blonde poker forum

Interesting piece of research at 2+2.

A guy has looked at the total action of all the players who played the 50K.

http://forumserver.twoplustwo.com/29/news-views-gossip/wsop-results-top-players-24-11-roi-1223086/

Worth a look.

Cliffs: Guy calculates ROI of some of the best players over a completely insignificant sample size.

By this method if the series had ended after event 50, christopher brammer would have been declared close to the nut worst poker player in vegas

Next case...

FYP

It's not a league table, he's taking the players listed as a set of indistiguishable players and using that to take an roi over a more significant sample. At least that's how I read it.

Yeah but he has still just taken a mean average of everyones ROIs and used it to suggest that 30% is the ROI for the series that a 'top player' should expect. Its obv absolutely riddled with inaccuracies. There is some selection bias before it even starts because the list has been pulled from the picks in a high stakes fantasy league so whoever ran well in the last few years and won multiple bracelets is a lock to get in whereas the online sicko that has never had a live cash because he has only played 6 live events lifetime is never going to get picked when actually there is a decent chance that his true ROI is higher.

What we can infer with the highest degree of statistical significance is that the results of this study are insignificant when trying to calculate the 'true ROI' of players at the WSOP.

This.

Some of the mark ups being asked for in the 2+2 marketplace were LOL

For example Galen Hall asked for 1.61 in the 10k 6 max.

I'd be interested in this fields ROI in the 1k and 1.5k NLHE events.

Not sure where I got that this was the field for the 50k from.

id love to know the actual ROI of all wsop's for erik siedel

Oh dear. Well in that case I'd say it was a nice idea but very badly executed.

Yeah meaningless so much selection bias I don't know where to begin

Would be cool if someone wrote out 300 sickos and then took lots of samples of like 50 or something then looked at their results in all the 1500s/1000s. I don't think the 5ks+ are especially interesting, everyone obviously has a small ROI in them IMO with the exception of the main event. This is partly what makes his numbers a heap of shit, it's all skewed by the bigger buyins - if he used average ROI instead of $ ROI then it might be a tiny bit more interesting (but still not really)

i thought sharkscope tracked wsop events?

I wrote a few times in that Galen Hall thread. His markup was a joke and his attitude was even worse.

link? He comes across as pretty cocky

Yeah but if we had to have a decent sample size all the time then you could never come to any conclusions about live poker. This statement just let's young kids who don't have any flags spout shit about poker and get away with it. (obv don't mean you coz I respect your flags. Mean these idiots who think it's good to 5 bet jam light on grandad/women/ or people putting in there chips shaking like a shitting dog).

How's that for trolling coz they locked the Marc Wright thread before I could reply. :)

Let's assume a player has a large enough sample that we can say his ROI in a certain tournament is 50%. What would be a fair/good price for him to set for staking?

You are hard as nails JP :)

If on average he turns £100 into £150 in each tourney he plays then, by definition, a fair price would be 1.5 - he would charge stakers £150 for each £100 of tourney fees.

Equilibrium price would be 1.5

But how's that fair? Let's say the stake is split 50/50.

Player puts in £25
Staker puts in £75

Cash for £150 and take £75 each.

Player has made £50
Staker has broke even

I didn't say it was fair, just what is optimal for the seller.

Economics doesn't really concern itself with what is 'fair' since its impossible to define.

Its hard to put it into context for me because I rarely play MTTs where I think that my ROI is <50% but I guess if I did I'd go somewhere around 1.1-1.15

Gotta remember though that sellers have to set their markup at whatever price point the potential investors will pay given the investors perception of the players ROI. This may or may not be anywhere close to the players true ROI.

He has made expected value so the stake broke even

"or people putting in there chips shaking like a shitting dog"

 :)

Not all sellers and buyers are concerned with roi. Lots of people are willing to make a bad financial investment because of other factors and lots of people are not afraid to ask for an extortionate mark up to take advantage of the other factors :)

Ty dung beetle.  Not sure they were Mike Caro's exact words ( probably something like "fear the trembling bet"). But at least physical tells ain't changed much in 30 years. Joe Navaro is some sort of native Indian do he probably uses bears and rabbits in his explanation

But all the profit went to the player. The staker broke even. So the markup is too high? Why should the staker even bother at 1.5?

If we know for 100% that a player has an ROI of 50% in a tournament, then he should, imo charge AT MOST 1.2 : 1.

