Liar Dice - Diary of a Bean Counter!

Thanks, champ. Look forward to talking on skype - love those 18 mans

Latest Video

My new PokerStrategy.com video on Spin & Gos is out! Hope you guys enjoy it

Looking forward to the Christmas wind-down, now. Just a few more hours to coach, a few more games to play and then it's time to pull off my stockings and dry them by the fire. Wait. I didn't mean that.

Merry Xmas, everyone! See you in a few days and, for those of you are not too politically correct, here's a naughty Xmas elf to bring you some festive cheer.

♠_drink:

Nice vid Tim with the spins and go. I've got a question for you, when you grind those, do you 5 table them or is it too much action at same time? Also, in HM2, to track my EV in those, the proper stat it Chips (EV Adjusted) or something else?

Hi Vinny80 and happy new year! Thanks for the comment

I usually 3 table the Spins which is probably similar to 5-6 tabling 6max sngs, in terms of decisions/second. I think many players are comfortable playing more tables, but I always prefer quality vs quantity.

In HM2, normal tourney EV isn't that useful (though it gives you a rough idea). The reason is that, when you lose, it ignores the prizepool you were actually playing for and compares your EV to your actual buyin. But when you win it looks at the specific prize pool for that game. i.e it doesn't correctly quantify that losing a flip in a $30 game with a 5x multiplier is more "expensive" than losing a flip in a 2x multiplier.

I prefer to use chip EV as a general guide to my overall profitability in those games, therefore, and EV adjusted will show whether you are running well or badly overall.

New Year Aspirations and December Wrapup

So, it's time to put down some thoughts about 2015. First exciting thing is that we're finalising details for a live coaching series on Spins/Twister/Jackpot type sngs in the near future. Oooh. Purple, eh? I didn't think I had it in me! My recent Spins vid was the most watched PokerStrat vid for a few months ♠_drink: so there is a lot of interest in these games, regardless of what the haters say.

In terms of my own development, I'm going to aim to play a decent amount of cash this year, in the hopes of once and for all establishing myself as a winning nl100/200 player who doesn't tilt every time he loses a pot Since I don't tilt at SnGs, there's no reason I should tilt at cash, so the key is to find out why I did in the past and fix the underlying psychological weakness.

I'll continue to grind SnGs as my main playing focus, mixing limits/formats/sites in the aim of achieving max $/game and largely ignoring rakeback, since most sites have significantly reduced rbk for UK players, as a form of tax-sharing. Hoping to add some new stakees and students as well, of course.

December
...was a quiet month. I had to look after my invalid mother after her hip operation so volume was very low (shock horror). Managed a final total of $2,698.42 including the usual diverse income streams.

With a bit of luck my health will stay comparatively reasonable as I enter the last year of my 30s.

Have a great year, all!

Another SnG champ looks to make the transition to cash...:censored: ok, I'm over it...

to you Tim in your 2015 goals!

Live Coaching:

As promised, my new live coaching slot kicks off on Thursday at 8.30pm GMT with some Jackpot SnGs.

Be there or be ... unhappy and unfulfilled 4evr.

Hello

Now, the Numbers:

ROI based on chips won: ~9.8% (51.7* $0.0576 per chip = $2.977 per game)

I think, you made a mistake:

First of all rake on 30$ is 5% so:
90$ - 5% = 85,5 / 1500 = $0.057 not $0.0576

And now $0.057 * 51.7 = 2,9469 - ROI 9,8% - it looks reall nice but is wrong and unreal.
You have to subtract 5% rake for each tournament: 2,9469 -1,5 = 1,4469 = 4,8% ROI based on chips won

To achieve 10% roi you need to earn 79 chips for each tournament.

Just look at my graph:

I won 310$ in 111 tournaments - that gives 2,7929 per tournament.

