Not making any comment about whether or not this MU is +EV or not, but it strikes me as a bit unfair that threads get people saying OMG SCAM this MU is too high!, but no thread ever gets derailed with people saying STOP seller, you are getting scammed, you are worth MUCH higher markup!!

Why are buyers who buy at a MU they know (or believe) to be advantageous for them over the seller any better than a seller who sells at a MU they know (or believe) to be advantageous for them over the buyer? If the seller bears the brunt of too-low markup, shouldn't the buyer bear the brunt of too-high? After all, they both voluntarily entered the agreement. There is nothing that places the seller in a more advantageous negotiating position.

Level?

Your condition is obviously not going to be satisfied.

fwiw I don't agree with the use of the word scam for this, and I don't think the OP is doing anything wrong.

but I do think the rigged structure of the marketplace makes it pretty fair to voice the opinion that it is overpriced.

Hypothetically, if he has a 26% ROI and sells at 25%, then it's not the worst deal in the world for investors.

An opportunity to invest in a 1-3 day event with a roughly 1% ROI isn't the worst thing ever, you could easily invest something like 0.05% of your bankroll in this.

Of course in reality, the opportunity cost makes it laughably unviable, since there are so many other good opportunities in the MP, and of course also the time frame is much longer than 3 days since you book so early and get money back so late.

Also, since most people have zero understanding of expected growth, they're basically almost going to over invest in themselves and their shares, which leads to an expectation of eventually being broke.

The deeper question is what should the standard function be to determine an equitable split of equity between buyer and seller? Does it vary based on purchase size and necessity? How should the marketplace function to determine this? As-is, it's definitely designed to be a seller's market.

That said, there are lots of people selling at 1.2 or 1.3 for events where they are probably NEGATIVE at that ROI, and those threads get a lot less criticism.

QFT

All--

This thread made me very curious, so I whipped up a simulator to attack these questions empirically. As I write this, a fake PrimordialAA, who has a 55% winrate on average against the opponents he faces, is playing 250,000 128-man HU tournaments.

This is a very simple program, but I'm confident that it yields a much better estimate than most of the ways in which people estimate ROI's. (Better players advance, so fake-Primordial's winrate decreases in a somewhat natural way as the tournament progresses. I used the log5 rule to estimate winning chances; I'm skeptical of the method but it ought to be close enough for these purposes.)

Some questions:

(1) Are people curious about fake-Primordial's ROI in these tournaments?
(2) I faked up fields by randomly giving people winrates between 40% and 60%. Is that close to realistic, or would another quick and dirty method work better?
(3) [Possibly embarrassing.] Can someone point me to the algorithm the WSOP will use to determine payouts? Right now I'm guessing.

All my best,

--Nate

about number 2

More than half the field will be above 50%, while there will be some very bad players with much lower winrates. I'd give the majority of the field between 45-57 with a couple massive fish under 40.

Could be wrong about that since I'm not used to the structure or field, I'm sure there are others who can say more accurately.

Also last year had 158 entries

I don't think anyone should have a 60 WR... I think similar to ibav there should be a handful of ~38 to 43s, then a handful of 50s and a bunch of 51-55s or something.

sounds interesting though, would like to see results

ibavly--

Thanks for the response.

I know last year drew 158; since this program is something I whipped up in an hour, it can only so far handle field sizes that are powers of two. If I have a few extra minutes, I'll try to improve on that.

You're probably right that the low winrates will be lower than the high winrates are high. I'll try to incorporate that somehow, but my guess is that spreading the winrates out in [.4, .6] isn't terrible.

All input welcome.

All my best,

--Nate

JDalla--

Thanks for the input!

FWIW, I'm not withholding results in order to bait people into begging for them and making me feel good; I'm furiously testing the simulator. I'd rather not release incorrect numbers that could cause Primordial to face unnecessary abuse.

All my best,

--Nate

All--

Do these look reasonable?

Winrates:

10 percent of the field: 39%
10 percent of the field: 44%
30 percent of the field: 50%
20 percent of the field: 52%
20 percent of the field: 54%
10 percent of the field: 55%

(For technical reasons these winrates might be something like .05% off, but they're very close.)

Payout structure:

1st: 25% of the prize pool
2nd: 15%
3rd-4th: 10%
5th-8th: .05%
9th-16th: .025%

Rake 6%.

PrimordialAA has a 54% winrate. (The distribution above is a probability distribution for non-Primordial players; it would be easy to run the sim with Primordial having a 53.5% or 54.5% or 36% or whatever winrate.)

Thanks for any guesses you can give,

--Nate

What percent of the time does a 55% player win vs a 52% player?

Fritzy--

I'm using the "log-5" rule (I linked to a document about it above). Again, I don't think it's that great, but Bill James has endorsed it, and I think it ought to be close enough for these purposes.

Using that rule, a 55% player will beat a 52% player ~53.012% of the time.

All my best,

--Nate

assuming the payouts should read 25, 15, 10, 5, 2.5 right?

Yes! Sorry! (FWIW, part of my little test suite for this simulator checks that payouts add to 100%.)

All--

The first run of the program (winrate ~= 54%; see first round winrate):

The second run (winrate ~=55%):

I'll keep cleaning this up and trying to improve it. I think the machinery is general enough to use for other questions too. Sorry if I find mistakes tomorrow; I sometimes make silly programming errors late at night.

