Most of the big names travel the domestic circuit, and even the world.. exotic and expensive destinations..

I've never even played a 1k$ tournament, but it seems to me that it's reasonable to expect that a 50k buyin would have a much tougher field, and that players would naturally have a lower ROI in such a tournament. Are you (anyone who wants to calculate ROI in a way that's not weighted by buyin size) arguing that this is not the case?

[QUOTE=NickMPK;33811481]It seems pretty silly to be deducting for expenses in this situation. A sizeable number of these players live in Las Vegas. And if you are just talking about general food and rent and stuff like that, everyone has to pay this...you dont usually deduct this beforehand when estimating your income. Everyone has to pay taxes too.[QUOTE]

But we're talking, at least in part, about the viability of being a pro tourney player. And being a pro MTT player has certain features that other careers simply do not... First, unlike most jobs, it is possible to lose money over long stretches of time. Second, while all income is taxed, my understanding is that the tax code is pretty harsh on gambling income as opposed to a lot of other sources. Finally, while many players do live in Vegas, full time tourney players do have to travel quite a bit in order to play in big series like WPTs, EPTs, the LAPC, etc.

So, as a data exercise, I agree that we shouldn't be trying to net out taxes, living expenses, etc... But, when discussing the viability of playing MTTs for a living, you have to acknowledge that the live MTT lifestyle has a lot of overhead...

It's actually slightly more than 2000 tournaments entries...although this doesn't reduce the variance of the estimate that much compared to 1500 entries.

But I don't really agree that this is a "laughably" small sample size. As I posted above, you can actually get a fairly confident estimate of certain statistics, like cash %, from this sample size. And the cash % we see is consistent with an ROI of about 30%.

In terms of actually estimating ROI, the bigger issue than the absolute sample size is the vastly different weightings that tournaments have in the posted stats based on buy-in level. The $50k tourney by itself is having a huge influence on the stats, and this sample size is truly minuscule. When this happens, your effective sample size is much smaller than your actual sample size. We would probably be able to calculate a much more confident ROI estimate if winnings were reported as # of buy-ins won, so that each entry could plausibly be weighted evenly. Of course, I would also expect this to increase our ROI estimate, since we would imagine any player's real ROI is going to be high in a soft field $1.5K NLHE tourney than in a $50K tourney with 100 entries.

2009 Top 25 Earners according to Hendonmob at this years WSOP
No 24& 25 didn't play many events/ the search couldn't find them.

What it actually tells us is that if someone whose real cash % was 14.7% played in ~2000 tournies, they would cash between 13.2% and 16.2% of those tournies 95% of the time.

It also tells us that if the group's real cash % was outside this range, we would observe this cash % very rarely, so it is thus probably unlikely that the group's cash % is outside of this range. For instance, if the group's real cash % was 13.2% (on the lower bound of the interval), they would only cash 14.7% of the time or more over 2000 entries about 2.5% of the time (half of 5% since the test is 2-tailed).

Yeah, to me the most interesting conclusion from this data is that to play poker tournaments for a living you really need to be a winning cash game player as well. The reason is that you have to have cash game winnings that you can use to write off your losses in the years that you lose. If you did a thought experiment where you assumed that each player in the list represented a year of tournament player for a "poker corporation" the corporation would end up paying 80-90% of its winnings in state and local taxes if you assumed a combined tax rate of 35%!

Another way of looking at this could be estimate # of buyins won per event to get around the entry fee magnitude effect. It's not necessarily better, just a different perspective. Regress # of buyins onto player rank tiers and throw in a control for field quality (or field size / buyin could be a good proxy for that) so we can still see the effects of event type.

And people saying sample size should take note that the Levitt paper basically has the same ROI estimates and presumably ~same N, so actual size is probably double what OP is using. Not that 3k is a ton of events for MTT, but it's getting closer.

If you track bbs/100, you can know a player's true win rate (and therefore extrapolate ROI) with much greater certainty than by just looking at $ win/loss. This is obviously hard as **** to do live, but not impossible.

As an aside, people don't swap / get staked nearly as often as they should mathematically, but nobody really cares about math so whatever.


No, that's not what people are saying.

Regardless of it yielding any definitive conclusions (our Cal/WVU fan above covers it well), it was still a pretty interesting chart to read.

The database includes just this year and last, right? I'd be curious to compile this chart with the same player field only with 2011 data. Also, I ran a WSOP pool this year -- perhaps it would be interesting to plug in my league's field to see what comes of it.

You could try and divide old school from new school. Hellmuth, Ivey, Negreanu, ... being old school, and internet players being new school.

If someone compiles 2011, they should use the list of players drafted prior to the 2011 series. This list (players drafted before the 2012 series) includes a lot of names that are only on the list because they ran extremely well in 2011, so it is almost certain that the results among this set of players will be better in 2011.

I would love to see a version of these numbers done with the 50k excluded.

"Vanessa Selbt is a beast. Invested less than 100k and got over 600% ROI...it's just amazing.
Investment lesson 101: Do not invest 72k in the stock market for an average return of 12% annually when you can play poker and have a 600% ROI in less than 2 months. Conclusion: learn how to play tournaments and stop calling it donkaments just because you can't compete in today's poker world. "

Yes indeed, enough with the donkaments crap. Just become a great tournament poker player and run like god. And **** the stock markets.

just for the record, would taking out the 50K improve the roi or not? I don't know, but since a "top pro" won it likely it helped the roi.

obv the drafters didn't pick the real top 88 players, that has to have some effect here also.

