Sunday, September 28, 2008

Phase 2

I upgraded my PC a few days ago, which actually also was quite funny to do (well, not the assembly, but the buying of the parts I needed). I walked into a computer store, across the place I work at myself, and asked them if they had the Q9300 processor. They answered no, so I thanked them for their time and walked outside. So I went into the next store, where they answered 'yes' to my question, and followed by asking me if that was all I needed. So I answered I also needed an asus mainboard but wanted to see their boxes (and made my pick after seeing them choosing the only one with the P45 chipset), 4GB of ram, an LG or Nec DVD burner, a SATA harddisk of 250gb (have enough external drives lying around) and a 550w power supply from Coolermaster or Antec.

Somewhere along the line they asked me what I was going to use the PC for, I believe it was when I informed them that the memory they were trying to sell me wasn't much better than the memory which was only 1/3rd of it's price (which was the only thing I bought where I even bothered to inform about the price, which also probably surprised them). I still remember the looks on their faces when I answered 'Poker' to their question.

So after finally bothering to put my PC together a few days later and reinstalling everything I decided it would be interesting to start with a complete clean database, and even not import my .dat file from Full Tilt which contains all the notes I have about players.

The reason for doing so was something I realized after my trip to Barcelona. I played a few days online and ran/played pretty bad, losing about 6 buyins or something. Nothing out of the ordinary, but I was getting less and less satisfied with my overall game. Further, my notes on players were cluttered with examples about hands that were way too specific, yet failed to describe the kind of players they were.

My NL2 winnings were something about 3.5ptBB/100 or something longterm, and that is not something to brag at, especially when I consider myself a decent player who invests a lot of time doing his homework, and who does not play too many tables.

I've been doing an extensive review of my old pokertracker database the last few days and came to the conclusion I missed out on quite some important things while playing my normal game, which was something that used to be a large part of my game before: Pure mathematics.

Here are a few examples to describe what I mean:

$1/$2 No Limit Holdem
6 players
Converted at

UTG billyk38139 ($406.00)
UTG+1 YRRLYZL ($237.00)
CO blinnd ($336.10)
BTN Hero ($216.85)
SB ChungBuck87 ($311.50)
BB Klaus47 ($54.25)

Pre-flop: ($3, 6 players) Hero is BTN

2 folds, blinnd raises to $7, Hero raises to $24, 2 folds, blinnd folds

Final Pot: $17

Hero wins $34 ( won +$10 )
blinnd lost -$7.00

Villain in this spot is a solid regular. He plays 23/19 attempt to steal 36% with 28% attempt to steal from cutoff. His fold to 3bet is 61%, but this is obviously weighted based on position, history etc. His continuation bet % is only 61%.

So here, even while it does not look that bad to just call, I do not think I can exploit him that much with his low cbet %, and I may get in a tough spot with an easily dominated hand like KJo.

My old standard was to polarize my range here vs players like these with QQ+, AK and suited one-gappers which exclude a T. That part of my range is still the same vs him, but KJo has some great preflop bonuses in this spot:

If he opens with 28% of his hands, and even with the history between us which makes me believe he will defend somewhat lighter vs my 3bets, I still only expect him to continue with hands like 99-AA, AJ+, KQ and JQs/JTs. That is 8.1% of the total hands which he could be opening with, and it's about 29% of his total CO opening range. Assuming I will get owned badly if called, or won't put a chip in postflop, when seen as a total bluff this needs to succeed 24/34=70,6% of the time while he will generally fold 71% of that range. Even while that seems miniscule +EV, there are still two people sitting behind me who may wake up with aces.

Yet if I add the card removal bonuses for having KJo here, his range shifts. Before he would be continuing with 29% of his total CO range, but if I calculate the odds for him to hold a hand he can continue with when I hold blockers to some of those hands, things shift drastically.

So what seems a likely 28% cutoff opening range? Assume something like this, which is actually 28,5%: 22+,A2s+,K8s+,Q8s+,J8s+,T8s+,98s,A7o+,K9o+,Q9o+,J9o+

That leads to 13 pocket pairs (78 combinations), A2s+(48 c.), K8s+(20), Q8s+(16), T8s+(8), 98s(4), A7o+(112), K9o+(64), Q9o+(48), J9o+(32).

