Stars SNG Leaderboard Questions

[kleinfeld]

Ok, I did some messing around with some formulas and came up with the following, and in my opinion, it’s pretty damn good, but let me know what you all think.
The following formula takes into account three things.
1) games played (G)
2) $ limit (L)… and,
3) Percentage of times in the money (P)

And here it is… G*(LOG(L/3)+0.2)*P… and the number you come up with is that players ‘score’ for that week, month or year.

So…. If one week a certain player played 300 SNG’s, at an average stake of $25, and made the money 42% of the time, the formula would look as follows…
G=300, L=25, P=.42
300*(LOG(25/3)+0.2)*.42
=141.22

You probably won’t be able to form an honest opinion of the formula unless you put it into an excel type program and play around with it, but if you can’t be bothered doing that, I’ll give some examples here, all of which are done for an imaginary weekly leaderboard…

250 games, make the money 33% of the time for varying limits…
$5, score 34.8... $10, score 59.64... $30, score 99.00... $100, score 142.14... $500, score 199.8... $1000, score 224.64.

By looking at the above scores it may seem impossible to beat the high limit guys, but that is untrue for a couple reasons. Firstly, I doubt 250 $1000 SNG’s are played on pokerstars on a weekly basis. So a low limit player should be able to play many, many more SNG’s than the higher limit guys. Secondly, a good player will presumably finish in the money a higher percentage of the time in the lower limits, increasing their ‘score’… To prove that anybody can win, I’m going to create an imaginary week for an imaginary GREAT player, then show a couple other ways that he/she can be beat.

Great Player… 370 games, average of $350 SNG’s, make the money 47%…. SCORE = 394.22
How to beat him…..
550 games, average of $50 SNG’s, make the money 52%… SCORE = 406.65
860 games, average of $10 SNG’s, make the money 64%… SCORE = 397.87

So, if this ‘great player’ exists, he/she will definitely be tough to beat, but it could be done…

… It’s 3am here in Australia, and I’ve said enough. I can’t believe I spend my nights doing nonsense like this, but I love it…. Let me know what you all think…

Kleinfeld… aka, MelbourneJoe.

[zipppy]

[ QUOTE ]

1) games played (G)
2) $ limit (L)… and,
3) Percentage of times in the money (P)

And here it is… G*(LOG(L/3)+0.2)*P… and the number you come up with is that players ‘score’ for that week, month or year.


[/ QUOTE ]

No formula will be perfect, but I feel like this stresses volume WAY too much, especially when you don't factor in any difference for 1st/2nd/3rd. For example, using your formula:


<font color="blue"> Person A plays 200 $30 tourneys during the month.</font>

<font color="red">Person B plays 1000 $30 tourneys during the month.</font>

<font color="blue"> Person A has an amazing month and WINS (1st place) every single toureny.</font>

<font color="red">Person B isn't very good...he places 2nd or 3rd in 250 tourneys, and OOTM in 750 tourneys.</font>

Person B scores more leaderboard points.

Obviously no one is going to win every tourney they play in a given month, but hopefully this illustrates why this formula gives too much weight to pure volume. Factoring in different buyin levels I'm sure would amplify problems further.

[zipppy]

[ QUOTE ]

I would like it though if, as someone else said, it's not strictly volume related. Maybe a points-per-game ratio with a minimum number of games?


[/ QUOTE ]

I was thinking about this, too. Could it be a leaderboard based on ROI with a minimum number of games played to qualify?

One (large) pitfall to having a leaderboard with a "minimum games required" is that it favors part time players, or players with multiple accounts. If someone runs well at the beginning of the month and doesn't need to continue playing, they can stop to preserve their high leaderboard rank as long as they've satisfied the minimum games played.

[Melchiades]

The leaderboard HAS to award volume. That should be pretty obvious.

[zipppy]

[ QUOTE ]
The leaderboard HAS to award volume. That should be pretty obvious.

[/ QUOTE ]

true...it's good for pokerstars and winning players.

[ QUOTE ]
The leaderboard HAS to award volume. That should be pretty obvious.

[/ QUOTE ]

Yea, I think it does even though that screws me as I'm a small-time player relative to others. I play maybe 100-150 SNGs a month.

However, maybe avoid making volume linear? Some sort of exponential function where someone that plays 100 SNGs isn't at a 10x disadvantage to a 1000 SNG player, instead maybe just 2x disadvantage?

Say points per game * f(volume), where f(volume) is logrithmic?