All my best,

--Nate

EDIT: For the curious, here's the results for a winrate in between 54% and 55% (almost exactly 54.5%):

Why would anyone have a higher winrate in round 7 than round 6?

ZJ--

That appears to be an quirk of that set of tournaments. I just ran 100,000 more for the 54.5% (assuming that's the one you were talking about):

ROI: 29.602688%
100000 tournaments played.
Winrate in the first round: 0.54515
Winrate in the second round: 0.545171053838
Winrate in the third round: 0.537550471063
Winrate in the fourth round: 0.534051076615
Winrate in the fifth round: 0.52543366151
Winrate in the sixth round: 0.525987062235
Winrate in the seventh round: 0.519083969466

I'm going to sleep on these questions now. I'll return to the thread in the morning. Again, sorry in advance if I made any silly errors.

All my best,

--Nate

He clearly has the cluth gene

In before Primo ships the event

i created a formula for markup in which profit gets split equally between backer and horse based on ROI. just gonna copy paste what i wrote in my 90% finished book that i never got around to finishing and has sat around for 13 months collecting internet dust:



for online play i think its completely fair to split the profits equally between backer and horse. however, for live the player does deserve the lions share due to expenses in flights, travel, etc while the investor pays nothing for those things. i think for live play the horse should receive 2/3 of the future profits. i dont think higher would be fair based simply on increased risk of ruin for backer.

formula for 2/3:
2*[(ROI+100)*(1/X)-100] = (1-1/X)*(ROI+100)
2* [(ROI+100) / X] - 200 = (ROI+100) - [(ROI+100) / X]
3*[ROI+100)/X = ROI + 300

X = [3*(ROI+100)] / (ROI + 300)

ROI Markup
10 330/310 = 1.0645:1
20 1.125
30 1.182
40 1.235
50 1.2857
75 1.4
100 1.5
200 1.8


estimating ROI is the tough part of course, but at least this is a decent guideline to follow for online/live

Awesome post, Daut. Thanks

If the goal is that profit should be split equally between backer (more accurately: BAP'er) and horse, markup should be exactly 1, regardless of ROI. With any MU>1, an investor dollar will always have a lower ROI than a horse dollar.

For example if a 75% ROI player sells 50% of a 10k event @1.25 MU, the backer has an expected profit of (10k+10k*.75)*.5-10k*.5*1.25 = 2500 or an ROI of 2500/(10k*.5*1.25) = 40%. The horse OTOH has an expected profit of (10k+10k*.75)*.5-(10k-10k*.5*1.25) = 5000 or an ROI of 5000/3750 = 133%.

So either the goal is not to split profits equally, or I'm missing something?

Daut, it seems that if there are two investments with equivalent ROIs but A has a lower standard deviation than B, the buyer should capture more equity in B than in A. Would you agree? Therefore you can't just determine equity split with a raw ROI calculation, you need sdev as well.


That said, even if we can't estimate ROIs, we can at least align them a bit in packages which sell shares of multiple events. I see packages where people charge 1.15 for a 5k at the series and then 1.25 for the millionaire maker, and that just seems absurd.


Also, "needing to sell" matters as well, but that's an ex-post market-based adjustment I guess.

you wouldnt be the first person to believe that tournaments should never be sold at markup. i dont think markup is an outright scam in general, but i do think that a) people way overestimate their ROI in tournaments and b) they use a markup that is too close to their ROI in general.

but, youre definitely right that the investor/horse will never have the same ROI with markup. but my formula was based on the old school "ill put up all the money and you get half the money cashed" type deals from rounders and from stu ungar's 97 world series. i wanted to find a way to evenly split the profits of the tournament if someone put up all the money for a horse based on his ROI in the tournament since a lot of good players are too broke to buy in to a series with much if any of their own $.

perhaps the backer putting up all of the buyin as the base formula isnt accurate though. maybe people shouldnt play tournaments if they cant afford to put up a certain amount of the buyin.

lets create a formula that splits the profits 2/3 to horse, 1/3 to backer (since as you said, it is impossible for them to have equal profits due to markup) based on the backer putting up 50*x of the buyin and the horse putting up 1-(50*x) of the buyin.

cashes for ROI+100
backer puts up 50X
horse puts up 100-50X
backer gets .5*(ROI+100)
horse gets .5*(ROI+100)
backer profits = .5*(ROI+100)-50X
horse profits = .5*(ROI+100)-(100-50X)

2*[.5*(ROI+100)-50X] = .5*(ROI+100)-(100-50X)
ROI+100-100X = .5*ROI+50-100+50X
.5*ROI+150 = 150X
X = (.5*ROI+150) / 150

ROI markup
10 1.0333
20 1.0667
30 1.1
50 1.1667
75 1.25
100 1.3333
200 1.667

unlike my original 50/50 profits with horse putting up 0% of the buyin example, this one allows for markup over 2:1 when the horse has an ROI >300. but at the low ROI, it has much lower markups than my first example.


we can come up with any number of these formulas/tables, but whats really going to turn the tables is when we can start estimating ROI for more than just HU tournaments. based on large sample sizes of many great online players we can deduce that ROIs in general are smaller than people think. Even a boss like munchenhb who has a million dollar score online doesnt have a 100% ROI on pokerstars.