League Formula Drawbacks? (1 Viewer)

[efdenny]

I know league points formulas get discussed ad nauseum here, but... wanted get thoughts on the following formula, specifically what drawbacks there may be:

(number of entries / rank)

So for 20 entries:

1st: 20
2nd: 10
3rd: 6.67
4th: 5
5th: 4
6th: 3.33

and so on... TIA

 

[CraigT78]

I’ll start - too big a spread between 1st and 2nd imo. For a home game - is winning first worth THAT much more?

 

[efdenny]

Yeah, that was kind of the point... Two 2nds = one 1st, three 3rds = one first, etc

Just trying get the right spread. Right now, a guy with five fifth place finishes is ahead of a guy with two wins... seems wrong, but maybe not?

 

[Eloe2000]

This is my first time hosting a ranked "season" for our tournaments. We got together and voted on the ranking methodology and we have big spreads between the top finishers and I really don't like it. We made this a short season to test the scheme and I already am going to lobby hard for a tighter spread next season.

This is what we do for a 10 person tourney...

1st: 2xBuy-ins + Rebuys (10 players + 5 rebuys = 25pts)
2nd: 1.5x Buy-ins (10 players = 15pts)
3rd: 1x Buy-ins (10 players = 10pts)
4th: Number at which you busted (6th to bust = 6 points)
5th: Number at which you busted (5th to bust =5 points)
...
10th: Number at which you busted (6th to bust = 1 point)

 

[bsdunbar1]

How many players do you typically expect?
Do you allow rebuys or add-ons?

I guess Dr Neau's system is my opinion.

We have used many different ones over the years. None of them necessarily right or wrong since everyone has the same chance to collect the structured points.
But if you figure in the variable of Joe winning with a rebuy & 16 players vs Tom winning with no rebuy but 13 players, and Frank finishing 2nd both times but with an add-on etc etc, Dr Neau's is what we have settled on to manage all the variables of what should be better than what.

 

[APatHand]

For your consideration; a formula for assigning points based on the number of entrants and the number of spots paid (both being variable) in each tournament.
This formula has been refined over at least ten years. [Win a bigger tournament, get more points, win a small tournament, get fewer points.]

http://ceemeck.org/BARGETrivia/crunchy-points-and-the-best-all-around-player-baap.html#formula

 

[Gobbs]

I’ll start - too big a spread between 1st and 2nd imo. For a home game - is winning first worth THAT much more?

Agreed. I’m not a fan of the one first equals two seconds equals three thirds...method.

Also, for 20 players, I like giving points to the Top 5 or 6 players, not all 20 players.

 

[efdenny]

Thanks everyone for your inputs, I appreciate your thoughts!

How many players do you typically expect?
Do you allow rebuys or add-ons?

We are typically between 20-25 players each tournament, no rebuys or add-ons. Only 10 tournaments per year.

I guess Dr Neau's system is my opinion.

I've considered Dr Neau, but wanted a system that players could easily understand where they would need to place to potentially win the season, especially in the later games.

For your consideration; a formula for assigning points based on the number of entrants and the number of spots paid (both being variable) in each tournament.
This formula has been refined over at least ten years. [Win a bigger tournament, get more points, win a small tournament, get fewer points.]

http://ceemeck.org/BARGETrivia/crunchy-points-and-the-best-all-around-player-baap.html#formula

Thanks, I'll look at this a bit more too.

Also, for 20 players, I like giving points to the Top 5 or 6 players, not all 20 players.

That's actually a great idea... may help me out quite a bit.



Thanks again!

 

[BGinGA]

10/6/3/1 or equivalent based on field size

 

[TexRex]

For a freeze out, I find that the point spreads being the same percentage apart seems to work the best. If you are going to give points to say the top 5 or 7, lowest gets say 1.x those who just show. Each additional spot gets 1.x more than the next.

Example:

	Place	1.25	1.4	1.5	1.6	Fib
7​	8​	4.768​	10.541​	17.086​	26.844​	34​
6​	7​	3.815​	7.530​	11.391​	16.777​	21​
5​	6​	3.052​	5.378​	7.594​	10.486​	13​
4​	5​	2.441​	3.842​	5.063​	6.554​	8​
3​	4​	1.953​	2.744​	3.375​	4.096​	5​
2​	3​	1.563​	1.960​	2.250​	2.560​	3​
1​	2​	1.250​	1.400​	1.500​	1.600​	2​
0​	1​	1.000​	1.000​	1.000​	1.000​	1​

Winning a 20 player tournament is more difficult than a 19 player tournament. For measuring the number of players, I use 2 different things. The primary one is 1.04x the previous spot (an idea BG gave me). The other is a formula: # players/10. After testing both making all other criteria the same for several years, I found both worked and always found, when that was the only factor changed, that the results for our very top players were the same, meaning they finished in the same order using either formula. Since all I've ever looked for is the top 10 or so, that's worked well. I like the "BG" formula better though. My original formula was a variation of I think a Bluff magazine formula, but maybe it was Card Player magazine, I think very similar to what BG suggested above.