[zipppy]

[ QUOTE ]
[ QUOTE ]
The leaderboard HAS to award volume. That should be pretty obvious.

[/ QUOTE ]

Yea, I think it does even though that screws me as I'm a small-time player relative to others. I play maybe 100-150 SNGs a month.

However, maybe avoid making volume linear? Some sort of exponential function where someone that plays 100 SNGs isn't at a 10x disadvantage to a 1000 SNG player, instead maybe just 2x disadvantage?

Say points per game * f(volume), where f(volume) is logrithmic?

[/ QUOTE ]

I don't think pstars would do anything like this either. Players aren't going to want their 200th win to be worth less than their 1st for the month.

[blinlk182]

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
The leaderboard HAS to award volume. That should be pretty obvious.

[/ QUOTE ]

Yea, I think it does even though that screws me as I'm a small-time player relative to others. I play maybe 100-150 SNGs a month.

However, maybe avoid making volume linear? Some sort of exponential function where someone that plays 100 SNGs isn't at a 10x disadvantage to a 1000 SNG player, instead maybe just 2x disadvantage?

Say points per game * f(volume), where f(volume) is logrithmic?

[/ QUOTE ]

I don't think pstars would do anything like this either. Players aren't going to want their 200th win to be worth less than their 1st for the month.

[/ QUOTE ]

I don't think this is what OP is saying here. I think the suggestion is that you gain more points based on volume, not less. Something like for the sngs that I play, the points earned towards the leaderboard are multiplied by some volume factor equation.

The volume equation has to have an upperbound so that if someone plays an infinite number of games, there leaderboard points aren't multiplied by infinity.

In addition, it would also make sense to me that whatever the average number of sngs played by a person on stars in a month is, this number would equate to a volume factor of halfway between the upper and lower bound for the volume equation.

So, someone who plays a single game has their leaderboard qualifying points multiplied by a factor of 1 for volume.

Someone who plays an infinite number of games has their leaderboard qualifying points multiplied by some upper limit for volume. I think a fair upper limit would be something in the range of 1.5.

And someone who plays the "average" number of games has their leaderboard qualifying points multiplied by [lower bound + upper bound]/2. So if the lowest volume multiplier you can recieve is 1 and the highest you can receive is 1.5, the average player would get 1.25 in the volume portion of the equation.

[nuggetz87]

asking for a leaderboard - especially a SNG leaderboard - to reward skill over volume is really stupid. if you don't play very many just accept the fact that the leaderboard won't really do anything for you.

oh, you ran well for a 20% ROI over the 200 $60s you played this month? awesome. you shouldn't get anywhere near as many leaderboad points than someone who grinded 2000 $114s for a 5% ROI (not that you'd have made the same amount of $ as that person but you get the point).

i know you're just suggesting that the marginal worth of each SNG decrease the more that you play, but i think someone who plays 5x more than someone else should get 5x the leaderboard points.

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
The leaderboard HAS to award volume. That should be pretty obvious.

[/ QUOTE ]

Yea, I think it does even though that screws me as I'm a small-time player relative to others. I play maybe 100-150 SNGs a month.

However, maybe avoid making volume linear? Some sort of exponential function where someone that plays 100 SNGs isn't at a 10x disadvantage to a 1000 SNG player, instead maybe just 2x disadvantage?

Say points per game * f(volume), where f(volume) is logrithmic?

[/ QUOTE ]

I don't think pstars would do anything like this either. Players aren't going to want their 200th win to be worth less than their 1st for the month.

[/ QUOTE ]

No, the 1st and 200th would be worth the same and in fact they'd both be worth more due to a volume bonus. I'll try for a concrete example...

Take the sum of all points over the month/week (whatever time frame they use for leaderboard awards), then multiply the volume function against that sum.

So, player A plays 100 SNGs for a total of 500 raw SNG points.

Player B plays 200 SNGS for a total of 400 raw SNG points (B isn't as skilled as A, obviously).

Now, add in the volume muliplier. A's multiplier is 1.0. Since his volume is fairly low, every match counts as normal points. B's multiplier is 1.25. Since B plays so many matches, he gets a volume bonus, but not a 2x bonus to avoid volume dominating the equation.

A's total is 500 raw points * 1.0 = 500.
B's total is 400 raw points * 1.25 = 500.

So, both players score the same points, but B got there through volume while A got there through skill.

Finding the right function to get that volume multiplier is the tricky part since you want to reward volume, but not make volume so dominant (or pointless) that volume is all that matters.