From Lead to Gold

Well lately things haven’t been going too well, I haven’t played that many games and I suspect this is due to a mental leak, a tilt in reducing my play when bad luck strikes.
Anyway never mind about that I have spotted an interesting effect in the AllinEv graphs that I didn’t realize would be there. I admit it is still early days and I might be seeing some random effect as a pattern in limited data but I’ll show you what I mean.
Ok I’ve been plotting my AllinEV vs Normal for lots of 180 seat tournies for a bit of fun really. I have plotted two lots.
(i)    Just the hands involved in allins – by filtering for allin by the turn or earlier. (Well there will still be a few a few hands with no calc done as other players might not be allin by then)
(ii)    The complete set of hands so in this many are not effected by an allin adjusted calculation.

Here is the latest version of (ii) for 180s only for this year.

 This is the graph that behaves in a inverted saw tooth form, it is quite easy to image that this shape is being caused by the nature of the tournament game, it is only when you come first that you get to keep all the chips.

Below is the graph of only the hands that can have a sensible allin adjustment calculated.

An odd thing I have noticed in this is that a remnant of the inverted saw-tooth shape seems to still be visible in these hands. Initially I expected the green and the red lines would simply bounce around the common mean and not exhibit this tooth shape, I hadn't really thought much about it.
It is only a small sample really and I am not certain that these ‘teeth’ I’m seeing are not tricks played by randomness but it seems connected to the question ‘Why Do I Feel So Unlucky?’
From the look of it if you play the bigger field games you will typically get long stretches of poor allinEv followed by relatively sudden jumps back up. So playing these you have to get used to this effect. The bigger the tournies you play the bigger stretch of bad allinEV you should expect, any early luck you get in a game tends to be overshadowed by a later hand - you have many chances to get unlucky before the end of each tournie.
I, and most others, probably feel the pain of the long stretch totally out of proportion to the joy of the sudden burst of good run-good and overall you feel on a bit of a downer.
This may not be true at all but it’s my current theory and I think I might try to train myself to expect this effect and perhaps it will help with my low volume tilt.

edit:

One thing I forgot to add that may be of interest to someone is the difference in variance in the green and red graphs above. It is easy to see that both do seem to be heading in roughly the same direction, I suspect if I plotted it for several years of data (the above is just this years so far) both green and red would end up relatively close to each other.

For the first graph, 'All tournie Hands', the sample variance of the green is 3.4 times larger than the red.

For the second graph, 'Allin Turn or Earlier', the sample variance of the green is 5.9 times as large as the red.

Yesterday…
It was the best of times, it was the worst of times, it was the age of wisdom, it was the age of runner runner suckouts ….

So things didn’t go too well but on the plus side I did manage to hit ‘the zone’. I actually arrived there via the difficult route. The hands ran out so badly that after a while I begin to feel numb to the pain, I knew I was playing really well and a thought kept running round my head, “Well sod it, I know I’m playing well, just keep going there is nothing you can do.” I then continued to play really rather well for another hour or two. The luck didn’t improve and I ended down by 25 BI’s but I know I played very well and also in that strange, very focussed state.

Last month was a better month overall than of late and I do think some adjustments I have made recently are paying off slightly. I didn’t play anywhere near as much as I would have liked but I did put some effort into improving.
I have found it quite useful to concentrate on small areas in order to find improvements and plotting out my performance at blind levels does seem to provide some areas worth looking into. Here is the latest graph of the data. The ‘thin’ plots start from the time I started specifically trying to make some post ante adjustments, about 4 or 5 weeks so far. I didn’t mention it last time but when looking for leaking areas I think it is much better to use the EV adjusted values as these at least take some variance out.

 I do need more data to be sure but I do think I have improved slightly, the combined result is moving upward but I won’t be too certain until about 4 or 5 thousand hands at each level are played. For this latest period my roi is +5% higher but only over 700 extra 180s so still too early to be sure.

