Never Ending Math Equation

I've been quite the nerd lately.

First, I fooled around on excel for the first time since we did rat surgery analysis in Animal Phys. I wrote a program to calculate win rate, variance, standard deviation and ROI for SNGs, but have had trouble with figuring out how to calculate risk of ruin. Either way, if anyone wants it just lemme know (IM me, email, PM on +1/CP) and I can send it to you.

Today, I used the ProPokerTools tournalyzer to attempt to find an optimal solution (or rather, as close as I know how to get) to the following problem: On the bubble in a SNG, with all equal stack sizes of 10BB, what should the player in each position be moving in with?

I started by assuming that the first player jams ATC and that the other 3 players only call with aces.



Then maximized the equity for the second player. Then used these values and maximized equity for the third player, then the fourth. Then I changed the first players jamming range to be maximized, given the other players different calling ranges. I kept doing this over and over until the ranges didn't seem to change.



I think that's probably close.

For the record,

12% = 77+,A9s+,KTs+,QTs+,JTs,ATo+,KJo+
10% = 88+,A9s+,KTs+,QTs+,AJo+,KQo
8% = 88+,ATs+,KTs+,QJs,AJo+

The non-UTG ranges are actually much looser than I expected.

Of course, I had to run into a problem.

In equity terms, let's say UTG jams and it's to the SB to call.

Needed to Call Risk Reward
64.65% $25.00 $13.67

He's risking 25 in equity to win only another 13.67 extra. So he needs almost 2-1 on the call.

Given the propokertools analysis from before, this means that the 8% range should be a ~2-1 favorite over the 12% jamming range from UTG.

equity win tie pots won pots tied
Hand 0: 45.260% 42.72% 02.54% 10978562908 654077116.00 { 77+, A9s+, KTs+, QTs+, JTs, ATo+, KJo+ }
Hand 1: 54.740% 52.19% 02.54% 13414965900 654077116.00 { 88+, ATs+, KTs+, QJs, AJo+ }

Clearly, this isn't so. So now I'm troubled. I'm feeling like I did something wrong in my pro poker tools analysis. I'm ponder it some more later...

(Not) OK Computer

Signed onto Full Tilt and started three SNGs up. I played for a few minutes, treading water. I got kicked off. No big deal, loaded it up again and continued to play. Built up some chips in each of the SNGs over the next five or so minutes. Full Tilt then froze up. I closed it again and couldn't log back in. Fortunately, I'm not the only one this happened to, so hopefully I'll get a little refund. It was still frustrating though because I felt like I was playing well.

Signed up for 3 SNGs on Stars with the mindset to right all the preflop wrongs I've been making. I played really aggressive, and kicked my preflop aggression up a ton. I avoided some people and cold decks, cold decked one guy and managed to pull off two seconds and a first.


Confidence, as well as the bankroll, are beginning to increase steadily.
I have a med school interview in a week at West Virginia University. Should be interesting. I hope they don't quiz me on bio though, 'cause I'm a little rusty (to put it gently).

After that, I gotta drive home as fast as possible in order to make it to the Girl Talk show in time at the Grog Shop in Cleveland. Should definitely be a good one.

Finish Your Collapse and Stay For Breakfast

Life is hard. Poker isn't.

Hand #1

Full Tilt Poker Game #1669128727: $6 + $0.25 Heads Up Sit & Go (12235206), Table 1 - 15/30 - No Limit Hold'em - 23:52:14 ET - 2007/01/24
The button is in seat #2
*** HOLE CARDS ***
Dealt to theMFfulton [As Ac]
theMFfulton raises to 90
2s2e2a2n2 raises to 1,500, and is all in
theMFfulton calls 1,410, and is all in
2s2e2a2n2 shows [Ah 5s]
theMFfulton shows [As Ac]
*** FLOP *** [Th 2c 6h]
*** TURN *** [Th 2c 6h] [6s]
*** RIVER *** [Th 2c 6h 6s] [6d]
2s2e2a2n2 shows three of a kind, Sixes
theMFfulton shows a full house, Sixes full of Aces
theMFfulton wins the pot (3,000) with a full house, Sixes full of Aces
2s2e2a2n2 stands up
theMFfulton stands up
*** SUMMARY ***
Total pot 3,000 Rake 0
Board: [Th 2c 6h 6s 6d]
Seat 1: 2s2e2a2n2 (big blind) showed [Ah 5s] and lost with three of a kind, Sixes
Seat 2: theMFfulton (small blind) showed [As Ac] and won (3,000) with a full house, Sixes full of Aces

+1 PXF Tourney, 24 Songs and some Blog Changes

Last night there was a +1 tourney on Stars in which everyone gave lucko their hand histories and he loaded them into a Poker X Factor replayer so that we can see everyone's hands as they play.