This is not 100% correct off course, since I do not know his complete range, but it's a close estimate. Further, it's much more likely that hands like 56s are in his CO opening range than K9o when I am sitting behind him, but I'll leave it at this range for now.

So what this tells us is that he opens 430 different combinations, and he can continue with 29% of that range if I had no blockers, which are about 125 combinations. Me holding KJ gives me some blockers on his hand distribution if I look at what hands he'll continue with, which I described as 99-AA, AJ+, KQ and JQs/JTs. Holding KJ reduces the probability of him holding AJ+ with 8 combinations, KQ with 5, JQs/JTs with 2, 99+ with 6.

This sums up that while normally he could continue with 125 of his possible combinations, he can now only continue with 104. Me holding KJ has a significant impact on the chance I get called here, and while it's incorrect to say he will now only continue with 24% of his hands and will fold 76% of the time (since me having two blockers to hands he will continue with also reduces the chance he opened from that spot in the first place) it does have some noticeable impact.

This is an extremely marginal situation, but in this spot, combined with his low cbet % and some lunatics sitting behind me who could do something stupid like squeeze with A7o even without knowing how to spell that word, I figured 3betting was better in that spot than calling.

Hand 2:

$1/$2 No Limit Holdem
6 players
Converted at

UTG FIVETHREE53 ($82.40)
UTG+1 zer_cool ($227.75)
CO Hero ($230.70)
BTN Neffer2 ($76.20)
SB clopper lover ($110.00)
BB basouli ($438.75)

Pre-flop: ($3, 6 players) Hero is CO

2 folds, Hero raises to $7, Neffer2 calls $7, 1 fold, basouli raises to $29, Hero calls $22, Neffer2 folds

Villain in this hand played quite sporadically. He ran at 40/31 with a 3bet % of 8.8%, and I already made a note about him in earlier observed hands which made me think he was extremely squeezy-happy. So right here, I again have to fix a little mathematical dilemma.

I probably open about 25 or 30% from the CO on this table, and this player is going to 3bet tons and is definitely going to use most squeezy opportunities. If I fold almost any hand vs him except a widened value based range, he is going to exploit me tremendously preflop. This is why I need to continue with about half the hands I open from this spot to not be exploitable preflop, either by calling or 4betting for value/as bluff. 9Ts in position is definitely playable, so I call.

($66, 2 players)
basouli bets $45, Hero calls $45

In a normal raised pot I rarely float with just a gutshot and backdoor flushdraw and prefer a bluffraise (though on this board texture, I prefer a float), but in a 3bet pot the stack sizes become a bit too small to raise here. Further, I cannot represent much if I raise since I would not raise a hand like AJ either. His large cbet makes me quite comfortable about my options later on, so I decide to float here.

($156, 2 players)
basouli checks, Hero bets $63, basouli folds

Standard after floating flop. vs tighter players I would take the free card, but the chance of him having a made hand is so small that I prefer to take the pot down on the turn, cheaply.

Final Pot:

Hero wins $216 ( won +$79 )
Neffer2 lost -$7.00
basouli lost -$74.00

Hand 3:
$1/$2 No Limit Holdem
5 players
Converted at

UTG Big0308 ($541.60)
CO Hero ($401.10)
BTN blinnd ($197.00)
SB sorry_its_me ($121.75)
BB easyway777 ($352.45)

Pre-flop: ($3, 5 players) Hero is CO

Big0308 raises to $7, Hero calls $7, blinnd raises to $31, 2 folds, Big0308 folds, Hero raises to $66, blinnd calls $35

A 35/17 fish opened UTG, so I just flatted, since I do not really like raising with AQo in this spot much. It becomes more interesting though when someone who plays 23/19 with a 3bet % of 7.5% decides to squeeze. Yet again this person made some plays earlier which made me think that he pounds on preflop weakness a lot, so I assumed he was quite capable of squeezing here. I do not mind a call, but I figured AQ was ahead of his range, and since I will be playing AA/KK/QQ/AK the same in this preflop situation, I put in a tiny 4bet. Villain obliged by calling.

($142, 2 players)
Hero checks, blinnd checks

I hated this flop kinda, since even while I rarely ever cbet in a 4bet pot OOP on dry boards, I do not expect to get much action. I took the same line with the rest of the hand as I would with KK in this spot.