Oddly, the one line I expected to improve most has stubbornly refused to budge and that is the 250 to 300 bbs one. I think I will have another look at my play at this level as there must be some decent leaks here. I would expect to be at least a few extra chips up during these levels rather than just flat-lining.
I don’t really know what’s going on with the bb2000 to 3000 line, quite a roller-coaster. I like its current direction but what the hell was I up to in December/January?
The costliest one, the bb4000+ is about flat but I haven’t really got enough data yet to see any proper trend with this one.It is worth checking out the play for these final table hands though, there must be plenty of value in getting good in these spots.

I’ve discovered that viewing my performance by different levels is quite interesting and may show some areas where my play is poor. This is great if true as I should improve with a bit of directed study. I am still a little unsure if the data is firm enough to consider this as showing a big leak. Is it normal to lose about 3bb/100 at the final tables of 180 tournaments? I think you should always be even or better at any level but 1 am unsure if there will be some sort of bias, I would expect it to be the very high variance part of the game and when you bink a few you get an up-shift..
So here is a graph of my chip performance split into a few groups of post ante levels. The thinner areas are from the point where I decided I had found a bit of a leak in this post ante play and made some corrections. It should be really interesting if I keep plotting these for a week or two to see any trend develops.

 The graph is a bit misleading in that it shows it from a chip view and not in bbs so the green line (1200 to 1600) is probably as bad as the orange. In the future I might convert the data to show this as bb’s won or lost.

You can see from the orange plot (levels 4000 to 10000) that I seem to be haemorrhaging chips at the late stages. I did think I would be fine in this area as I am coming at these from a reasonable STT background. I am not sure if I have made an unconscious change but in the last 2 or 3 months I am doing well.
I haven’t shown the pre-antes levels as there gets to be too much going on in the graph and at these levels I seem to be doing very well anyway.
The orange level is just all late final table hands with plenty of HU play. As I mentioned I thought I would do fine here but I think I mustn’t have adjusted well to the players at these, they seem less solid than the STT players so it could be that I am calling closer to Nash than I needed to and could perhaps push wider. Overall it may be just that I have been a bit too tight, I’ll have to try to look at this more closely to find the more obvious leaks.

Last week I jokily put in a link to a midi file that was supposed to sound like OMGClayDoll playing the 180’s. Ok it was Bach, well it wasn’t Bach playing 180’s – I think he was a more of a Heads-Up specialist, it was one of his Cello suite preludes and you can hear it played properly by Yo Yo Ma here.
http://www.youtube.com/watch?v=PCicM6i59_I
That is how a 180 session should sound, ending with a nice flourish as you take down the last couple of games.

Also if any cash players are interested in seeing their results analysed Gaussian style they could download this excel sheet. It’s quite simple and similar to the one I did for Tournaments, you only have to put in your own win-rate, number of hands and stdev to get some graphs. I would suggest that you use your latest several 100,000 hands to build an approximate w-r estimate and then use some fudging, increasing or decreasing your own estimate of w-r to adjust for run-bad or excessive run-good, and comparison with the average w-r figures. If you plug this in you can see what sort of result spread is reasonable over your next several thousand hands. I suspect most of us think we have been running bad.

Here is the link to the cash spreadsheet:

http://speedy.sh/Sb94g/CashSpread.xls

edit: I noticed the link had stopped working so I have re-uploaded to the above one.

Here is the 180s Ev Adjusted allin pre-turn graph for the current year so far, it is quite weird really. As you can see I have been nearly exactly even for the ‘luck’ but the odd thing is that over the whole time I have been very consistently losing flips except for about 10 hands of HU. It is very important to get some run-good in HU, it might even be the most important time for it but having a constant drip of bad luck over the period has been hard to take. Without these 10 hands the green would be running something like 45 degrees below the red and that’s quite bad over such a large sample. So there you are, I’ve run very nearly perfectly even EV wise but I can still find ways of complaining!