Ok lets fallow your way of thinking:
2,7929 / 1500 = 0,001861
I won 6500 chips / 320 = 20,3125 per tournament.
So 0,001861 * 20,31 = 0,04 per tournament. That means that i should have 4% ROI but i have -3% just because i have to subtract 7% rake. 0,04 - 0,07 = -0,03 = -3% ROI.

what is this math? looks interesting!

diabelb, I`m not sure if I remember correctly, but I think if you want to look at your chip ev line, you should filter just for hands that were raised all-in, looks like you have all hands there.

The graph should look something like this.. (would be better obviously if the green line went above yellow )

Just simple calculations

If you would like more look at this:

[Edited by Lazza61]

My chart include all hands from tournaments. Of course i can filter only allin situations but it really tells nothing:

Thanks, diabelb for all that info!

I will take another look at those calculations - it did seem too good to be true. Although one thing to point out immediately is that when I ran the numbers originally they were based on the lower rake in force. $30s had only 4% rake originally - Stars increased it to 5% in November.

One other thing that you can probably clarify quickly is that, when we assess the prizepool as being "90$ - 5% = $85,5" aren't we already taking into account rake? Why do we need to deduct another $1.5 rake from the $2.94 figure? I mean, if we do that then we are counting rake twice, aren't we?

I'm no mathematician and make more mistakes with numbers than someone doing my job really should, but that was my initial thought. Can you confirm?

Thanks and nice to see you dropping by guys

This is weird and unintuitive. My first calculations was similar as yours. But than i realised that something is wrong - my chip won line goes up but i'm loosing money. I do not know how to explain that.

Lets assume that after 1000 1$ tournaments your chip chart shows 0. So we are break even? (0 * 0,001861 = 0$ per tournament :f_drink:). No we are not break even. If after 1000 tournaments we earn 0 chips and loose 0 chips we are just getting back our BI reduced by rake.

It's entirely possible for your chip line to be going up and still be losing money in the short to medium term, because turning chips into $ is dependent on 1). getting a "fair/reasonable" distribution of good multiplier prize pools, and 2) having decent luck in those particular games.

I agree with your second point - if you made no chips after 1k games you will be break even minus the rake you paid (although of course even that does not follow, because you might have had a 1200 multiplier and won it!). But just because we know that we need to earn a positive number of chips just to stay ahead of the rake, doesn't mean we should count rake once when we work out the average prizepool and then again when we work out our winnings.

I had a look at Aarnimetsa's article and it looks right to me. As I said, when I originally calculated it rake was 4%. Stars increased it to 5% (i.e. 25% increase in rake ) so it's definitely true that 51.7 chips/game will be equivalent to a significantly lower ROI than before the rake change.

I'm alreaddy almost sure my calculations. I tried small sample from Fulltilt jackpot:

Date:
tournaments: 82
Net won: -8$
Chips won: 1000
1st place: 28
Prize won: 74$

So first of all distribution of prizes was not perfect - i won 74$ in 28 tournaments - that means that ave won was 74/28 = 2,6428$

Now we can calculate rake: 1 - (2,6428/3) = 0,1190. That meant that because of bad multiplier distribution i payed 11,9% rake.

Ok so i payed 0,1190$ rake x 82 tournaments = 9,7619$. I payed 9,7619$ total rake.

Now we can check the value of chips. I won 1000 chips. Ave won is 2,6428$ so 2,6428$/1500*1000 = 1,7619$

Summarizing i won 1,7619$ and payed 9,7619$ so my net won is 1,7619$ - 9,7619$ = -8$
Bingo ♠_drink: - we've got the same number that PT shows to us

Similar we can calculating net expected won just by taking "net expected chips won" and ave won 2,85$ (5% rake on FullTult)

Yes, I agree. You can notionally reflect "bad multiplier distribution" as being increased rake. Your calculations make sense to me, if that's what you are doing. They demonstrate nicely that most samples will show notional rake as being much higher than "actual" rake. Indeed, that's the case with my results, too

But if we had a sample with perfect distribution, we would have a prizepool of $2.85 exactly for each $1 Jackpot we played. And in that case we would be entitled to divide the prizepool by the number of chips (1500) to get a value for each chip ($0.0019/chip). If a player on average won 50 chips/game they would win $0.095 per game assuming that perfect distribution.