($142, 2 players)
Hero bets $69, blinnd folds

Final Pot: $142

Hero wins $208 ( won +$73 )
blinnd lost -$66.00
Big0308 lost -$7.00

Hand 4:

This is a hand I misplayed terribly, but I do not mind my play if I had something like KJ or A9o, since then I would at least have some blockers.

$1/$2 No Limit Holdem
6 players
Converted at

UTG shxlfxm ($281.80)
UTG+1 doktore4 ($158.55)
CO 2michael8 ($173.60)
BTN Hero ($207.85)
SB Bule Gila ($691.40)
BB Boodabor ($219.00)

Pre-flop: ($3, 6 players) Hero is BTN

1 fold, doktore4 calls $2, 1 fold, Hero raises to $10, Bule Gila raises to $28, 1 fold, doktore4 folds, Hero raises to $52, Bule Gila goes all-in $691.40, Hero folds

Villain in this hand is quite solid, played 21/16 after 1.2k hands, and had been owning me constantly at this table. Him sitting there would be a reason to actually leave the table, but since I want to get ready for 2/4, I do not mind some tough opposition at times.

What was interesting is that the villain in this spot 3bet me, while he generally used to just call my isolation raises in position when a fish was in the pot, even with hands like QQ+.

Because of that, I had to assume that he was probably more likely to be making a total bluff, compared to having an actual hand, since I did expect him to just call with all his playable and premium hands in this spot to keep the fish in, especially in position.

Sadly, I was wrong, and he told me after the hand that I made a good laydown, claiming he had AA, which translates into him saying he had 99 or something and just leveled me...

Final Pot:

Bule Gila wins $747.40 ( won +$56 )
doktore4 lost -$2.00
Hero lost -$52.00

Hand 5:

This has nothing to do with the point I'm trying to get across, but it was an interesting hand vs the same villain as in hand 2, and I am quite happy about the way I played it.

$1/$2 No Limit Holdem
6 players
Converted at

UTG Neffer2 ($233.45)
UTG+1 clopper lover ($171.60)
CO basouli ($269.40)
BTN FIVETHREE53 ($35.70)
SB zer_cool ($195.00)
BB Hero ($388.80)

Pre-flop: ($3, 6 players) Hero is BB

2 folds, basouli raises to $7, 2 folds, Hero calls $5

Again the 40/31 player opens, 52% steal, and I hold crappy-connectors in the big blind. Generally this is a fold, but slightly over 100BB deep vs a bad player who may be tilting a bit after the last hands, I could not find the fold button.

($15, 2 players)
Hero checks, basouli bets $12, Hero raises to $32, basouli raises to $102, Hero calls $70

I flop the nuts or cooler, and vs people who don't cbet too bloody often I just lead here. Vs him, a CR is more optimal.

After his 3bet on the flop, I consider my options, and I decide that if I get my money in now I will get it in as a decent favorite, but if I just lead the turn when it's safe, I'll probably have better equity overall. So I stick to that plan.

($219, 2 players)
Hero bets $61, basouli goes all-in $160.40, Hero calls $99.40

River: ($539.80, 2 players)

With the history between us, I expect him to shove a turn lead extremely light, and I manage to get the rest of my money in pretty well. Obviously a shove would accomplish the same vs his set, but I do not think he would play a bare Ad or even Kd that different vs me.

Final Pot:
Hero shows:
basouli shows:

Hero wins $536.80 ( won +$267.40 )
basouli lost -$269.40

Overall, even while it does not show that much from these examples, I am going to to focus a bit more on mathematical edges when I can find them, and will probably tighten up quite a bit in spots which are extremely marginal vs non-regulars. Though vs regulars, I will probably still continue to do a bunch of crazy stuff for 2+ buyins when I think it's breakeven+, but that's good for metagame and image anyways.

Saturday, September 27, 2008

EPT Barcelona

Like the standard has been for me lately, it has been a while between updates again.
Only this time I do think I have a somewhat valid excuse, which is having spent 2 weeks in Spain granting myself a well deserved vacation.
Should any observant reader notice that the timespan between this entry and my most recent one is more than 2 weeks, please keep it to yourself since I do not need someone to tell me how lazy I am. Which yet again is proven by how little of the entry about my vacation is somewhat interesting to add to my blog, while the largest part of the vacation was spent at the local swimming pool with a nice cold beer.

