Most of it is shitty win one, lose one zigzagginess. None of it seems to bear any relation to whether my opponents were any good or not, even the fools who check every button won their fair share. One of them today made an awesome laydown though;

Admittedly shoving that river seems remarkably spewy since there aren’t many hands that would call, and he was also extremely passive – but that passivity was so extreme that I was convinced he really really liked his hand (I know turn bet was small, but *any* turn bet after he called was big strength from this guy). He tanked for quite a while before folding. I dunno if I would have had that strength…

That game was annoying in the end because I had him pinned down to < 500 chips for ages and he kept on clinging on, winning some flips back to equal stacks and then my 66 lost to his AK.

Here's another hand that shocked me. Different player, this time someone who raised their buttons and you'd expect more regular lines. Maybe flat calling a 3bet pre with aces is a much better move that 4betting or shoving?

Anyway, to the point of the title of this post – essentially I had two heaters this year, one at $10 games and the other at $30. The money from the $30s is blown away now, and my wins show I’m back to the point after the 4man $10s heater. It’s a pretty solid decline for the last 3 months. Those heaters raised my expectations, and it’s painful facing the reality that I’m not really winning at heads up.

I wish I knew where the holes are in my game. It was a lot more fun when I was winning…

Despite the lack of play though I’ve been really piling on the learning & evaluation side of things though – that downswing might have been the best thing to happen to me as I’d stopped learning and assumed I’d got the ABC side of poker down pat and was either winning or losing (50/50!). Since I was previously winning it all seemed great, but now I think it was just lucky really (as I’m not a complete idiot it was luck biased towards winning, but obviously that’s very precarious still).

Moving back down to $10s to concentrate on how I was playing (rather than printing money) was feeling a bit pathetic, but then Yegor reminded me I was still learning, and wasn’t going to be rich tomorrow. So very true, I need to keep telling myself this until I’m ready to return. This was reinforced today when I found a three year old post on 2p2 saying exactly the same thing. Cwar was saying when 20 buy-ins isn’t enough – most of the reasons were to do with tilt & poor mental state, but the final one was;

-You are learning, no need to move up faster than your knowledge will allow you, going from the $10s to $220s quickly is not going to happen without you needing to learn and adjust. If you breeze through a couple limits fast I would recommend slowing down a bit just to see how you adjust to the better play at higher limits.

The same point twice really, but it rams it home nicely. Adding on top of that the fact that the learning I’m currently doing is really opening my eyes to a lot of new things means it all feels constructive, and I’m not wasting my time, even if I’m not moving up (despite the BR telling me to).

I know what my problem is, and have really known it all along TBH. I run good, get confident, and then start to embrace the fun side of poker – the gambling!

So I chase draws, make big cock swinging hero calls, thin value bets when I have no value, and generally discard the fold button. All with the obvious consequences.

At the weekend I buckled down and started to squash these urges, and I managed to stall the downswing. It’s not turned round, but the latter half of this month has been breakeven which is better than before.

I’ve been reading 2p2 intensely, going over my HH, having other people go over my HHs, talking strat on AIM, and generally becoming an obsessive poker bore. It’s feeling good too.

After a long discussion last night with Yegor (ok, most of it was laughing about roseeker’s downfall) I decided to drop right down to $10 games and set a strong volume goal. I’ve currently 240 games at $10 with a ROI of 13%. I’m going to play another 260 to bring it to 500 games and see if I can nail my game down to a cold impassionate cyborg crushing of these $10 morons. No more hero calls

So far so good, although I’d forgotten how bizarre $10 games can be. For $20 and $30 games the players have become far more rational about their lines, and betsizing is at least in relation to the pot size. $10s it’s still kinda crazy. For example I had 72o in a limped pot. The flop had been checked back so it was a $40 pot on a TJQK board. Villain fires out a bet of $200. Somewhat obvious what to do.

I’m still squishing some gambool tendencies though – same villain 3bet me pre while I had KJo. Against him I felt a call would be ok, but something in me said that he was just going to jam the flop with his A rags if I did. So I 4bet, which KJo isn’t really good enough for, and he shoved. So I thought ‘whatever’ and called to see that he did indeed have A6o. Here was the difficulty of $10 – he had no idea that a 4bet looks really strong and he was supposed to fold A6o. But that’s what we all say in these positions – ‘how could he shove with that!’. Ultimately these are mistakes which make us money. Anyway, neither of us improved, I went down to $900 and then had to fight another 10 minutes in order to finally win.

It was so easy to review them that I’m tempted to record games by default. The only thing that makes me hesitate is that while recording it makes Windows a bit glitchy occasionally. I’m already plagued with disconnect problems, so dunno if I want to add more crappiness to it.

I started out recording using the trial version of camtasia, and that was pretty good once I figured out how to set the frame rate & compression. At £200 for a license though it’s too expensive to buy (I’m not installing any warez – too much trouble with malware).

So then I tried UltraVNC screen recorder which is free, and it worked perfectly. Slightly more glitchy than camtasia, although I’d upped the framerate which was probably the cause. Here’s the result. I tried uploading it to YouTube and it came out really fuzzy which is a shame as the camtasia ones worked quite well, but I’m not bothered by that.

Since pokerstove already exists there’s not much point recreating the 5 high straight, so I decided to aim higher and crunch some numbers for Omaha.

My first goal is to calculate relative hand strengths preflop for a HU game. Plenty of people have done this for hold’em because there aren’t that many combinations (with two hole cards there’s only 1326). However with four hole cards the number of combination is massive, and to calculate the hand strength of a particular hand you need to run it against the entire range of those combinations. Plus if you want the perfect result then for each hand vs hand combination you need to run it against every possible board. That’s just way too much number crunching, so I’ve taken a few shortcuts.