He cannot charge 1.5, that is ludicrous. If I buy 50% of someone in a £1000 tournament at 1.5 where the expected return is £1,500, then I will be paying £750, and expecting a return of.....£750.....what I am doing is I am paying the seller £250 so I can flip on the variance of his tournament play, sometimes he his full buyin, sometimes he wins 5 buyins or w/e so I'm just blindly punting on which time he wins/loses, this is not a profitable investment for me.

He on the other hand is investing £250, and expecting a return of £750, which is now an ROI of 200%.  This is such a common misconception that ROI should = MarkUp.

I think that the "fairest" way to decide would be to half the action, so you, and your investor get to spit your ROI equally, so our 50% ROI player enters his £1000 tournament, I buy 50% at 1.25,  costing me £625 and am expecting a return of £750, I make 20% on my money, which is a good investment. He invests £375 and is expecting £750 in return, increasing his ROI to 100%.

Then I think you should reduce the mark-up by about 20%, to account for the times you don't play your best, are feeling a little ill and so on. the 50% ROI you are claiming is based on your top game, it's rare you will exceed your own level of ability, but qiute possible you wont be 100% on it, so it's only fair to reflect this in the price.

Some of the mark-ups at this WSOP are jokes, people like Galen Hall are so deluded it's ridiculous. Galen is a really good player but the foundation for his own opinion of his actual value in a poker tournament is based on results that are hugely above EV.. Self addmissoin of >100 live tournaments and 2 years as a pro, meaning he could have played MAXIMUM 1 $100k PCA high roller and 3 PCA main events.

PCA fields generally accepted to be the strongest out there, so lets go mental and give him a 150% ROI  (which is literally IMPOSSIBLE) in the main and a 100% ROI in the High Roller (again not even close to being possible)

His total profit from 4 tournaments (likely to be at LEAST 5% of his lifetime sample) is $2,672,900 whereas his "EV" of those tournaments (giving him an ROI likely to be 3 times what his true ROI is) is +$145,000 so he has run over $2,500,000 over EV in what is about 5-6% of his lifetime sample. I know this is true of ANY tournament and tournament winner but the point that me explaining it clarifys is that piece taking in tournaments is nothing but a HUGE, HUGE PUNT.

Providing you're not buying pieces of Galen Hall in the toughest tournament in the world at 1.6 : 1 and you know a bit about tournaments + tournament players you'll almost certainly be getting the best of it, but thats almost irrelevant.

Good post Dave, and this is what I was trying to get at. I've not quoted all of it but am gonna mention a few other bits and end on a question.

There are a few reasons you might stake someone but ultimately it's done to make some money. The main reasons to get staked in my view (and it's why I ask sometimes) is to reduce variance and reduce the cost if the tournament is a bit too expensive.

I don't charge a mark-up when doing staking. (I think I did once but I never have done since, nor do I think I will again, unless my ability level skyrockets). I think I am +EV in the tournaments I play, but wouldn't claim to be one of the top boys. But seeing how my stakers are doing me a favour by essentially allowing me to play the tournament in the first place I don't want to eat at any potential profits they may have.

So I just find it a bit odd when I see people putting posts up at 1.5+. Quite a few people asked this for the main event. It's a soft field, sure. But it's also expensive, and a huge field, hence high variance. Given these are good reasons to be asking for the staking in the first place it seems unfair to me to be charging this much and essentially passing on the high variance to the staker.

Obviously supply and demand comes into it. If people are willing to pay a mark-up, then by all means charge one.

So, the question I was gonna finish on. And to keep it simple I'm going to consider the WSOP main event only. Obviously it's really hard to do, but any realistic estimates on what a top player's ROI in this event would be?

"But seeing how my stakers are doing me a favour by essentially allowing me to play the tournament in the first place I don't want to eat at any potential profits they may have."

Indeed.  I never stake players other than at par for exactly this reason.  Poker is the only industry where the person effectively providing credit lines gets charged interest!

If someone really is 1.5 over the field, you'd really expect that they are in the situation where they can win consistently over a large sample that they don't require backing.  If they haven't played a big enough sample then I don't see that they can justify the fact that they are 1.5 over the field.

But hey, if people are willing to buy people at 1.5+ (and this seems to be fairly standard in the poker community) then you'd be foolish not to sell at this price.  It's a free market - just my view.