 Last week I also showed a graph of 180s treated as though they were cash games, winner-take-all fashion. Playing around with data can often make useful insights possible and since last week I’ve been struck by the shape of this (reversed) ‘saw-tooth’ graph. As I have some background in Physics I realize that graphs of any shape can exactly be reproduced by combining various frequency sine waves. When I say ‘I’ I mean a clever guy called Fourier realised this way back before internet poker was even born. In musical instruments a saw-tooth wave is often used to model the Violin family of bowed instruments. I guess 180s are something of a cello, 9s a violin, and monster MTTs must sound like a double bass.

 I thought I would see how this would 'sound' if I put all my 180s hands through a sound generator and played them as a tune. Well as you would expect it doesn't sound nice so instead I got hold of all of OMGClayDol's - now he can really play these 180s and here is what his sound like OMGResultsPlayedAsMusic.

If you have read my previous couple of blog entries I apologise, I put in a few fairly gratuitous graphs that didn’t really make much sense. So if you were stupid, I mean interested enough to look at those you might as well carry on and see what they look like now.
I’ll start with the most bizarre one, viewing the 180 seat chips like crazy winner-takes-all tournaments.

Hey look I’ve gone on a heater.
Here is the important one for, the ‘Winning’ for this year so far.

And here is a graph if you filter for the allin hands

 Nice heater there and I did manage to convert this sunshine into some nice hay. I ran at 160% roi for the last 3 sessions.

So I do know the logic that you just play enough and the luck part evens out but even so I still do find it emotionally painful during the bad runs. The more years I play the more hardened I’ll get to it and I’ll keep trying to work on the mental side to try to slide straight through any bad patches in future.

I'm still trying to find my edge at the 180s and so far I'm about at a 20% roi overall, not great but I hope to push this well past 30% in the micros to give some room for level jumping.
With allin stuff it is usually best to view it for a very large batch and here is my current db allin chip graph. I am officially lucky again, I only momentarily dropped below ‘lucky’ for a very short spell. Why complain? Well the previous db was a different story but that’s all in the past and it’s best not looking, given time it should all work out fine.

My mental game is still not close to perfect. For the past 3 or 4 weeks I’ve been running pretty bad, I’ve managed to keep a +ve roi but not by much, only 6% and that’s not good at micro 180’s. I do think I have now found a decent leak and this should help in the future, I think I have missed a little bit of value as the antes kick in. This is what good bad runs are for, time to have a hard look at your play. If I were great at taking all this bad run I would totally ignore the immediate results and just review the hands afterwards. If I stacked the games it wouldn’t be so much of a problem but I overlap them and so see the dirty effects as they happen. I actually prefer overlapping as I can auto-fold earlier to save time and also I can see some of the other actions like big allins etc.

On the mental front the emotion is still building and yesterday I managed to play my A game for a couple of hours but I got pounded on and decided to stop early. I should really have been able to take it and play on as I knew my game was good but my heart wasn’t in it. I played 23 games and lost one buy in overall.
I feel I should be better at this by now as I’ve had plenty of bad runs and I know how the numbers work, you just need to play enough to get your just desserts. Somehow it’s difficult to accept that the good runs are there as often as the bad but your mind seems to notice the bad so much easier.

Last week I added a graph of allin and EV adjusted allin, it was not really of much use except you can plot several months worth to see how what seems like a nasty bad run up close is just a blip when you display the big picture. So to keep this trend up I have decided to show an even more useless graph. This one is for this years 180 seat games, I’ve played 772 and only binked  4, slightly under average expectation. Anyway I wondered what a graph of chip value and EVAdj would look like with all hands including blinds and folded ones – the same as the typical cash ones. I had thought it should look a bit like a saw tooth graph and sure enough this is what you get.

This is what my HEM graph would look like if I played cash games where you bought in with $1500 and couldn’t leave until you had stacked 179 others. The only time you get to keep any chips (money) would be when you acquired 179 fellow scalps before you had bust out, oh and in this weird cash game the cost of each blind would rise every so often. Brutal for a cash game but I suppose this is what winner-takes-all tournaments are like, they are crazy. It’s all very silly and doesn’t really make sense but from this viewpoint I am down 300BI’s in 772 games ;).

