Shocking findings while analyzing some stats

[sirio11]

I was analyzing all my tournament stats from 2004 to the present and I decided to make a comparision between my 2004 stats and my July-November 2005 stats (this is after the WSOP).

One of the main changes in my game between both periods was that I think in 2004 I was less aggresive than today, I was very money conscious, always keeping an eye in the changes in the money, making considerations to make the money and then to make it to the next money change; and in the final table, very, very conscious of how to play my hand depending of the stacks and the money changes; in general I had a step by step approach to the money. Today, I barely care to make the money, my approach is to make the final table and of course I have had several bubbles in this span.

I remember when somebody mentioned that you should just care about first place and not about placing, I always thought of this as a macho-attitude in poker, since I had so good results without that attitude. But gradually I have been changing to that side of the spectrum, I still don't have that 1st place or die attitude, but I feel in the practice my game is moving in that direction.

The majority of 2+2ers and in general in the poker world think the 1st place or die is the best approach to win; they even critizice, often harshly to those who have a conscious making the money step by step approach.

So, this change in my game looked like it was all good, I'm joining the majority with my approach to the game and I have had a lot of 1st places finishes in the last months; so it should be right.

But then I decided to compare my ROI and I had a shocking surprise.

ROI(2004) > ROI(Jul-Nov 2005)

ROI(2004) >>> ROI(2005)

1st places finishes Jul-Nov 2005 >> 1st places finishes 2004

And of course

ITM(2004) > ITM (Jul-Nov 2005)

I understand that maybe the number of tournaments is not enough and that the numbers could be some king of statistic aberration.

#tourneys 2004: Around 600

#tourneys 2005: Around 900

#tourneys Jul-Nov 2005: Around 350


Still, I want to present these facts for discussion. I think it would be a very interisting one.

For now on, every time somebody write something about how he could care less about making the money and how he just play to win (whatever that means) my answer will be:

"Don't talk too loud man, just show me your ROI"

ROI rules.

David

Interesting numbers to say the least. I'd have expected the win-or-die attitude to give a better ROI in the long run. And I'd expect 2005 to be a great year for you with the stats you've posted on your winnings.

Something that might explain some of it could be that you've changed the way you play. You play more tournaments now. Do you multitable more? Do you play against tougher opponents in higher buy-in tournaments?

I also think that while ROI is interesting, in the end it's your earnings in dollars that's the really interesting number (maybe expressed as dollars/hour) and the number you should try to maximise. I'm pretty sure you could get an awesome ROI playing $5 tournaments, but I doubt you could make a living doing it.

[sirio11]

[ QUOTE ]
I also think that while ROI is interesting, in the end it's your earnings in dollars that's the really interesting number

[/ QUOTE ]

Yes of course,

$$(2005) - $$(2004) = $3,000

With about 300 tourneys more.

Yes, multitabling could be an issue definitely.

Obviously, playing entirely for first (i.e. pretending tournies are cash games) is inoptimal because the payout structure clearly encourages surviving. The question is how inoptimal is it.

When the big stack has 1M chips, you have 100k, and shorty has 1, at 5k/10k blinds, clearly playing for first is inoptimal. Someone just posted a great and much less extreme example from the WPT.

There is a mathematical answer to every problem in poker. The frustrating thing about MTTs is that the math is frequently too complex to do easily because your equity in a tournament is related to every single other stack in the tournament. (Contrast this with STTs in which the math is solved.

I think that the reason most successful MTTers tend to argue that that playing only for first is correct is because there are two forces that change optimal play in MTTs--the payout structure and difference in the play of your opponents. The fact that your opponents tighten up around bubbles means that you should frequently loosen up.

[gumpzilla]

I will say that in the SNG forum, where people bang out thousands of SNGs a year in a game with substantially less variance than MTTs, a couple thousand tournaments is seen as a pretty good (but not great) sample for determining ROI. I'm guessing for MTTs, you're going to need many, many more before you can really nail down how good your ROI is. So please expand on how much >>> is; I think there's an excellent chance that it's not statistically significant. Nevertheless, I think there are many situations where it is absolutely correct to focus more on making the money than "#1 or bust." sheets's JJ hand is an excellent example of this, to my mind.

[PrayingMantis]

[ QUOTE ]
Contrast this with STTs in which the math is solved

[/ QUOTE ]

I wouldn't say it is solved. The link sends to the ICM calculator, which is a way to get some "solution" for a question asked in a context of an accepted model, not more. It surely doesn't mean that the game is solved. In reality it is quite far from being solved.

[ QUOTE ]
[ QUOTE ]
Contrast this with STTs in which the math is solved

[/ QUOTE ]

I wouldn't say it is solved. The link sends to the ICM calculator, which is a way to get some "solution" for a question asked in a context of an accepted model, not more. It surely doesn't mean that the game is solved. In reality it is quite far from being solved.

[/ QUOTE ]

The math is solved. The game isn't because we don't know exactly how our opponents play. In MTTs, the math is solvable by ICM, but the problem requires a ton of inputs and changes with different pay-out structures and field sizes.

[gumpzilla]

[ QUOTE ]

The math is solved. The game isn't because we don't know exactly how our opponents play.

[/ QUOTE ]

This is a misunderstanding. ICM is not the exact equity of the game; as is popular to point out, it doesn't even incorporate blinds or position, so it is obviously missing parts of the game that are relevant for determining equity. It is a model which matches well with the intuitions of experienced players. PrayingMantis is right when he says that the true mathematics of the game are not solved.

[ QUOTE ]
[ QUOTE ]

The math is solved. The game isn't because we don't know exactly how our opponents play.

[/ QUOTE ]

This is a misunderstanding. ICM is not the exact equity of the game; as is popular to point out, it doesn't even incorporate blinds or position, so it is obviously missing parts of the game that are relevant for determining equity. It is a model which matches well with the intuitions of experienced players. PrayingMantis is right when he says that the true mathematics of the game are not solved.

[/ QUOTE ]

Fine, fine. It's just really really close to solved.