I thought I played awful last night, but seeing everyone's hands (R.O.T.!) I don't think I played as bad as I thought.

One thing that stuck out immediately was that all of the good players stole and restole a LOT. Stoner pushed me around real bad. It gives me some things to work on, I suppose.

A few days ago on +1 I made a challenge to make a soundtrack to your lives that lasts 24 hours long. Pretty much everyone thought this was too much and that 24 songs was better. So I did this... eventually I'm sure I'll do the 24 hours long one; I have the songs in a list but not in order.

Anyways, here's 24 songs in the life of Nick Fulton



I've made song blog changes. I updated to the beta version and now it's being a huge fucking problem to log-in. Not sure why. I also figured out how to post videos on here, if I ever have the desire to do that.

A SNG Problem

Lucko in a recent post showed a series of hands he played at the end of a SNG. There was a hand where he jammed into a bunch of shorter stacks with 63o. He said he thought it was close, but +$EV. Stoner thinks that it was a bad move.

So I tried to figure it out. Take the following for granted, because a lot of it is probably wrong.

Here's the hand:

Seat 1: jbreuker (3,850)
Seat 2: lucko21 (4,950)
Seat 4: thomas a (920)
Seat 6: vengven (1,855)
Seat 8: NYpokerguy (1,925)
vengven posts the small blind of 200
NYpokerguy posts the big blind of 400
*** HOLE CARDS ***
Dealt to lucko21 [6d 3d]
jbreuker folds
lucko21 raises to 4,950, and is all in
thomas a folds
vengven folds
NYpokerguy folds
lucko21 wins the pot (1,000)

I typed everything into the ProPokerTools Tournalyzer calculator.

From the short stacks POV, lucko could be jamming any two. So if he calls with AA-22, AK-A2, KQ-K2, QJ-Q2, JT-J2, T9-T8, T7s-T3s, 98s, 97s, he maximizes his $EV, increasing his value from $81.40 if he folds to $90.40 if he calls.

If the SB knows the shortie is calling with this, he suffers the least (no matter what, he is always losing a little EV but this is the range he performs best at) calling (or overcalling, I suppose. Part of the reason this isn't exact is because going through overcalls would've taken forever) with AA-22, AK-A2, KQ.

If the BB knows the shortie and SB's ranges, he suffers the least with this range: AA-55, AK-A4.

Stack/blind posted/range/Post-hand value/Pre-hand value.
3850 0 $27.29 $26.30
4950 0 ** $30.03 $30.25
920 0 AA-22, AK-A2, KQ-K2, QJ-Q2, JT-J2, T9-T8, T7s-T3s, 98s, 97s $9.04 $8.14
1855 200 AA-22, AK-A2, KQ $16.42 $16.77
1925 400 AA-55, AK-A4 $17.22 $18.55

Okay, so if he knows this and is playing (close to, I suppose) optimally, lucko should reduce his range from ATC to 81% of the hands.




Stack Blind Hands Post-hand Value Pre-hand Value
3850 0 $27.34 $26.30
4950 0 81% $30.27 $30.25
920 0 AA-22, AK-A2, KQ-K2, QJ-Q2, JT-J2, T9-T8, T7s-T3s, 98s, 97s $8.78 $8.14
1855 200 AA-22, AK-A2, KQ $16.37 $16.77
1925 400 AA-55, AK-A4 $17.24 $18.55

This range of 81% of the hands does NOT include 63o, but it's really damn close: 22+, A2s+,K2s+,Q2s+, J2s+,T2s+,92s+,82s+,73s+,62s+,52s+,42s+, A2o+,K2o+,Q2o+, J3o+,T4o+, 95o+,85o+, 75o+,65o,54o

So you can keep going with this. The shortstack can now call with T2s, 96s and 87s to gain a little value (about 20 cents worth).