Having those two factors be consistent percentage intervals (getting progressively further apart) I think leads to better results. I've been using a Fibonacci sequence, and after a few places, it gets to a very consistent 1.618x. It's a naturally occurring sequence that is used to measure a lot of things. There are some variations in the first several places though. Whether they have a significant impact, I don't know.

I like Dr. Neau's formula in 2 cases. [1] You have a small group and a single table. His formula starts at the last place guy. The problem that leads to is when you have multiple tables (I'll use 3 tables w/10 players per table as an example), if those 30 start at different times, it gets quirky.

• If you start with 29, 4 are eliminated, and another player shows up (let's assume he was just late), he could get KO'd on the first hand and finish ahead of the 4 who maybe played an hour. Did he really perform better? He's getting more points.
• The 3 tables will play at slightly different paces, maybe even radically different paces. First KO is at T3, then 5 minutes later a KO at T2. Did the guy at T2 really perform better? He's getting more points.
• I honestly think trying to measure down that far is unreliable at best.

[2] I've modified his formula for use this year to test. We no longer do a league, but I use formulas to choose things like POY, etc. Ours are paper awards, but players have told me not to underestimate the value of bragging rights. I modified it by counting everyone who didn't make the final table as the same score. I've not even run totals and won't until after our Dec. tournament to see how it compares.

I tried measuring KOs. I did that for 3 years. They are easy to measure, and it was easy to see that over time, better players tended to have more KOs. In theory, if there were 30 players every time, 29 KOs would be equal to winning 1st place once if you measured points and KOs equally. By equally, I mean points count 50% and KO's count 50%. I accomplished that measurement by adding all players' points, then measuring the percentage each player got, and then doing the same for KOs. Then add those 2 percentages together. You could add points to KOs, but due to KOs being limited to those at a particular player's table, there is no real way for a player to collect all the KOs. If one player did get all the KOs except the last one, I think it would be silly for him to come out ahead of the guy whose only KO was the guy who got all the others, but finished 1st.

Even with large differences between players though, KOs made no difference in outcome of our top places. It might have more value in a single table tournament, but with 3 tables, even your top player in a night won't get that many. KOs get spread among all the tables. In 3 years, the only way it would ever have mattered in outcome among our top places would have been for all other factors to be pretty much identical, which never happened. So it's a measurable, but I couldn't figure out a way to make it really count for anything unless you had it count for a disproportionately large part of your formula.

BG might point out that not all KOs are equal. KOing a guy who has already been crippled isn't the same as taking out a "near peer" player where you would be crippled or out if you lost.

I'm always looking for new ways to measure players, so I'm interested in formulas others use.

 

[Geremie]

I use the following:

1st - 235
2nd - 175
3rd - 144
4th - 129
5th - 121
6th - 106
7th - 75
8th - 15

It is based on a 1000 point system where from the mid-point to the top increases at a decreasing rate and the mid-point to the bottom decreases at a decreasing rate. Both slopes from the mid-point have the same slope

 

[BGinGA]

I use the following:

1st - 235
2nd - 175
3rd - 144
4th - 129
5th - 121
6th - 106
7th - 75
8th - 15

It is based on a 1000 point system where from the mid-point to the top increases at a decreasing rate and the mid-point to the bottom decreases at a decreasing rate. Both slopes from the mid-point have the same slope

Difference between 4th and 5th looks whacked compared to everything else.

 

[allforcharity]

Difference between 4th and 5th looks whacked compared to everything else.

It's because the inflection point is at 4.5th place. The numbers could be transformed to a 100 point system by just dividing by 10 and rounding, rather than having to deal with 3 digits. But stats people love stats.

 

[BGinGA]

It's because the inflection point is at 4.5th place. The numbers could be transformed to a 100 point system by just dividing by 10 and rounding, rather than having to deal with 3 digits. But stats people love stats.

Understand. But still miserably fails the logic test.

 

[Geremie]

Understand. But still miserably fails the logic test.

Finishing 4th over 5th really shouldn’t be rewarded that much. Focus should be finishing in the top 3 rather than grinding out a 4th

 

[BGinGA]

Finishing 4th over 5th really shouldn’t be rewarded that much. Focus should be finishing in the top 3 rather than grinding out a 4th

And similarly, finishing 5th over 6th shouldn't be rewarded so much, either.

 

[Geremie]

And similarly, finishing 5th over 6th shouldn't be rewarded so much, either.

It’s more of a penalty to finish 6-8 rather than a reward to finishing 5th

 

[BGinGA]

As stated, fails the logic test.