Anyways, back on topic, so somewhere during that vacation me and the buddy who went with me decided to go to Barcelona. I lost the count about how often his father must have told him to not go to the casino/be careful at the casino/don't gamble at all as if he somehow thought the two of us were some gambling degenerates who only wanted to go to Barcelona to visit the casino there.

He couldn't be more right, but we did make some pictures with my phone from other random stuff so we could show him we did not spend all of our trip to Barcelona in the casino. One picture even got me in it, and even if it's a silly picture of me sitting in front of a pretty sad looking but tasteful desert, you have to take into account that these are the kind of pictures you want to show to people who think you are a degenerate. It screams "See, I do go outside at times." while it leaves away information like that particular restaurant being only about 500 feet away from the casino.

When we entered the casino, which was very early in the evening, it was already packed. Reason for that being that the European Poker Tour was taking place there, and with the amount of players left in the tournament, the final table would probably be starting that evening. Perfect day to play some cash games vs some EPT busters.

Our plan was somewhat disturbed because one of the differences between online and live poker: waiting lists. Online, you get onto a few lists, get a coffee, and get back in time to click sit down on a few tables. Yet live, during the EPT, you have to wait 4 bloody hours before you can claim your seat.

So after finally getting my seat at an NL1k euro table, it did not take that long to bust. I found myself within the first half hour after sitting down in a situation where I held two overs + nut flush draw in a multiway pot getting about 3 to 1 pot odds, snapcalled failing to hit my outs vs two pair and some random crap hand I do not even remember, instead of dragging in a nice 2600 euro pot. That's poker, but I am quite happy that I snapcalled there instead of bothering to consider the amount of money involved.

After that I decided to watch the 'bubble' before the final table, which is quite an experience to see from the tribune compared to some online broadcast.

After the final table had started I noticed Kara Scott, the EPT hostess, from her accent on the Cardrunner's videos, when she interviewed someone in the tribune. So when I saw her in the press area and she was unoccupied, I decided to chat a bit with her.

It did look quite foolish though when she pointed out to me that as a Cardrunners member I should be recognizing the person who was sitting at the table I was standing at, who was Brian Townsend himself.

It was quite fun to run into Brian. After introducing myself to him and talking to him a bit, he seemed a genuinely spontaneous and nice guy, as you'd expect him to be from his videos/interviews etc. I did not bother him for long though, but since I was a tourist that day, I did ask for a picture for my blog. That was no problem at all, and Kara was kind enough to also join the picture. Sadly it was a bit out of focus though thanks to taking it with a cellphone.

I left the casino somewhere around 2am, even while the action on the final table was getting quite hot and only a few players were left. Me and my buddy, who is just as poker degenerate as yours truly, had the dilemma to either watch an EPT final table from the tribune, or go out to party in Barcelona throughout the rest of the night. And after going out till like 6 or 7am I can definitely vouch that the nightlife of Barcelona is quite a blast.

Monday, September 01, 2008

August Results

Last month I posted my winnings and bragged about it being my first 3k+ month (which actually is incorrect if I include donkaments, but when looking at cash games, it still is. I nearly doubled that this month when I include rakeback. The stats above are from the 30th of July till the 1st of September.

But what was more interesting in this sample is the amount of huge adjustments I made to my game, and how much I have progressed as a player in about a month time because of the play at NL200 vs NL100. The winrate isn't that impressive, but that is mostly my own fault. The hands I misplayed this month taught me quite some things about beating the game as did the hands I won, and those I observed.

What stands out most to me are the discrepancies between the regulars at NL100 and NL200.

At NL100 there are some good/decent/bad regulars who are working hard on increasing their games, generally not playing too many tables.

At NL200 the average regular is much better. Their play is much better on average and they are used to countering other regulars in certain ways.

But what is almost mind-boggling is that almost any regular which I mark as good or very good are just playing way too many tables, which means that even while they may be better than me, they lose their edge because they turn into hud-bots and adjust in a somewhat predictable matter.

One example I could take are solid regulars on the button. They open a wide range, some of them even over 60%, and when they get 3bet by the BB they just look at BB's stats. In my example, it would show something like 22/18 to them with a 7% 3bet. The first conclusion they will make is that I will be 3betting light often enough, and each regular has their own way to adjust.