The SB can't add anything profitably but the BB can add A3s and A2s, KQ and KJ-KTs.

Of course now lucko has to tighten up a bit. Once he tightens his range to 76% of the hands(22+,A2s+,K2s+, Q2s+,J2s+,T2s+,92s+,83s+, 73s+,63s+,52s+,43s,A2o+,K2o+, Q2o+,J4o+, T6o+,96o+,85o+, 75o+,65o) he can start jamming profitably, although he performs better as his range tightens (although to what extent I'm not sure, because I'm tired and going to bed).

I hope I wake up in the morning, look over this and have it all make sense, because if not I'm gonna be pissed I spent so much goddamn time on it.

On a brief sidenote, I've been rollercoastering in the FTP games, in SNGs, HU SNGs and cash games, but played really well in the UB cash games today for a decent profit in a pretty short amount of time.

VP$IP and Table Talk: The Matusow/Smith Quandary

Enough about math; let's talk some psychology.

Those who have listened to The Circuit since its inception know that its first host, Mike Matusow, and it second group of hosts, Joe Sebok and Gavin Smith, professed very different styles of play.

Matusow preached (although I have no idea if he practiced this) a pretty tight style. He'd continually say you don't have to play a super-loose, super-aggressive style because most players were bad enough that you can just play solid, wait around for a good hand and trap.

Smith and Sebok (although Smith moreso) play a much looser style of poker. In internet poker terms, you could say that they play a LAG style.

So what is this a function of?

Let's look at some other players who play each style.

Tight (or perhaps even supertight):
Hellmuth

LAG:
Negreanu

So besides playing style, what do Matusow and Hellmuth share that Negreanu and Smith do not?

I think a large part of it is their table talk. Smith and Negreanu are both known for being jokers at the table. They both talk a lot, but they're pretty friendly. Their game plan, essentially, is to be everybody's buddy.

Then there's Matusow and Hellmuth. They whine, cry, talk shit. Basically, they make themselves targets. I would imagine that players like Tony G and Shawn Sheikan are also pretty solid players.

So the Matusow/Smith Quandary is this: the nicer you are to your opponents, the more hands you'll be able to play because they will let you steal more often and will gun for you less. The more of a prick you are, the more action you will get.

Of course, this effect is probably close to negligible against most players and for most people; I would imagine you'd have to pull this (whichever style you are attempting and are more comfortable with) on a pretty rank amateur and even then, you would have to be pretty excessive whichever route you opt to take.

SNGs

I hate playing SNGs for some reason. They just don't do it for me. That being said, I nearly busted my FTP account a couple weeks ago and planned to transfer from funds back there from UB. When Neteller shut down, this couldn't be done, so I've been playing $5 SNGs on there to replenish it the old-fashioned way.

There's one thing about SNGs that I absolutely love (besides the fact that they're fairly mechanical-- which generally I hate but it allows me to four or five table them and the variance is low) is the end game when the blinds are ridiculously high. You can open-jam every hand and people will still lay down to you. Yesterday, I was even getting comments from people like "you got somewhere to be?" but they'd still fold way too often. It looks like a stupid, maniacal play, but if they keep mucking and playing way too weakly, there's almost no risk at all in doing it.

Not much else exciting going on. I have been focusing on not being so lazy with my open-raise sizing (i.e., making smaller raises from the front and bigger raises on the button and in the cutoff) instead of just hitting 'bet pot' but it's a real minor difference.

I won't get into the open-raising thing too much because it's been done before by a million people, but here's a good link:

Chris Ferguson

Phil Gordon's Little Green Book also covers this concept a fair amount.

Chipp has also broken it down for me pretty well in an old thread called "varying opening raise sizes".

That's all for now, hopefully something exciting'll happen soon.

Hey Neteller....

...fuck you.

