Taking a shot at mid-stakes

[Leader]

[ QUOTE ]
Missing the point. I didn't say "Leader is a slimeball and will probably play like a spew hoping to run it up on a freeroll" I said the structure promotes this. There are lots of ways to structure a stake that doesn't do this. A very simple example is that (assuming you play 5/10 now) you play 10/20 and Veganmav takes half your action. If you need initial money to play you borrow it from him. Your hourly rate drops a bit because your edge is probably bigger at 5/10 but you gain the experience, Veganmav's profit/incentive is obvious, and you share the risk.

[/ QUOTE ]

The problem with this logic though is that lending me money is incentive for me to just keep it. I mean if someone going to be devious about it, that's a much simpler way no? So if your going to trust someone that far you mid as well trust them to play the best they can. In any case, I do understand your point, but I'm not sure if you understand mine. Whatever the structure encourages, you should only stake people you trust to do the right thing.

[ QUOTE ]
Any staking deal is better for the backer than the horse because the horse could probably achieve the same hourly at a lower limit unless he is totally busto. If you want to do the deal to gain mid limit experience than you sacrifice the hourly in order to do that. Maybe we need to understand the motivations of you both in order to structure it better.

-DeathDonkey

[/ QUOTE ]

The goal of the deal was for it to be +EV for both. You can do this mathematically though it's difficult. You get the staker's EV by multiplying the probability of a win times the staker's % of a win times the average win and adding that to the probability of a loss times the staker's % of a loss times the average loss. The staker's percentages are set by the parties. The probability of a win/lose is attainable using basic win rate calcs. The average win/loss is harder. You have to fine the mean of a truncated normal distribution. I found this paper pretty helpful in grasping the basics.

[mjkidd]

[ QUOTE ]

The goal of the deal was for it to be +EV for both. You can do this mathematically though it's difficult. You get the staker's EV by multiplying the probability of a win times the staker's % of a win times the average win and adding that to the probability of a loss times the staker's % of a loss times the average loss. The staker's percentages are set by the parties. The probability of a win/lose is attainable using basic win rate calcs. The average win/loss is harder. You have to fine the mean of a truncated normal distribution. I found this paper pretty helpful in grasping the basics.

[/ QUOTE ]

So based on all this statistical analysis the optimal staking deal is the one that knish offers Matt Damon after he goes busto?

[veganmav]

[ QUOTE ]
[ QUOTE ]

The goal of the deal was for it to be +EV for both. You can do this mathematically though it's difficult. You get the staker's EV by multiplying the probability of a win times the staker's % of a win times the average win and adding that to the probability of a loss times the staker's % of a loss times the average loss. The staker's percentages are set by the parties. The probability of a win/lose is attainable using basic win rate calcs. The average win/loss is harder. You have to fine the mean of a truncated normal distribution. I found this paper pretty helpful in grasping the basics.

[/ QUOTE ]

So based on all this statistical analysis the optimal staking deal is the one that knish offers Matt Damon after he goes busto?

[/ QUOTE ]

I was wondering how long it was gonna be till someone mentioned this.

Hey, man, let me stake you. Standard deal, you know.
Fifty percent of your winnings. If you lose, it's on me.
I'd just throw it away.
- You still got the truck? - Sure.
Come on.

[emerson]

Weekly settlements are much too frequent. It should be like quarterly. And then, don't split all the profits. Take out perhaps half the profits and split them. At the end of the year settle up.

20,000 hands is not close to a long run. With arrangements that short the investor should get a bigger cut. It should converge upon 50:50 as the sample becomes meaningful.

[Leader]

[ QUOTE ]
Weekly settlements are much too frequent. It should be like quarterly. And then, don't split all the profits. Take out perhaps half the profits and split them. At the end of the year settle up.

20,000 hands is not close to a long run. With arrangements that short the investor should get a bigger cut. It should converge upon 50:50 as the sample becomes meaningful.