Just intuitively, 50 chips won per game would be 10% ROI because your starting stack is 500. The reason it's not actually 10% (and only 9.5%) is because of the rake which reduces the prizepool by 5%.

Here, maybe, is the problem: In your example, you worked out the notional rake ($9.76) which was 11.9%. You then deducted that from your winnings of $1.76. Because that gives you a loss of $8 (which PT shows) you have justified the method of deducting rake from winnings. But when you commented on my original calculation, you wanted to deduct $1.5 (i.e. the actual rake) from average tournament winnings:

Originally posted by diabelb

And now $0.057 * 51.7 = 2,9469 - ROI 9,8% - it looks reall nice but is wrong and unreal.
You have to subtract 5% rake for each tournament: 2,9469 -1,5 = 1,4469 = 4,8% ROI based on chips won

But we can't do that because, according to your method, we don't know what the actual rake is until after the sample is available. Note also that, if a player is lucky enough to hit a high multiplier in a small sample, their winnings/tournament would be higher than the average and this would make rake a positive amount!

Well now my brain really hurts - somewhere the truth is out there Time for my bed!

P.s. On reflection, calculating rake as you have done is probably misleading and, I suspect, innacurate because it relies only on the prizes you actually won. It ignores any games you played in where you played for a big multiplier (I.e. bigger than 3x buyin) but did not cash. In the 82 game sample you had, we can't know whether you really paid more or less than 5% rake without seeing the total prize pools you actually played for. And unfortunately neither PT nor HEM show this.

1) Chip EV is a flawed measure of performance in SnGs, even winner takes all SnGs, because, if you run as badly as I do, you can easily have a chip EV of more than 1500 chips at the end of the SnG, even though there are only 1500 chips in play. If you're looking to whinge in a quantitative way, I think you're probably not going to be able to.

2) Actual chips won is completely equivalent to ITM. Why? Because in each SnG you either lose 500 chips or win 1000, irrespective of however wiggly wobbly your chip count is during the SnG. If you plot a graph that goes up 2 for a win and down 1 for a loss, give or take some stretching and squiggling, it'll be the same. The cash equivalent is up $55.5 for a win and down $30 for a loss.

3) 'But what about those pesky multipliers?' I hear you say. Well, Pokerstars don't take anything other than the rake out of the prizepool so, in the long run, you can ignore it in calculating what you'll win. It just adds a gigantic dollop of variance on top, which, I'm sure we can all agree, makes Spin n Gos so much fun to play, at least for fish like me. You pros can go fret about your poxy variance. I just like watching the little numbers spin round so that I can piss my pants with excitement when I get a 3x multiplier.

My conclusion is that there's no need to use chip count and fiddle about converting it into $ when you can just use $55.5 for a win and $30 for a loss and work out your ROI from that. In addition, since you're pros, you don't care about your EV ROI because you have a Zen mindset.

Much appreciated, JB. I like it when a trained mathematician comes to my blog and tells me I'm talking nonsense because I can stop agonizing over it. So there you have it: use your chip EV as a very rough guide as to whether you are playing profitably at Spins, in a similar way that you can use EV bb/100, but for estimated long term ROI, stick to ITM (as long as you have a reasonable sample - over 700-1k games, I'd guess).

I was thinking that someone should write a little simulator for these things to illustrate how bad the variance is for a given ROI. It's a five minute job if you know the probabilities of the various multipliers. I suspect that the really big multipliers suck out so much of the prize pool and happen so rarely that the results will be terrifying. I don't care, so I cba to do it myself.

We could ask Pokerstars to do it? I'm sure they'd be only too willing to help out.