Poker is goooot

For the first time in a few weeks I feel like I played well today. Of course I made mistakes, but I've felt pretty unconfident lately. I don't think I've been reading people's hands well and paying off way too much. It's been pretty frustrating considering how much work I've been putting into studying the game.

In the car today, I was thinking about the book Word Freak by Stefan Fatsis.

Very interesting read about the world of competitive Scrabble players. There's a part of the book where Stefan is talking with Joe Edley, a three time National Scrabble Champ. After leaving Edley's, Stefan is encouraging himself as he's struggling with becoming more experienced in the game. He keeps telling himself "Make the best play. Winning is a byproduct."

Poker's no different.

Here was a fun hand I played today:

Ultimate Bet
No Limit Holdem Ring game
Blinds: $0.50./$1.
5 players
Converter


Stack sizes:
Hero: $96.70
CO: $99.90
Button: $132.30
SB: $119.10
BB: $146


Pre-flop: (5 players) Hero is UTG with
Hero raises to $3, 2 folds, SB calls, BB folds.


Flop: ($7, 2 players)
SB checks, Hero bets $9, SB raises to $34, Hero raises all-in $93.7, SB calls.


Turn: ($194.4, 1 player + 1 all-in - Main pot: $194.4)



River: ($194.4, 1 player + 1 all-in - Main pot: $194.4)



Results:
Final pot: $194.4
Hero shows Kc Ts


The HH won't show it for some reason, but my OPP ended up having AJ.

Anyways, I raised in the hijack with KTo. This seems standard, as does the raise amount.

The flop comes down VERY coordinated. Any time there are two broadway cards, there's a good shot someone has a hand. Three broadway cards makes this even more dynamic, and the two hearts just adds to that fact.

From what Chen/Ankenman have been saying, since the board is so dynamic, I can get away with making a bigger bet. In retrospect, due to my stack size being ~13x the pot-size, I should probably be betting closer to the pot-size here ($7) due to the three-street geometric growth considerations. In other words, I bet 7 here, he calls, pot is $21 and I have 84 left. Next street I bet 21, he calls, pot is 63 and I have exactly enough for a pot-sized bet left.

Oh well. Live and learn. Plus, going through this work afterwords means that I can gain from my mistakes without having to pay for them.

Anyway, my thought process was to overbet the pot. It looks a skosh weak and it's likely that he has some kind of hand that he may either try to value-raise with (like a set, two pair hand like he had, AK) or may try to make a big semi-bluff with a KQ type hand or pair + flush draw.

He raises and there's enough in the pot that it can be argued I could take one off here and jam a blank turn in order to punish flush draws that miss on the turn. But I opted to jam. Honestly, given 2-1, even if the K or T of hearts comes and counterfeits my straight, I'm likely gonna pay off even if he hits. So if I always pay off, I'm losing some money from my opponent when I jam a blank turn and the flush draw is forced to lay down.

Additionally, if a K or T comes, it might kill some action from other made hands (like the type of hand he had). Given the hand that he had, a Q is also a pretty shitty card that most players are going to muck to.

So I jam, he ponders for a while and makes the crying call. SHIP IT!

Chapter 21, Part 2

Okay, more about The Mathematics of Poker.

After mentioning the situations I mentioned in my last post, Chen/Ankenman mention check-raising. Because of their fold equity, weak draws gain from check-raising and can even go from -EV (if the CR wasn't available) to +EV. Now if you just CRed with weak draws, you'd be incredibly exploitable. Fortunately, there are two other types of hands you'll want to (occasionally) CR with. One is big hands; the reason for this is simple-- pot building. As players like Gavin Smith are famous for saying, you want to play big pots with big hands and although is statement encourages very exploitable play, the point here is that when you have a strong hand and you can build a big pot by check-raising, do it. The next type of hand in this range is the strong draw (i.e., open-ended straight flush draw). The reason for this is if you get reraised, you can just jam, as drawing hands always benefit from getting the money in early anyway (and very often you'll be a favorite in the hand).

So, according to the authors, check-raising with some weak draws, some strong hands and strong draws gives one a balanced strategy in this situation.

This leaves some medium strength hands, some strong hands and some weak draws to check-call with. This group of hands is also balanced because due to the strong hands in the mix, your opponent can't exploit you by increasing or decreasing his betting frequency.