OK. Here are the totals WITHOUT the $50K Players Championship included:

Buyins: $7,536,500
Cashes: $10,467,262

Combined ROI not including the $50K event: 38.89%

(there were 58 entries in the $50K for 8 cashes totaling $3,167,569 from the 88 players in the pool)

he's on the list in third position from top

BTW, the 2011 list is available at http://www.25kfantasy.com/2011/playerList.php.
There were 120 players drafted. That list includes # of entries and # of cashes, but not prize money won. It looks like the 2011 players only cashed 12.5% of the time, out of ~2500 entries.

Oh, yeah sorry, nevermind....I'll delete my previous message.

Of course I am not saying that. 1k ROI's dwarf 50ks.

I am saying that in terms of understanding what peoples edges are in fields, it's more relevant to determine ROI in terms of buy ins won, not money, since there is a massive range of buy ins.

If I played 1k 1ks, and profited 1 million, then I played one drop and lost, my ROI would be 0. Does that mean what it should mean? (at least in terms of future action selling, determining actual attainable edges, bankroll management, etc).

great work OP, thanks a lot!

Many small backers aren't going to put money into the biggest events, so this new average is more accurate for your average investor.

Not sure if it works the same way in the States, but they also probably deduct some of the spendings in their tax report.

A related thread started questioning whether it is worth playing tournaments due to high variance. One way to make tournaments better is to lower the variance by paying more spots and reducing pay at the top.

Tournaments used to be winner take all in the beginning, but as fields got bigger they began to pay more spots due to player demand, but unfortunately, the flattening at the top has not come down nearly enough to match the huge increase in field sizes.

Looking at this years main event at WSOP with 6598 players, a more reasonable top award would be in the $5 million range, not $8.5 million. And paying 12.5% of the field instead of 10% would help keep more player in action as well.

Here is how it could look:

WSOP Flatter
Actual Pay 1/8 Field % Difference
1 $8,527,982 $5,285,727 -38.02%
2 $5,292,889 $4,019,031 -24.07%
3 $3,797,558 $3,073,477 -19.07%
4 $2,850,494 $2,363,986 -17.07%
5 $2,154,616 $1,828,861 -15.12%
6 $1,640,461 $1,423,155 -13.25%
7 $1,257,790 $1,113,972 -11.43%
8 $971,252 $877,125 -9.69%
9 $754,798 $694,752 -7.96%
10-12 $590,442 $553,598 -6.24%
13-15 $465,159 $443,783 -4.60%
16-18 $369,026 $357,911 -3.01%
19-27 $294,601 $290,417 -1.42%
28-36 $236,921 $237,099 0.08%
37-45 $191,646 $194,766 1.63%
46-54 $156,293 $160,987 3.00%
55-63 $128,384 $133,899 4.30%
64-72 $106,056 $112,070 5.67%
73-81 $88,070 $94,394 7.18%
82-90 $73,805 $80,014 8.41%
91-99 $62,021 $68,259 10.06%
100-162 $52,718 $58,608 11.17%
163-234 $44,655 $50,648 13.42%
235-306 $38,453 $44,056 14.57%
307-378 $32,871 $38,574 17.35%
379-450 $28,530 $33,999 19.17%
451-522 $24,808 $30,166 21.60%
523-594 $21,707 $26,945 24.13%
595-666 $19,227 $24,231 26.03%
667-747 $0 $21,939 100.00%
748-825 $0 $20,000 100.00%

Some may think $5 million is "too little" to pay for 1st, but think of all the benefits:

You can afford to pay an extra 159 spots.

Only 27 spots pay less, 798 spots pay more. And odds of a player with an 80% ROI finishing in the top 27 is only once every 120 years or so.

A 10%+ drop in payout only affects 7 players, but a 10%+ increase in payouts affects 735 players.

A whopping 43.9% of the pool is paid to just 9 players out of 6598. Don't kid yourself, this is a lottery structured payout. That drops to 33.3% under the alternative payout. Still a large amount if you think about it.

The gap in pay from 1st to 2nd is unusually large at 1.61x that of 2nd place. All other gaps are usually in the 1.1-1.35 range. This should apply to 1st-2nd as well. The jumps in the alternative payout are generally in the 1.1-1.31 range including only a 1.31x gap from 1st to 2nd.

While the difference from 1st drops by $3.25 million, it only drops by $2.1 million after taxes. (based on USA tax code, single filer status) In fact, all taxes paid by the winners total $14.9 million vs $13.4 million under the alternative payout. This $1.5 million difference may not sound like much, but it amounts to an extra $2331 per player for the 666 players that originally cashed.

The average tax rate drops from 24.14% to 21.64%.(add in vig of 6% and inability of USA players to carry losses forward in losing years, and you can see why many tournament players are going broke)

The ROI for a good player does drop some, but not a whole lot. A player with an 80% ROI now has a 70% ROI. A player with a 41% ROI now has a 35% ROI. But this is offset by a huge decrease in variance. The current variance for the main event drops from 209.37 to 108.45 under the alternative payout.

To put this variance in prospective, the variance on many video poker games is around 25. The variance on blackjack is about 1.1, and if you consider the variance on NL cash games based on buyin amount, it seems to be between 0.4-1.0.

Bottom line, paying more spots and flatter payouts will help to keep more players in action than the current situation.

An interesting question this chart brings up is whether one should even go to vegas or simply play online? Referring to mtt players here mainly- could well be argued that there are a lot of players going that would be better off playing online- agreed?

flatter payout structures result in lower ROIs and higher risk adjusted returns, this shouldn't be a surprise