I am interested in data and poker does provide so much, I am thinking of other perhaps more useful ways to display some of this. One way I would like to see my results is a graph like set of random walks of all the individual stacks in the tournies, it might have some interest to see if there is a common spot for my tournie life to shuffle off it’s mortal coil. To do this I will have to try to sql into HEM again rather than just exporting csv’s and the HEM1 db schema seems so impenetrable that I may never do it.

I’m still trying to improve my Mental-Game..I have never been the type to wildly tilt off due to running bad but this doesn’t mean I can’t get better in this area and over the last two weeks I do seem to have improved.

I think that an important part of the modern on-line game is being able to play a few tables simultaneously. It is possible to make money playing one or two high stakes cash games but for tournaments to maximize profit you do have to play more although as some sections of the game are straightforward you often don’t lose much edge in having a few open. There is a balance in this and sometimes it’s good to concentrate on just one or two.

If playing several it is important mentally to have decent decisions come to the fore quickly and the only way this is possible is to use the power of your subconscious. Trying to consciously think through a hand will use up way too much time. In the Jared Tendler book he mentions that too much emotion can hinder this process and dwelling on previous hands will scupper it, your ‘A’ game will disappear and so you will struggle, I think a lot of this mental process is caught up with playing on your instinct and this instinct is actually your subconscious brain doing it’s job.

Over the last couple of weeks I have been making an effort to avoid any emotional build up and I have seen some improvement. I have had a pretty bad time as far as ‘luck’ goes but I do feel like I can let this luck wash over me and not get tied up with it. This seems a much more comfortable way to play poker and I also think more profitable. Although the cards haven’t fallen well at all I have managed to stay profitable and during this bad run I’ve still managed +15% roi.

Here is a graph of the last two weeks of allin ev pre-turn.

 And here is a graph of roi during this period.

It’s important to realise that in the long run any bad luck does even out and so just let the luck take care of itself and play in the moment.

To show the effect of time I can show a plot of my all the pre-turn allins in my current db (only includes games since last September).

 As you can see the red line and the green will eventually appear to meet if you have enough examples and over this period I have actually been lucky. Well it’s about a 0.5% difference.The variance of the adjusted (red) line is about a fifth(**) of the green and HEM seem to have managed to get the calculations un-biased – it’s from only my own recent results but it seems reasonable to infer they will appear to meet given a large enough sample.

The allin ev for toiuranaments is not a perfect indicator of roi but overall it is clear that not winning flips often will hurt your roi.

Another interesting thing about the allin graph is its slope, is there an optimal slope to have? This db's data is nealry all from180s and I suspect that the tighter you are the more steep your lines get. If it is too steep you are probably missing some push or call spots.

(If anybody is interested I had to export the hands as a csv file and process this to get the all-in graphs as HEM doesn’t show this for multi-table tournaments.)

(**) Edit: I corrected the calculation of the variance so the green is actually 5.17 times the red, the stdev of the green is 2.28 times bigger.

Sometimes you get into a zone that allows your A game to shine through for long periods and sometimes you don’t. I suspect emotions play a major part in this process and I do wonder if the Mr Spock, purely logical approach would work.

Yesterday I had a pretty bad session and I dropped about 30BIs playing 180s. The really good thing about the session was I kept quite calm and steady and I do believe that this way you have more chance of finding ‘the zone’.

In Jared Tendler’s book, “The Metal Game of Poker”, he describes the Adult Learning Model (ALM) and I feel there was a lot of useful insight for poker in this. If you can get plenty of Unconscious Competence into your poker brain it would help greatly in multi-tabling and also the general application of your A game and this may also create a zone to get into.