Some of them will start 4betting light, or 4bet bluffing often. The frequency they 4bet with will give you different options to exploit them. One way is to 3bet with a value-based range which does well vs their range in an all-in situation assuming you get enough folds after 5bet shoving, and expand that range based on how often they fold.

Another way to adjust is by not 3betting light at all. Most of these players will just be looking at your 3bet % stat, and if it's high enough, they make an automated decision.

One key factor from these observations is that this group of regulars will NOT adjust their opening range at all. If they open A3o from the button, and the BB is an aggressive player who defends his blinds quite often vs an aggressive button, they will still not consider folding A3o the next orbit.

Here's one example from today about this type of regular:

$1/$2 No Limit Holdem
5 players
Converted at

UTG ($200.00)
CO ($487.40)
BTN ($520.85)
SB ($92.00)
Hero ($200.00)

Pre-flop: ($3, 5 players) Hero is BB

2 folds, BTN raises to $6, SB calls $5, Hero raises to $28, BTN raises to $56, SB folds, Hero goes all-in $200, BTN folds

Final Pot: $118

Hero wins $262 ( won +$62 )
BTN lost -$56.00
SB lost -$6.00

Now, while this situation looks bloody standard, and this seems like a squeeze spot, the interesting part is how easy some other hands can be played vs this type of opponent:

With 1.1k hands on him he plays 23/17, attempt to steal 36% with 48% steal from button. So far, he folds to 3bets 56% of the time and 4bets 19% of the time.

If his 4bet % was not so high I would just call here to keep the fish in, and play fit/fold, and I think that would actually be more profitable in this spot. But assume for the moment that the fish in the SB had not called, what hands could I profitably 3bet here?

Definitely not a normal value bet range or a very polarized range. So 3betting KQ or air is suicide, so is 3betting any hand if you intend to fold to a 4bet. A better range in a button vs blind situation vs this player would be hands like ATs, AJs, AQ+ and any pocket pair including deuces with the intention to shove over a 4bet. Villain will be folding enough of his 4betting range there for that range to be very profitable vs him.

Now there are also players with a wide opening range who just fold to 3bets way too often OOP yet do not adjust their opening ranges much. Here are two examples:

$1/$2 No Limit Holdem
6 players
Converted at

UTG ($265.60)
UTG+1 ($239.80)
CO ($197.00)
Hero ($206.85)
SB ($121.85)
BB ($220.75)

Pre-flop: ($3, 6 players) Hero is BTN

1 fold, UTG+1 raises to $7, 1 fold, Hero raises to $26, 2 folds, UTG+1 folds

Final Pot: $17

Hero wins $36 ( won +$10 )
UTG+1 lost -$7.00

This regular is someone I really respect, and I do not want him sitting to my left at all. His postflop skills are very solid, and he will be getting the best of me quite often. But preflop, in position, I can make a lot from him if I adjust well.

What do we know about villain? He plays 26/23 after 4.6k hands, attempt to steal 41%, and he opens 26% from his hands from MP. His fold to 3bet % is 80%, 4bet% 7%. Further, he cbets 95% of all flop, which is way too high yet he succeeds because people just don't adjust well to him.

So what kind of hands would I 3bet vs him? All kinds of offsuit connectors, small pocket pairs (I cannot setmine profitably vs his range, and floating with unimproved small pairs is something I really hate vs a solid player), suited hands with a few gaps like 58s, the lower suited connectors etc. I still want some postflop possibility because I do expect him to just call a 3bet OOP with hands like AQ and JJ/TT.

But what would I flatcall with? All premium hands and hands which play very well in position, like mid-pairs, medium and high suited connectors, and all premium hands.

This may sound exploitable but it's not. If villain cbets so bloody often and folds to many 3bets, there's little value in 3betting with premium hands. And villain plays way too many tables to notice I won't 3bet premium hands vs him in position. Should he start to notice, and readjust, I will notice it after a while and respond appropriately.

