Hey

New HU SnG proposal:

The idea is that you buy-in for a variant where you and your competitor play a predetermined amount of games (e.g. 10) and the one who wins the most games wins. First to 6 wins in this case or if it's 5 all you keep playing until a winner is decided.

I think the benefits are that for the better player it's a good way to build their roll quicker... Say you have a $1,000 bankroll, instead of playing say 10 $52 (assume $2 rake) games with an edge and win say 6/10 = $80 profit over 10 games. Instead you may conclude that you are actually a really good heads up player and over a larger sample (say 10 or 20 games) you'd back yourself to win a decent amount of the time.

Effectively your roll would allow you to buy-in at a higher level under my proposed structure whilst reducing variance seeing as the outcome is determined across a larger number of games.

So the new structure may allow your bankroll to afford the $200 + (10 x rake) level and if you win 6/10 you're up more than $80 (assuming $5 rake you'de be up $150 at the proposed variant $200 level v $80 at $52 under 10 single-game HU SnG's) and it can take you less than 10 games to make this extra money.

Obviously this idea could only work if you used it to increase your buy-in and therefore expected return (if you're a winner) if you believe you have an edge. At the same buy-in level it wouldn't be of benefit.

I just thought it could be a better way for the better players to build their rolls faster.

Do you think this idea has merit?? Obviously the rake charged would have to be considered.

Sounds interesting, but I was wondering if this would equally attract enough number of fish?

Good point mela. I doubt it would, not by a long shot i'd imagine. I do feel it would be a profitable variant for the poker rooms to put in place though. I think they could charge a slightly higher than normal rake given the structure.

By this i mean if they were to charge 10x normal rake then most games will finish before this 10 game point anyway so the poker room benefits. Not saying that they should charge 10x but hopefully you get the idea.

Would be good if you tie in this variant with the capability to challenge a particular player.

U know those times where you're playing someone who thinks they are great and just can't face the fact that u may be better than them. When they say stuff like "Fish" etc you could always say oh well then play me in game XXXX best out of 10 wins!! (at a higher stake of course)

If they accept this could be a profitable ploy.

Sometimes peoples ego's get in the way a bit too much. In which case them pre-booking a best of 10 match-up before realising they shouldn't hit the accept rematch button could make some extra $'s.

Stars would not integrate this future because someone accept a 10 match game and disconnect after the second game. Now He would loose the other 8 games or what?

this is a great idea, recreational players tend to favor games where the luck factor is reduced, in general they don't seem to like gambling, look at hyerpturbos, they are basically dead.

You are well aware that variance will increase this way?
Suppose you normally play $20 games. And now you decide to play 10 games vs somebody for a total amount of $200.
Normally if you win 4 and lose 6, you’d be down $40. (Disregarding the rake). In this case if you’d win 4 and lose 6, you’d be down $200.
So if you’re up against a complete fish, this approach is nice, but if the level is fairly equal, swings can become horrible. (In chunks of 10BI) So your bankroll should be way bigger as well. Or you should decide to drop in stakes, but than your hourly will be affected.

Is variance the right term for that? If you think of them in sets of 10 games with 1 buyin, then winrates would be higher for good players than in normal games, which would reduce variance. You'd swing for larger amounts of money, but saying that's more variance is like saying a player with 10% roi in $100 reg speeds is going to experience more variance than a player with 10% roi in $10 reg speeds.

That said, I'm probs totally wrong so someone who understands stats feel free to tell me how

I think its a good idea to establish a best of x games format.
Makes reg wars easier (if rake stays same or slightly higher) only thebest regs will be able to opensit them ( because weaker ones cant run after 1 game) makes the format more like a sport challenge overall, especially for hypers which dont take too long i think it will be a nice option.
IMO it doesnt need to be 10 games, but best of 5 or even best of 3 will be nice.

ohly trolling :P yer nice idea but i don't think that it won't catch on, fish don't like the idea that they might get stuck in a match vs a much better opponent and then have no way of leaving etc

Great idea, vote for me in the coming election and I will get this done for you buddy.

the idea is good if its not a winner take all format.
you win whatever matches the number of games yo won, same if you lose.
say you play 10 10$ game and wins 6 of those games. you would have won 10$.
idea is to favor adaptation to opponent by playing him in more matche than you would usually do. no one can escape. you are locked in a cage until the 10 (or 5, or 20) matche are finished, and the amount is withdrawn from you cashier at the beginning of the game.

Having game sets available to sit is not a bad idea, particularly at the higher stakes.

If done properly, it might allow the poker room to squeeze out more action and rake total out of two warring players if they each have to commit to a certain amount of games (and maybe receive a per game rake discount to give them incentive to commit).

This idea has been proposed in similar forms before, but the big downside is that it is not all likely that popular for a majority of players.

Just keep in mind that if you're playing this form with a winner-takes-all formula (which was the original idea), fish won't probably like this very much, as they quickly notice how difficult it gets for them to win some money.
So, is this a format you'd like to play against a fellow reg? In that case: prepare for nasty swings imo.

And don't forget other variance comes into play, and that's the order in which you win or loose games. Suppose you play 3 sessions of 10 games, and it goes 6/4 - 6/4 - 2/8. Player B won 16 out of 30 games, but he's still down a nasty 10BI. Player A won 14 out of 30, and he's up 10BI. Sure this is what you want?

This is why i proposed a first to x type setup. E.g. 10 game HU SnG would be decided first to 6 wins. And i like the winner take all idea best. I really think it would be a promising variant and would love stars to bring it in.

But i do understand the lack of attraction for the fish.

In saying this i'm pretty sure the reg's would smash me in this variant anyway. I just thought it could be a better way to increase the chance of more correlation between skill and profit. But if the player pool changes between normal HUSnG and this variant where fish don't take to it then this would be mitigated somewhat.

Best I can tell, if we had a 55% win rate (4.75% ROI, $5+.25) then we have a 59% chance of winning a best of 5. This gives us a ~1% chance of worse than -2% roi over 200 sets of 5. However, with a 55% win rate over 1000 games, we have only a ~.5% chance of -2% roi. That said, our ROI is 12.38% with the sets of 5. which would make our mean return much higher, lowering swings quite a bit after the first 1000 games or so.

I used ROI simulator for the simulation runs.

Are we always playing an odd number of games ?

If not, what happens on a tie ? Money back ? Play another N games ? A single tie-breaker game ? Who pays rake ?

Btw. speaking of variance, I don't think you can calculate that as it has been attempted to do so here.

Someone needs to look at the section starting with "Suppose that the observations can be partitioned" in http://en.wikipedia.org/wiki/Variance and also http://en.wikipedia.org/wiki/Law_of_total_variance and http://en.wikipedia.org/wiki/Variance_decomposition.

I'm way too tired to even try to read that, was about to leave to the TV.

My gut feeling is that you're going from the per-game variance into sub-groups, but it may not be an equal split (variance of the means vs means of the variances, something like that).

And I also have the gut feeling that this format will "reward skill" since you can't run away from better players.