handhistories

There's been a lot of buzz recently about handhistories (isildur,BT,BH etc) so I've decided I should try something similar. after all it sounds like a good business plan:

1. collect handhistories

2. ?

3. profit

after browsing the web for some time I found an archive of 5million hands played on ipoker@50nl 6max this autumn (sept-oct '2009 probably). I dont play on ipoker and I know nothing about players there but I assume there's no much difference between ipoker and FT or PS (and also I comply with their TOS..). I've created a separate database and now I'm importing these hands. It seems that it may take forever - it's been exporting for almost 24hrs so far yet it's imported about 2.7 mil hands. import speed dropped down dramaticaly from 70-80hands/s in the beginning to 33hands/now so I may stop import some time tomorrow.

there's an interesting too that comes with holdem manager which is called 'holdem vision'. basicaly it profiles players at given limit(lag, tag, nits etc) and shows how different stats(VPIP, PFR, AF etc) correlate with their winrates. It also shows ranges in different flop and preflop spots, depending on actions(limpers/raiser/facing 3bet etc), board textures and number of players. I may play with this database on weekends and may write some sort of report here so if anyone has any specific questions post them in comments.

what's the use? well, for example there's an interesting post on 2p2 (link) on facong flop raise. cliffnotes from article that goes with holdem manager:

Post Summary

• take notes: Always important, but top pair would be easy to get away from if villain only min-raises with sets and you have no other draws. (Now I just got to find the time to enter all this data into my player notes).
• Not best hand: It's probably safe to say that when min-raised, your top pair hand is behind 75% of the time, or will be by the river without any draws. Maybe a little less than that for your overpair.
• No Fold Equity: 99% of the time the min-raiser will go on to see a showdown, so don't try to bluff him/her out unless you have a good read/notes.
• Great Implied Odds: This does give you GREAT implied odds if you have a good draw.

Now this sounds extremely important as I probably lose alot of money in situations like this. I'd also like to check balugawhale's "theorem" (fold when someone raises your turn bet) and examine winning players' tendencies.

ok, so for now some diagrams (based on 2.67mil hands and filtered only players that have min 1000 hands tracked)

breakeven range is +/- 5 bb/100

VPIP

15-20% vpip - 25% of players are winners

20-25% vpip - 33% of players are winners

25-30% vpip - 31% of players are winners

30-35% vpip - 31% of players are winners

35-40% vpip - 27% of players are winners

PFR

10-15% pfr - 25% of players are winners

15-20% pfr - 31% of players are winners

20-25% pfr - 36% of players are winners

25-30% pfr - 32% of players are winners

Post Flop Aggression

AF :: % of winners

1 :: 22

2 :: 27

3 :: 30

4 :: 31

5 :: 24

>15 :: 27

3-bet

2-4% 3bet - 21% of players are winners

4-6% 3bet - 27% of players are winners

6-8% 3bet - 34% of players are winners

8-10% 3bet - 37% of players are winners

10-12% 3bet - 21% of players are winners

12-14% 3bet - 32% of players are winners

WTSD%

15-20% wtsd - 29% of players are winners

20-25% wtsd - 31% of players are winners

25-30% wtsd - 26% of players are winners

30-35% wtsd - 29% of players are winners

35-40% wtsd - 16% of players are winners

40-45% wtsd - 24% of players are winners

so stats for most winning players:

vpip - 20-25%

pfr - 15-20%

af - 2

3bet - 4-6%

wtsd - 25-30%

another blitz

i've had problems with my ISP today so the session was only about 15 minutes. but! graph:


i know poker gods owe me $22(i was $22 below EV) but still my recent sessions look ridiculous:

Limit	Hands	Hands Won	Sessions	Sessions Won	Hours	Amount Won	BB/100	$/100	BB/Hour	$/Hour	Rake	MGR
$0.02 NL	809	142 (17.55%)	58	36 (62.07%)	1.38	$24.87	76.85	$3.07	449.46	$17.98	$3.85	$1.42

stats:

Limit	VP$IP	PFR	W$WSF	WTSD	W$SD	AF	AFq	3Bet	Fold 3B	Att To Steal	Fold BB to Steal	Fold SB to Steal
$0.02 NL	25.09	20.15	51.63	35.77	52.27	2.36	55.56	10.90	54.17	40.43	66.67	60.00

looks standard for me : 25/20 vpip/pfr. my 3bet is usually ~12% AFq~2.6, WTSD~31
but $17.98/hr playing 2nl was awesome. i want to replicate my success :)

preflop analysis

i tried not to change my strategy during my 10k test, 5k hands look more consistent(with average~10PTBB/100 win), while 5k hands look random. so here are my stats for whole sample(~10k hands), hands ~1-5k and ~5k-10k:

	10k	<5000	>5000
Vpip	25.29	26.53	24.66
pfr	15.86	16.83	14.9
W$wsf	44.79	44.1	45.48
Wtsd	26.79	26.37	27.2
W$sd	42.63	44.1	43
Af	2.65	2.65	2.65
Afq	53.67	42.25	53.79
3bet	6.38	5.62	7.19
fold3bet	57.41	58.59	56.07
att to steal	28.35	27.13	29.34
fold bb to steal	75	85	66.67
fold sb to steal	88	89.47	87.1

i started playing a little bit tighter and started 3betting more and calling 3bets more. the difference between vpip and pfr is quite big but it's for a reason. for example, i used to always open-raise, but at this level open-limping is still profitable (Harrington gives some examples when open-limping is acceptable in HoC). here's how i open-limp:

so...open-limp is profitable. oh, and i only limp in with speculative hands and i always make 3-5BB bet with premium hands.

my open range is pretty loose, however with limping in i may play even more hands.
here's how 26% range looks like:

here's how i play these hands.
hands 1-5k (-$11.12):


hands 5k-10k(+$26.53):

huge difference. i still have to find why. starting hands are the same, strategy is pretty much the same (according to stats). maybe it's pareto principle :) my 1st guess it's due to aggression(red line, drastic changes in AFq). my 2nd guess is even though my vpip% is the same, the actual cards i played are quite different. here's how i played cards that are outside this pokerstove range when i wasn't on the button or BB/SB.

hands 1-5k(-$8.82):

hands 5k-10k(+$2.08)

total number of hands outside 26% range that i played without position is pretty much the same however i played them more aggressively..or folded them earlier.