 

[Geremie]

Also, finishing 1&8 and 2&7 will have the same cumulative standings after two games

 

[BGinGA]

Also, finishing 1&8 and 2&7 will have the same cumulative standings after two games

Which is also flawed. Player with a win should have a slight advantage.

 

[Moxie Mike]

Seeing it in practice might help.

As a point of reference, we use the 25,18,13,9,6,4,3,2,1,1,1,1,1... system. All participants get at least one point.

The top 5 through 2 events so far (with respective finishes):

1	Peachi Patty (2nd place; 1st place)	43
2	Crown Royal Dave (1st place; 4th place)	34
3	Nasty Nate (4th place; 2nd place)	27
4	Triple J (3rd place; 10th place)	14
4	Hitchcock Blonde Lavonne (3rd place; 10th place)	14

I encourage you to apply different point systems to the players' finishes and see how that alters the order. If you want more data, I'll send you the final standings from past seasons of the Moxie Poker Tour that you can play around with. Just send an email to mike@moxiemeninc.com and I'll hook you up.

 

[Frogzilla]

I know league points formulas get discussed ad nauseum here, but... wanted get thoughts on the following formula, specifically what drawbacks there may be:

(number of entries / rank)

So for 20 entries:

1st: 20
2nd: 10
3rd: 6.67
4th: 5
5th: 4
6th: 3.33

and so on... TIA

Entries / (1+ rank) is hard to beat. I stole this from Craig.

1: 10
2: 6.67
3: 5
4: 4
5: 3.33
6: 2.86

 

[Geremie]

Which is also flawed. Player with a win should have a slight advantage.

But a person who finishes dead last should be penalized equivalent to the bonus for the win

 

[Geremie]

Which is also flawed. Player with a win should have a slight advantage.

I guess I’m also focused on penalizing being first or second out as well as rewarding being a top finisher

 

[Mr Winberg]

I have never hosted a league so these are just my theoretical opinions based on zero experience.

1) Finishing e.g. 14th is no better than finishing 15th, so they should have equal points.

2) If applying the logic of (1) fully, then only ITM finishes should get points, which would be boring, so I would reward the paid spots + a few more. NOT the entire final table, though, since it can lead to stalling (fake tanking). Perhaps final table minus 1?

3) Make sure to reward that which you want more of. For example participation, give participation points to everyone, then give more points according to (2). You want reckless play, then have a top heavy points system. No? Then a flatter system. Etc...

Also, I wouldn't focus too much on determining what the "fairest" points system is, because in reality, the only fair points system which fully reflects how successful the players are is having the points be the ROI. Paid 100 and won 700, that's 600 points. Won 700 after 1 buy-in and 7 rebuys, that's -100 points. This would, IMO, be a horrible system, but it's the only one that truly represents the success of each player.

My 2 inexperienced cents...

 

[BGinGA]

the only fair points system which fully reflects how successful the players are is having the points be the ROI. Paid 100 and won 700, that's 600 points.

Also take field size into consideration and you have a solid idea (net_cash/field_size = points). I also think your first two points are extremely valid. We typically pay 25% of the field size and award points to the top 33%.

Finishing e.g. 14th is no better than finishing 15th

Make sure to reward that which you want

Exactly. Paying points to the entire field encourages tight play for the entire duration of the event, since every single elimination is a chance to lose/win points. I try to encourage solid poker play and reward performance, not for 'hanging in there" for an extra point or two before running out of chips and going busto.

 

[Mr Winberg]

I also think your first two points are extremely valid.

Thanks! Coming from you that means a lot!

Also take field size into consideration and you have a solid idea (net_cash/field_size = points).

The field size is already built in, since the more players the more prize money.
I don't really get having the field size in the denominator, as it reduces the points as the field size increases. Am I missing something?

Two big problems with ROI is
1) Weak players see how much they're losing, and
2) since participating will most likely cost you points, there will be scenarios where your best move is to not participate. That's nothing I would want to encourage...

 

[toynoob]

For those of you that run these leagues can you explain the end game? What does the winner get? I've seen free buy ins to a final tournament, wsop ticket, etc just curious how you guys implement it. I wanted to start one in July.

 

[APatHand]

points per event = 1000 * log (field size/place)

Place	20 runners	75 runners	200 runners
1	1301	1875	2301
2	1000	1574	2000
3	823	1380	1823
4	699	1273	1699
. . .	. . .	. . .	. . .

We limit it to the places paid.

If you want, you can add a set # of particiation points (Every mother's snowflake gets a trophy!) and the formula becomes

points per event = (1000 * log (field size/place)) + participation points

This.cannot.be.easier.

 

[Alex Lundstrum]

+1 for Dr neaus. It's a bit more complicated than some people will understand, but it works out really fairly. Also helps even score if you count top 15/20 points, etc.