Firstly the obvious shortcut is that you don’t have to consider every suit combination – AKAK is the same as AKAK.

I then didn’t calculate for every possible board – I took a random sample of 100 games (monte carlo style really). I was a bit concerned that this isn’t that many, but it looks like it’s more than the propokertools omaha calculator is using, but I had to choose some number. (plus yes it’s all multithreaded and hammering nicely at all 4 cores of my cpu – I’m not a total coding fish)

So with those short cuts I started it running on Monday. Today it finally finished. Three and a half solid days of crunching! If I’d chosen 200 games I don’t think it’d be twice as long, but it’d be near enough.

I ranked the hands in order and the win% had a nice range of 20% to 70% which suggests it’s not a waste of time doing this exercise.

Unfortunately it looks like I might have missed some hands out – the worst hand was 5555 at 19%. What about 2222? Looks like it’s not in the list! I need to figure out why my program missed that.

Then also unfortunately it looks like I may have to increase the number of iterations. The best hand is AA66 at 74% – weird, isn’t is supposed to be AKAK double suited? That comes in 14th place with 71%. This is a result of only doing the 100 iterations – a massive error range of +/- 3% at least I’m currently crunching those two hands using the entire board range to give absolute accurate numbers, but that could take a couple of days to produce the answer.

So not super successful so far – I need to increase the number of game iterations, which is obviously a trade off against crunching time. I suspect to get a reasonable result it’s going to have to crunch for a couple of weeks…

I’m not posting about a single $20 win obv but about how sharkscope is a bit of a double edged sword. I sat first since all the tables were empty and when someone joined me I sharkscoped them just out of interest. It’s funny how a little frisson of pleasure (or relief) appears when you see that little jumping goldfish by their name So I thought this should go ok (damn, positive thinking failed pretty quickly there – should have been thinking, yeah! in the bag man! Oh well).

As it turned out the game was a real PITA, and I came pretty close to losing it a few times. The guy’s weakness was that he’d go to showdown with any piece of the board. The eye opener was when he called down with Q7 on a board of 34973. He also held onto strong Ace rags to the bitter end, hitting on the river several times after I’d barreled the earlier streets.

That meant I had to hold back and just hope I hit the board enough to value bet him out of the game. Bluffs were going to be useless. Fortunately I was given those opportunities and it went my way. Just as well really as I think if it hadn’t my positive state of mind would have taken a huge battering.

Somewhat shaken by that near miss I added him to my buddy list and saw that he’d sat down again at a $40 hunsg (martingales perhaps?). I opened the table to see he was already started against someone. I sharkscoped his opponent and this is what I saw;

He’d sat with the 16th biggest winner on the cereus network. Should be interesting I thought so I railed the game.

It went quite differently to my game. Vulcain999 played very small – minraising pre and folding easily to flop bets. He didn’t seem to make any big aggressive moves and didn’t commit himself to large pots. Chesta69′s tendencies that I saw in my game didn’t seem to arise mainly because they just weren’t going to showdown.

As the game progressed it became more aggressive, and most of the gains were made in preflop 3bets. Vulcain999 was never ahead in chips, and finally he made a strong move which was called down with a low top pair (on a 3 to the flush board) and showed a complete bluff. The ranked player made the losing player’s leak into the right play, possibly because he just hadn’t gained that information earlier in the match.

I think why I’m posting this is partly motivated by Simon’s post where he decribes a fish he played against 4 times and lost. In husngs understanding the opponent is the main thing – is he a station, a bluffer, a maniac. Will he fold to a cbet on a dry board or a drawy board etc. I’m saying the obvious here really. However, to objectify your opponent as simply a ‘fish’ doesn’t help. Even a fish’s bad move can be the right move if we play into it.

I did manage to sort out some problems with my game though (around tournament 15), but as you can see from the red ‘luck adjusted’ line I ran into enough ‘bad luck’ to not make any difference (tbh I don’t put any credence into these EV adjusted values at all, but like any good statistics; when they prove your point they must be right). It was pretty bad though, and one game I was up $2550 vs $450 when I disconnected – by the time I rebooted my shitty modem and reconnected I was the one with $300 left

Nevertheless, the problem I sorted out was my starting range. When I started out on husngs I was playing by instinct, taking my 6max range and widening it out to what felt right, and all was good with big rocketing graphs every day. I then found a starting range on husng.com and started implementing that.

Since that day I’ve broken even. Small sample, but it wasn’t right at all – I was getting into too many marginal spots, ending up overcommitting and spending most of my time fighting to come back from behind. So today I took a step back and thought what a heads-up range for SNGs ought to be. Basically since in a husng we’re playing shortstack poker most of the time, it makes sense that the Nash table is giving us the right distribution- after all, it’s calculated from the all-in preflop equity you have in a hand based on two players under shortstacked conditions.

So I twiddled around with the ratios until it presented 70% and 30% of the hands (so 70% is my opening range on the button and 30% is my calling range on BB). In this table the 70% is everything except the light grey, and the 30% is in bold;

The husng.com table varied in that it had the entire suited side for opening under the button with the junk as mainly the bottom section of the unsuited corner. TBH the husng.com table was for a cash player, and so that’s a very different dynamic, so I’m not saying it was wrong. But for husngs, it was totally unsuitable. My games instantly improved, with the number of awkward decisions vastly reduced. Shame the number of idiot calls that sucked out on the river didn’t.

Also I don’t care what anyone says about not needing hand charts. They’ll tell you that heads up is fluid and your range depends entirely on what’s best to beat your opponent. That’s true, but without having a solid and appropriate foundation from which to add or remove hands you’re just building castles in the air…