Here's another regular who I play almost the same way against in position:

$1/$2 No Limit Holdem
5 players
Converted at

UTG ($234.90)
CO ($206.35)
BTN ($207.00)
SB ($282.25)
Hero ($240.00)

Pre-flop: ($3, 5 players) Hero is BB

3 folds, SB raises to $6, Hero raises to $20, SB folds

Final Pot: $12

Hero wins $26 ( won +$6 )
SB lost -$6.00

I have 1.7k hands on villain which show him playing 24/18, with an attempt to steal of 45% from the small blind, 83% cbet and a fold to 3bet % of 90%. This player is very positional aware but just fails at adjusting to who is sitting to his left. Opening too many hands from the SB works well for him on average, which is why he does it, but when he's playing 16 tables he will take an incredibly standard line here when 3bet: which is only continueing with JJ+ and AK.

If you look at a short sample, the optimal 3betting range vs him is all junk, and optimal calling range all premium hands. There is only one problem with raising all junk, which is that he may start to notice it when I 3bet him almost every hand. Though I still 3bet him almost every time, and the reason I do flat him often with hands like T7s etc instead of 3betting is for the same reason: I do not want him to adjust to me, and those hands I do not mind playing in position, even while when looking at it in a vacuum, 3betting T7s has better EV than calling with it.

Another thing I notice is that the regulars adjust well to cbets compared to those at NL100. They have balanced their ranges for floating, bluffraising, semi-bluffing, valueraising etc. What this means if you play a straightforward style vs those regulars when you happen to end up OOP vs them, they got an edge.

What I am trying to do to counter that is to have a much more balanced checking range. What this means is that I cbet much less than at NL100, but checking does not automatically mean that I'll fold. I may CR some boards as bluff/semi-bluff instead of cbetting them, may check-call for value and to induce bluffs, and may CR turns on drawy board because most regulars are just incapable of checking behind a turn when they hold a draw (they hate to put their cards face up).

Here is one example of me balancing my range a bit:

$1/$2 No Limit Holdem
6 players
Converted at

Hero ($602.40)
UTG+1 ($209.80)
CO ($200.00)
BTN ($428.40)
SB ($282.65)
BB ($208.95)

Pre-flop: ($3, 6 players) Hero is UTG

Hero raises to $7, UTG+1 calls $7, 4 folds

The button here is a very solid regular if you look at his stats, but not someone I really mind having to my left because he's a bit spewy. So far I got him at 20/16, 8% 3bet, 55% fold to cbet. Playing straightforward vs him is going to absolutely kill my winrate.

($17, 2 players)
Hero checks, UTG+1 bets $13, Hero calls $13

As I mentioned earlier, I don't cbet that often. And especially vs a player like this, I balance my checking range in a way to let me exploit his tendencies more than he can gain from positional advantage.

Standard play here is off course to make a c-bet. I obviously do not want to give a free card on a board like this, and want to get value from draws.

But how will villain respond to a cbet here?

He will peel at least once with any heart + overcard combo he has and overpairs, will raise nut flush draw, made flushes and sets and will fold all other holdings like bare overcards.

So betting here gets some value out of his range, and getting it allin here on this board will be marginally profitable only.

But how will he respond to checking?

First of all, I expect him to bet his complete range when checked to because it looks like I am check-folding. I do check-fold on this kind of board vs him with hands like bare overcards all the time, but I would rarely if ever cbet here.

So if I bet here he can play his hand correctly vs me, if I check, he won't, because he will never get a free card. If I checkraise, I only get action from hands which will make a relatively small mathematically mistake.

($43, 2 players)
Hero checks, UTG+1 bets $34, Hero raises to $104, UTG+1 goes all-in $189.80, Hero calls $85.80

Now we get to the point where it becomes interesting. Villain gets the chance to make a small mistake only if I lead here and he calls, and he gets the chance to fold his air.

But how do I expect him to act should I check here again even with a somewhat scary card on the board?

I expect this villain to absolutely never check a flushdraw behind, because he just does not want to put his cards face-up. He wants to abuse folding equity, and likes to balance his ranges, and that is fine. So vs villains like him, when it's very likely he is on a draw, giving a free card is no risk at all, since he will never take it. So my best way to get money in on the turn vs him when he could be drawing is by checkraising.

($422.60, 2 players)

Final Pot: $422.60
Hero shows:
UTG+1 shows:

UTG+1 wins $419.60 ( won +$209.80 )
Hero lost -$209.80

Off course he did suck out, but that's fine. I did manage to put in the bulk of the money with 82% equity, while a more straightforward line would have got me to put my money in with only 58%.