No one will ever play enough live mtt's in their lifetime to prove there ROI is right or not.

variance is very different to ROI, Phil Ivey could not cash in the next 10 years worth of WSOP main events but that does not mean that his apriori expected ROI is not 500%+.

If a player puts a staking request out and it sells out then they priced it too low, if it does not sell out then they are too high.

Variance will obviously shape your actual ROI. For instance Phil Ivey's actual ROI (or anyone else who has made the final table/won it) in the main event is huge. He can play the rest of them until the day he dies and his ROI in the main will be big.

What i'm saying is let's say Phil Ivey plays the main event an infinite number of times. What might we expect his ROI to be?

"If a player puts a staking request out and it sells out then they priced it too low, if it does not sell out then they are too high. "

True. But this is a combination of things including forum popularity, reputation and being able to market themselves.  Someone selling out at 1.5 doesn't mean that they are 1.5 against the field.  In fact someone selling at 1.5 would mean that they are actually saying that they are even higher than this as the staker is taking on credit risk which would should be priced into the model (and will differ according to past actions of the stakee).

Very true and this is why Marc Wright sells at below his expected ROI due to perceived reputational risk.

I actually think Marc's prices are quite fair. his ROI is big AND because of his volume it's likely quite accurate, what Marc does which is the right thing is he makes sure people buying are getting a nice chunk of the roi as well.

Just so you know if you ever actually get down as low as having to tip the strippers in singles, you are defo good for 1.5 sell out round here.

fwiw what I posted is the mathematical solution to the problem and 1.5 is the price that maximises total EV. In that situation, all of the EV goes to the seller of course. That actually has very little relevance to tradition poker staking arrangements because the price (markup) is fixed which is a market imperfection.

To put it into a more poker-related context if my proven, real ROI was 50% in a tournament and I wanted to do an auction to sell for it then the final price would be 1.5

But why would anyone pay 1.5? What is in it for them?

Dan sold at 1.5. And People bid upto 2.0 on Keys on there wsop staking threads.  Pretty sure those that  bid/ paid were happy with their investments.

Actually I'm having a bit of a brain freeze as to why anyone would actually made the 1.5 bid, but getting to 1.49 is easy

If you know for sure that your true ROI was 50% and you ran an auction, investors are always going to bid higher because why would they let bidder A win a slice at 1.1 when they can bid at 1.11 and make a profit? This process continues until the price reaches 1.49 because everyone that bids is printing money.

I doubt very much it would.

if I am buying a piece of someone at 1.5, I would have to very confident his ROI was considerably higher, or else  there is no value for me.

piece taking in tournaments just inst a viable or strategically profitable endevour, so all the hours people spend working out who's mark-up is what, true roi is etc. to try formulate a system for piece taking is wasting their time.

It's a punt, you like the player, price seems good and you can afford it/feel like a decent little gamble, then you punt. That's how it works im afraid to tell everyone :D

Piece taking in cash games is however totally different and you can legitimately print money doing that.

If the price is there at 1.45 and you don't want to bid on something that has a real ROI of 50% then you are setting fire to money.

Two caveats to that would be (a) bankroll management or (b) opportunity cost if you think that a better spot is going to come along and gl finding that.

I'm not going to argue this point again because its an absolute bona-fide-oxford-english-dictionary-bank-of-england fact.

Ask a stocks trader if you don't believe me.

Finding out the true ROI of someone is virtually impossible. Working to 0.05% is unrealistic.

Plus there are other factors in play.. foreign exchange fees and a chance of being grimmed are 2 that spring to mind and I'm sure there are others.

If I think someone is 1.5, I would probably pay as much as 1.3 to cover margin of error and other factors.

Ha, that's almost a favour Keith, 1.25 max  :D

For the sake of completeness I read Cf's post as taking the assumption that we knew with 100% certainty that the players ROI was 50%

Obviously as our level of confidence in that number decreases then the maximum markup we would pay should decrease but if we know for sure that the players ROI is 50% then as I said, not buying at 1.49 is burning money. If you want to include stuff like grim risk and exchange rate fees (which should be pennies and zero respectively for a rational* investor) we might as well start compensating for plane crashes and air quality in the tournament room too.

*rational as defined mathematically, as in even if the horse binks a million you can still get it changed as spot because the investor has no desire to pay a premium to have all of the money right away in his bank account in his currency.

Obviously we have wires slightly crossed, you are talking from a purely theorectical and mathematical perspective and I am talking in the real world.

Obviously if I completely certain of a 1.5 ROI I would pay 1.49.