[/ QUOTE ]

Well this thread has kind of become about staking. [img]/images/graemlins/frown.gif[/img] In any case it doesn't matter how much you settle if you don't reset the running total. For example:

Week 1: I win 400. 200 to staker (staker +200)
Week 2: I lose 800. staker pays 600 (staker -400)
Week 3: I win 1000. 700 to staker. (staker +300)

Instead we settle after 3 weeks. I win 600 staker gets 300. So were settling after 20K weekly settle ups are so the money doesn't get to out of whack with reality. Note that this is similar to what is done in futures markets with margin accounts. As the deal gets closer to completion, the amount each party has coverages to the end result.

Also I don't understand how it should converge on 50/50. This doesn't jive with how I understand the EV of this situation works. Seems as the deal gets longer the stakers % should go down to the point were he makes the same on his investment that he could get elsewhere given similar risk. 150K hand 50/50 win:cover loss stake on a decently winning player is hugely +EV for the staker. If your not willing to take less, you're missing opportunities.

[danzasmack]

As far as converging - you should just have a buyout option. Like you can buy out after you make him 10k (at which point you made yourself 10k)

gl dude. 15/30 NE 10/20 but 10/20 EQ 5/10 for some reason. Crazy kids.

that's all the advice i have for you. you should be fine, surprised you weren't playing these stakes already.

[DeathDonkey]

Ok you took my point about the proper strategy you should use when playing to maximize EV and equated it to theft, so I'll just be done on that front. My post was hopefully informative to Veganmav as its much more consequential info. for his side.

Regarding a deal that's +EV for you both, this isn't a tournament backing scneario where there are entry fees or other costs. If you are a +EV player at these limits than the deal will obviously be +EV for you both no matter how you split it. If you are primarily concerned about EV though, it sounds like your EV would be higher playing 5/10 on your own than 10/20 where you get half the wins unless you somehow beat 10/20 for more than you can beat 5/10 (bb/100 not hourly).

That really isn't a good reason to do the stake unless you are busto and the choice is between 10/20 for half the win and 2/4 or something on your own. Don't answer to this part I don't care if you are busto its just a way to describe that if you can play on your own its usually better to do that unless there is a consideration beyond EV.

-DeathDonkey

[Leader]

[ QUOTE ]
Regarding a deal that's +EV for you both, this isn't a tournament backing scneario where there are entry fees or other costs. If you are a +EV player at these limits than the deal will obviously be +EV for you both no matter how you split it.

[/ QUOTE ]

Well if I play an infinite amount of hands on the stake, it's true. But if I say play 100 hands, it's not. Average win and average loss are going to be pretty close to the same in that case. The absolute difference between those values diverges the more hands you play. As well as a win becoming increasingly likely.

[ QUOTE ]
If you are primarily concerned about EV though, it sounds like your EV would be higher playing 5/10 on your own than 10/20 where you get half the wins unless you somehow beat 10/20 for more than you can beat 5/10 (bb/100 not hourly).

That really isn't a good reason to do the stake unless you are busto and the choice is between 10/20 for half the win and 2/4 or something on your own. Don't answer to this part I don't care if you are busto its just a way to describe that if you can play on your own its usually better to do that unless there is a consideration beyond EV.

-DeathDonkey

[/ QUOTE ]

I really think you don't understand this situation. There's some probability that if I play 5/10 I lose over the next 20K. Probably between 5 and 20% depending on my winrate. If you eliminate that possibility, I would be a substantially bigger winner on average. Therefore, if on average I make the same, then it's a net positive. I could probably figure out what win rate I'd need to have at 10/20 with no loss allowed and 50% of the wins to have the same win rate 5/10 with losses allowed. Also, 15/30 is included in the stake so the average limit will likely be more then 2x 5/10.

You're right though. It's not all about money, but I don't think experience quite covers playing 10/20 or 15/30 for 5/10 money i.e. just having someone take 50% of your action for 10/20. If someone asked me if they should take that deal, I'd tell them not to.

PS I'm not insulted by any of your posts BTW. I'm sorry if I gave that impression.

[Heisenb3rg]

GL dude, just dont play stars... Dont want to deal with you there [img]/images/graemlins/laugh.gif[/img]

Or just dont let me know its you, or ill turn into a huge spewmonkey and dump all my chips over...

[Leader]

[ QUOTE ]
GL dude

[/ QUOTE ]

Thanks for all the good luck wishes BTW