God for the Cake network.

Thursday, February 07, 2008

Two Years and Counting

Given that I played my first hand of online poker for money on February 7, 2006, it looks like I've been doing this for 2 years. So what better time to do some comparison between the two years? My apologies for the fomatting (Blogger does some weird things with tables).














Most BB won, day
Year 1Year 2
  • 4/8/2006: 70.6 BB

  • 6/17/2006: 63.8 BB

  • 5/5/2006: 53.1 BB

  • 1/5/2008: 63.1 BB

  • 9/2/2007: 53.3 BB

  • 1/18/2007: 51.7 BB



Looks like these are similarly sized.














Most BB lost, day
Year 1Year 2
  • 7/21/2006: -53.8 BB

  • 10/31/2006: -47.7 BB

  • 6/21/2006: -47.4 BB

  • 3/18/2007: -60.0 BB

  • 10/20/2007: -55.1 BB

  • 6/17/2007: -47.6 BB



Again, similarly sized. Looks like just a bad day on 3/18? Think again - look a few tables down....














Most BB won, week
Year 1Year 2
  • 5/1/2006 - 5/7/2006: 94.7 BB

  • 9/4/2006 - 9/10/2006: 78.4 BB

  • 12/18/2006 - 12/24/2006: 78.0 BB

  • 12/10/2007 - 12/16/2007: 112.3 BB

  • 10/29/2007 - 11/4/2007: 93.3 BB

  • 7/16/2007 - 7/22/2007: 76.1 BB



Nice week in late December, but in general these are similarly sized.














Most BB lost, week
Year 1Year 2
  • 3/27/2006 - 4/2/2006: -84.0 BB

  • 5/15/2006 - 5/21/2006: -55.8 BB

  • 1/22/2007 - 1/28/2007: -38.9 BB

  • 3/12/2007 - 3/18/2007: -98.0 BB

  • 3/5/2007 - 3/11/2007: -95.6 BB

  • 6/11/2007 - 6/17/2007: -76.2 BB



Here is where the real nastiness comes in. My first year, I only had two weeks where I lost even more than 40 BB. This year, I have some huge weeks of loss. And you'll notice the two consecutive weeks in March 2007 that caused me to take a break from the game for two months. Nearly 200 BB in 2 weeks. And if you "cherry pick" dates, the total size of the downswing before I quit was 282 BB starting on February 11, 2007 (just past my first year anniversary!) and ending on those two weeks above.

And for a truly sobering number, my total downswing since February 11, 2007 is 371 BB on non-Cake-network sites, since I did play a few days in June on Stars and FTP (and got killed LDO). I should just withdraw all my money from those sites so I am not even tempted to go play there again....














Most $$$ won, day
Year 1Year 2
  • 2/3/2007: $92.00

  • 6/17/2006: $81.00

  • 7/22/2006: $74.25

  • 1/5/2008: $380.90

  • 1/18/2008: $235.91

  • 11/10/2007: $181.75



Obviously bigger than last year. Best day is far and away better than my second-best day, and more than double my third place day.














Most $$$ lost, day
Year 1Year 2
  • 10/31/2006: -$107.35

  • 12/15/2006: -$103.25

  • 1/24/2007: -$85.00

  • 1/13/2008: -$241.50

  • 10/20/2007: -$212.30

  • 11/15/2007: -$203.00




Three days of more than -$200? Wow.














Most $$$ won, week
Year 1Year 2
  • 1/29/2007 - 2/4/2007: $177.00

  • 12/18/2006 - 12/24/2006: $146.75

  • 10/23/2006 - 10/29/2006: $98.85

  • 10/29/2007 - 11/4/2007: $447.31

  • 12/31/2007 - 1/6/2008: $335.48

  • 11/5/2007 - 11/11/2007: $295.95



Nice run from 10/29 to 11/11: nearly $750.














Most $$$ lost, week
Year 1Year 2
  • 1/22/2007 - 1/28/2007: -$166.75

  • 12/11/2006 - 12/17/2006: -$157.25

  • 3/27/2006 - 4/2/2006: -$85.87

  • 3/5/2007 - 3/11/2007: -$298.00

  • 1/21/2008 - 1/27/2008: -$225.75

  • 3/12/2007 - 3/18/2007: -$196.00



Note 2 of these 3 weeks were during the slide that had me giving up the game for a bit. Cherry picking the dates as before, the total slide starting on February 11 was -$923.40.

Well, that's about it. Back to the tables!