But it is impossible to know anyone's ROI in any indiivdual tournament.

And I am yet to find any bank which charges 0% fx fees!

I can assure you isn't talking from a theoretical and mathematical viewpoint

Dan needs to look up "Cost of Capital". Only a person with effectively infinite money could rationally pay 1.49 for a <1% ROI with the variance of tournament poker.

Better off backing 1.01 shots on Betfair.

What % of your bankroll would you stake on heads on tails if someone gave you 2/1 about predicting the correct answer?

(Obviously you know for sure it's 100% you aren't going to be cheated)

The last time I was at the Aussie Millions Roland offered m 5/4 red or black on the next flop.. What should I have bet?

... and risk premium in old school theory.

Buying costs should be deducted from any expected return, so that means the expected exchange costs from live poker clearly be in too.  This isn't the exchange rate movement which is a different thing.

..and grim risk should be factored in for some people.  If often somebody pays slowly, it would be reasonable to deduct for that too.

It is interesting Dan mentioned share traders, a lot of share traders wouldn't have a clue on the theory, they just "know" when to trade.  I expect Camel doesn't know the theory behind the price he quotes, just instinctively has a rough idea where it should be through experience.

Will give a better answer tomorrow.  Am up early in the morning.

 

There is a definitive answer to this question:

http://en.wikipedia.org/wiki/Kelly_criterion

That's good stuff.

I asked Tony Bloom the 2/1 coin flip question and his answer was 1% of his money.

Obviously the richer you are the less the utility of winning 2x your stake.

For Bill Gates the answer would probably be 0.

Yeah the mathematically correct answer (Kelly) is most likely way over what people would actually be prepared to gamble, it's sick variance. Making Kelly sized bets, you lose half your bankroll 50% of the time, so 7/8ths of your bankroll 1/8th of the time and so on, which most sensible people don't want to go through. Also, whilst underbetting kelly only has a small opportunity cost (making slightly less money than you could have on profitable bets), overbetting it leads to "gambler's ruin" and is a disaster in terms of long term expectation, so it's best to err on the side of caution. I had a friend at betfair that used Kelly/20, but he said Kelly/10 was more standard.

Not interested in 1.5 100% at 1.49. I think what David is trying to get at is some point where players ROI = backers ROI so markup should be sqrt(ROI) so for 50% ROI 1.5^0.5=1.22 would be a "fair" markup. My A level maths escapes me but I think variance would still be higher for the backer here. You can use a diminishing utility function within Kelly if you cared etc.

Good post.  Kelly suggests in Keiths example you should bet 25% of your bankroll which may be mathematically optimal but it is most certainly not life optimal.  I have used between half and a fifth of kelly based on my confidence in certain bets, stakes etc and have been broke a few times.  I would guess something like a fifth Kelly is workable but if I had a very big roll it would be less for sure.

That lecture must have been on a friday of a deepstack weekend ;)

I have a formula for staking, for selection of who to stake and how much to wager.

It's a sliding points system, derived from the following criteria.

1) How much do I like the person, out of 100

2) How much do I feel like he's "due" out of 100

3) How high has his morale been lately out of 100

4) Value at the price, out of 100

Anything above 325/400 is an auto buy for me. 250-325 very likely buying, if anyone has a particulally high 1) score Im often going to just buy anyways

How much to stake

1) How am I running recently - if it's 0-30 out of 100 = 3 points, 30-60 = 2 points, 60-80 = 1 point, 80-100 = 4 points

2) How Degen do I feel - 0-40 = 1 point, 40-60 = 2 points, 60-90 = 4 points, 90-100 = 7 points

3) Has the guy ever won before when I wanted to buy a piece but didn't, yes = 2 points, no = 0 points.

9-13 points = buy 3x what I can afford, 7-8 buy bigger than i should, 5 and below buy what I can afford, or maybe be a little nitty.

Pretty big loser buying pieces in tourneys lifetime, but I always get a good go for my money  :)

bookmarked

such a hero

:( and I thought it was a waste of time getting you to put me in the one drop. I just scored massive here

I would defo stake you in something as you literally smash every criteria + ive been losing lately and what a hilarious way to get out of it just sticking herbie in a live comp and collllleccccttttiiinnnnnnnng oioiiii

Lol Ty for the vote of confidence.

Love Lildave's staking criteria. I normally ask myself one question. 'have they sold out yet?'