Now it's time to look at bet-sizing and how it correlates with the board texture (my favorite concept in NLHE).

First, consider when there are two cards of one suit on the flop. Obviously, the threat of a flush draw is possible, so that should be added to the possible distribution of hands. (They also make note that this draw is "open" or, as Paul Phillips says, it is transparent information.)

Similarly, when the board is very static (the example they give is KK4r) you should bet geometrically such that the pot will grow in a predictable way on each street (specifically, in a way that a certain percentage of the pot is bet on each street such that the bet on the river gets all of the chips in)-- this bet can range anywhere from 1/3 the pot to roughly pot-sized. On this type of board, small bets are small enough to discourage any underdogs to try to draw out (because the hands will have only 2-5 outs).

They note that specifically "[o]n static boards, we never overbet the pot, preferring to take advantage of the static nature of the board to provide value to our distribution" (p. 275).

On a dynamic board, however, this is not true. In instances where we have a made hand and our opponent is drawing on a dynamic board, by the river he can play somewhat clairvoyantly because when he misses his draw, he will never pay off a river bet, whereas the reverse isn't necessarily true. Because of this, we benefit more from betting large in early streets in order to offset the disadvantage we have later (again, due to the partial clairvoyance the drawing hand obtains).

So on a semi-dynamic board (i.e., Kx7h2h) where we'd geometrically bet 1/2 the pot, they suggeste we bet slightly (maybe 50%) more, so roughly 3/4 the pot.

On incredibly dynamic boards (987 with a flush draw possible) we should bet even more. In fact, as opposed to situations where we attempt to bet geometrically so our stack is all in on the river bet (a three-street geometric growth) we may attempt to bet such that we're all in by the turn (or two-street geometric growth).

Of course, due to betting larger amounts, that means we should also be betting a smaller frequency of hands. As such, we should be checking a decent amount of strong hands in these situations. Of course, checking to give free cards can be pretty dangerous on dynamics boards such as this one. Therefore, the authors suggest that you check with strong hands that won't be hurt by free cards (i.e., on a Kc7c8c board, checking with AcKx). Doing so can also gain value when your opponents (semi-)bluff.

The authors then go into turn play and river play which I won't go into now. I may in the future. Again, I would definitely recommend picking up this book to any serious poker player. I feel like it has definitely already helped my game immensely and I'm not even finished with the book yet.

Chapter 21, Part 1

I've found what I believe to be the most important chapter written about NLHE for a player who has a decent amount of experience. Chapter 21 of Mathematics of Poker-- A Case Study: Using Game Theory. I won't go into the first half, which is regarding PF play and open-jamming. It's really interesting, but the second half of the chapter talks a lot about what I've easily spent the most time focusing on in my own game; PoF play when taking into account various board textures.

I've spent hours thinking and analyzing these types of situations just because there is such little info on it in most poker literature. Now it's out there for everyone to read (which helps me, but hurts my edge against those who read it; luckily for me, the math approach isn't a popular one so it's unlikely to become highly used).

For those who don't see me in person often, I like to highlight any book I read. My copy of MOP is easily the second-most marked up book I've ever owned, including textbooks (#1 being Catcher in the Rye). This whole chapter is almost entirely underlined and bracketed off.

The Flop Play section begins with mentioning the differenc ebetween static boards (few/no draws, i.e., K72r or AAA) and dynamic (draw heavy boards, like 9d8d7s).

They then mention that in situations where there is a PF raise from EP and a call out of the blind, the blind is more likely to have a weaker hand due to the fact that he's getting a discount on entering the pot. Because his range is so much weaker, he can basically autocheck to the PFR who will in turn usually have an autobet. However, the autocheck can be exploited by the raiser checking behind, but in general autobetting has a higher EV than attempting to exploit the check because his raising distrubution is strong enough that he's gaining value from calls (with his strong hands) and winning the pot immediately when the blind folds.

From the blind's POV, most hands-- weak and strong-- we can check and then have the option of folding (our weak hands) and calling/raising with the rest of our distribution. The authors then breakdown the different possibilities as a result of a checkraise from our blind.