Recently I also read an article in The New Scientist entitled “Zap your brain into the Zone: Fast track to pure focus”, by Sallee Adee. In this article Sallee goes on a sort of Shoot Them Up video simulation course and fails miserably the first time through. After being fitted with a TSDC machine to zap some current through her brain she later aces the simulation by what she describes as entering the zone. Well I suspect all us poker players would want one of these machines and maybe Nanonoko already has one. Sadly we can’t buy them – research organizations only and they cost $50,000 plus. So as an Ever Ready 9 volt battery isn’t going to do the trick I guess the best way is through some form of self-mental control.

The meditational approach of Tommy Angelo would help greatly or basically any way of avoiding the emotional build up that often happens during poker sessions. Well, yesterday I didn’t quite get into the zone but I did give myself a greater chance by staying calm and in the future I am going to try to practice reducing any emotional build up. The Mental book does cover this area and I do intend to re-read this soon. During the session many hands unfolded badly but I somehow seemed to be a bit better than usual at just seeing the hand as good play and moving on very quickly. If I keep trying to improve this area I think I will have a more comfortable playing experience, more success and more ability to handle the decisions quickly enough.

Yesterdays graph:

Sorry to get your hope up but maybe some just might like this spreadsheet I've knocked up, I am not an expert at excel so it's not too refined.

Tournie players have you ever wanted to know what the Standard Deviation of your results to six decimal places were, or see the shape of the Normal ‘bell curve’ of your expected future profit if you play 1000’s of games with this same distribution? Ok probably not really.

Perhaps you would like to know the 90% or 99% confidence interval of your results if treated as a Normal distribution. Ok it’s not of much value, if you want to you can open it now or wait for Christmas Day by downloading it later, but if you peek inside the later linked excel file you will find a spreadsheet. If you enter your own finish distributions for any specific sized tournaments you can get a few nice graphs, a Normal PDF (bell curve), a Cumulative Profit vs % Chance curve, a Earnings Envelope and well one I'll call the BaseMetal Trumpet of Roi.

Four graphs for the price of …. well nothing really, you don’t have to register or do anything – you will want to enter your own distribution percentages. You will then get tailor made graphs and you can put in any number of games to see your likely future earnings (Warning roi’s can go down as well as up).

Here is an image or two of the graphs. Everything should auto-adjust when you put in your own data, the data fields you should modify are the yellow coloured ones, it should be fairly obvious what to do. I have done a sheet for STTs, 18s, 27s, 45s, 90s and 180s but you could easily add your own or modify the ones present.

Don’t forget to virus check the file, it is a simple spreadsheet with no fancy macros or anything so it should be fine.I did full virus check the original at the online VirusTotal.

To download the spread sheet click on the "TournieSpread.xls" at the top of this page:

http://www.speedyshare.com/pGjDc/TournieSpread.xls .

Cheers and 'God bless us every onel' from Tiny BaseMetal.

I haven’t posted anything much for the last two weeks as I have been cruising on the Pokerstars SS Titanic and global warming doesn’t seem to have shrunk the ‘bergs any. Well the only good thing I can think to say about this last two weeks is that it seemed shorter than watching a James Cameron film.

I have never quite felt so unlucky at the tables and I have had a few bad weeks – I might have told you about them - take my advice, don’t read my previous bloggs.

This is a graph of the month so far, the tournies are a mostly $7 45s but the last 100 or so are 180s. Yesterday I did manage to play 90 of them with just 3 minimum cashes luckily I did hold my nerve and managed to come 4th and bink one in the last three. Phew! these 180s do toughen you up.

It has been a hard week but my mental game is strong and my game hasn't slipped much at all during this. We all have an A,B and C game but I do think that there isn’t that much space between mine, I do still play less when things go wrong but I am improving in this area. This new downswing does seem like a natural point to try some study to improve, I’ll try watching a few more vids and do a bit more reviewing. Why did I slip that far – was it ignorance or apathy? I don’t know and I don’t care.

Sorry about that last one - I’ll just go shoot another Albatross.