<3

I either stake people because i like them and want to see them win, or, i hate them and it would tilt me to death if they won so i need some kind of financial compensation.

<3

Seem to recall you staking me a couple of times in the past

Hope I fall into category A!  ;hide;

Brilliant.

Lildave has already won the thread, and this is all just theoretical.

Cost of capital and risk premium are pretty much the same thing.  I am old school so will stick to risk premium, despite having done some fairly recent cost of capital calcs in the day job.  And I expect if I produce a massive spreadsheet model for the WSOP then it might just be a touch geeky as well as time consuming.

Risk premium is the the extra amount over a risk free return you need for putting up with the variance of a particular stake/investment.

If I am investing I can put all my money in UK government bonds.  I can be pretty sure that I'll get the return expected at the date given.  People don't assume it is as risk free as it once was, but it is a good enough proxy.

If I invest in shares I am going to get a return, but I don't really know how much it will be.  There have been various studies that show that historically I am going to get 5 or 6% over bonds investing long run.  You can also have a good guess at it by starting from total dividend yields factoring economic growth and inflation in.  I remember lots of people saying it was too high around 2000, but you can guess how that ended.

People who invest in shares are at the riskier end of the investing public, so I think 6% is a good start for what the risk premium of investing in poker players should be.  You can compare the volatility of returns against that from poker tournaments and I'd be reasonable sure that your average stake has a lot more variance that your average share investment. 

Probably a good example of an investors attitude to risk is considering something like deal or no deal.  You get to the end and you have two boxes, £1k and £250k.  How close to £125k do you get before accepting the bankers offer?  You take 100k and your "deal or no deal" risk premium is about 25%.  I'd say if you were holding out for 125k you probably have an unhealthy attitude to risk.  If you hold out for 130k, because you "like the sweat" or #lovethegame, then you should be making your way down Gamblers Anonymous.  I am serious on this.

Poker staking is probably something between investing in shares and doing the deal or no deal example.  My instinct is that the risk premium from poker should be something like 10%.  Of course everyone's different and it depends very much on how much risk you want and how big your roll is.  This is purely tournament variance and doesn't factor in foreign exchange expenses, timing issues with payments or the risk of default from the stakee.

So if someone is a genuine 1.25 player, I don't think you should be looking at it unless you are offered shares at 1.15 or so.  Take some off that if you think there is going to be some foreign exchange risk and some more for your default risk.  Adjust if they aren't a 1.25 player, and add a bit if the tournament is soft, or live is soft.  1.25 was chosen because I think that is around the benchmark for a good internet player (sharkscope the ROIs of the better internet players in MTTs).  The 2 plus 2 thread suggest live is 1.3 or so*, and there is another study referenced which also got 1.3.  I just think that your average $1500 live event isn't really going to be that much fishier than your average $50 online donkament.

I expect quite a lot of the perceived edge in the WSOP main event can possibly be negated by the increased variance you get from that tournament.

On top of this, I think you can also factor in what kind of player you are.  If you can invest in yourself and get 1.25 long run, why invest in your equal if he is charging 1.25.  One has an expected return of 1.25, one has an expected return of 1?

I'll just add that you can never really absolutely prove anything using statistics, but

Big sample>>small sample>>unbiased opinion>>biased opinion.

On 2 plus 2, I did notice somebody complaining vociferously about 2000 tournament samples, while running a staking thread which emphasised hero's performance over 6 or so tourneys.   

Just my views

* There are flaws with that 2 plus 2 thread, but I don't think the number looks that far out.     

 :goodpost:

This is a really great post. Other thing I would emphasise that many have little idea of is the mind-boggling variance of relatively small +ve EV instances. I don't have much poker-specific knowledge of this, though it can't be that different to certain sports-betting models I have run, and I would say to a man everyone I have met tends, naturally, to underestimate variance.

One thing worth mentioning is that the costs of staking are generally pretty low. Most people would pay about £10 every time they want to buy or sell shares. It allows people to make far smaller investments but still expect a decent return.

Without digging too deeply in Alex's new venture, will it:

Charge both sides of the deal or just the stakee?

Offer financial protection against being scammed?

He said yesterday that there'd been a hold up. Then I saw on Facebook a friend was pushing chipmeup with Tony G

Chipmeup site looks a mess. Surely Alex could improve on that.

Wtf just happened was reading this up to page 10 of 14 and now only 5 pages

http://blondepoker.com/forum/index.php?topic=58457.420