1) If the opponent is calling with his strong hands (as opposed to reraising), our strong hands and our medium-strength hands are losing value. However, our draws gain in value because we're not forced to pay more money or fold.

2) If our opponent is calling with weak hands, the opposite is true. Our strong/medium hands are getting good value, but our draws lose value.

3) If our opponent is reraising here, our strong hands get a ton of value (potentially maximum value) whereas our draws are losing maximum value (when we're either forced to call off or forced to fold).

The task then is to balance our actions over the entire range of potential hands. I'll go into this in the next update.

some more poker, MOP

Haven't been playing a hell of a lot recently. Ran bad/played bad tonight but only ended up stuck like half a buy-in.

I noticed something strange on UB today. I've spent most of my poker playing the past month or so on FTP because of UB's stupid upgrade that stopped letting you save hand histories (or rather, more than four at a time; they fixed it but at the end of my session a few hours ago it seemed like the problem was back. Fucking ridiculous.) Anyway, I played on UB for the first time in a while. The play was SO fucking horrible.

Take for instance this hand:

http://www.pokerhand.org/?716101

Hand #37040068-96710 at Downey (No Limit Hold'em)
Started at 05/Jan/07 06:30:32

themffulton is at seat 0 with $173.40.
BigBruin007 is at seat 1 with $201.65 (sitting out).
eje63 is at seat 2 with $278.50.
Matthe is at seat 3 with $25.85.
laxlax is at seat 4 with $124.95.
conjenrick is at seat 5 with $91.60.
The button is at seat 0.

eje63 posts the small blind of $.50.
Matthe posts the big blind of $1.

themffulton: Jd Kc
eje63: -- --
Matthe: -- --
laxlax: -- --
conjenrick: -- --

Pre-flop:

laxlax folds. conjenrick folds. themffulton raises
to $3.50. eje63 calls. Matthe folds.

Flop (board: 4d 6h 3c):

eje63 checks. themffulton bets $4. eje63 calls.

Turn (board: 4d 6h 3c Kh):

eje63 checks. themffulton checks.

River (board: 4d 6h 3c Kh 6c):

eje63 bets $16. themffulton raises to $32. eje63
calls.



Showdown:

themffulton shows Jd Kc.
themffulton has Jd Kc 6h Kh 6c: two pair, kings and sixes.
eje63 mucks cards.
(eje63 has Ah 3h.)


Hand #37040068-96710 Summary:

$2 is raked from a pot of $80.
themffulton wins $78 with two pair, kings and sixes.

The value-raise on the river looks thin as hell, right? Well, it was a misclick. I meant to just flat-call. Maybe not making raises like this against these kinds of players has been costing me a little. You're really, really unlikely to be up against a bluff here, because so few hands are putting in another bet here for value you're going to see even fewer people trying to push you out.

But again, I digress. This wasn't even atypical for the dozen or so people I was playing against (3 tabling). It was all sick, maniacal play. People were betting pot-sized on every street, raising and reraising so many hands preflop and calling down with shit like ace high on a regular basis. Granted, people who play in the .5/1 game aren't usually the greatest players. In fact, most of them are idiots. But this was above and beyond some of the things I had seen and couldn't figure out why it was so prevalent.

Then I read the name of a guy up 7 buy ins at my table: PrtyPokersGay. Oh yeah. All the fish there had to migrate, and they've found me at UB. Good.

Getting further in Mathematics of Poker and am reaching a point where I understand it less and less. I've started to actually ignore decent amount of the math because it's way too over my head and instead just try to glean some information from the results that they find. It makes me feel like I'm cheating on a test or something, but even without understanding a lot of what Chen/Ankennman are talking about, it is still easily one of the best poker books ever written.

Since Christmas, I've had this fancy new laptop that works so hopefully I'll be able to play a lot more in the future. Work lately has put me on a really fucked up schedule (it's 8 a.m. and am planning on being awake for another 12 hours; and I didn't go to bed last night) so it makes it hard because I've been trying to avoid playing while tired, because I think that's when I'm the most likely to play with a "fuck it, let's gamble" mindset. I also started playing guitar again with some friends which has taken up a lot of my post-midnight time which is when I was previously playing the bulk of my poker.