JMC, I’m going to address your original questions on [1] attendance, [2] Kos, [3] rebuys, and [4] placement/cashing. I ran a league for a few years, but we abandoned it. However, I’ve continued to look at ways to measure players as we do give out some awards based on performance. I’m also just very interested in ways you evaluate something like poker players who don’t all play the same number of games, don’t all start at the same time, etc. I’ll start with a little philosophy (it’s mine).
Edited when I realized I mixed a system I looked at with the system I actually decided on for this year.
In baseball, there are 162 games. That’s a long season, and it’s enough to truly evaluated teams. But different parks play differently. It’s a long enough season to measure park differences. In the NFL, there are only 16 games. In college football, there are 12-13 (now counting 2020). The BCS has introduced us to the “13th datapoint” concept. At least they think 13 games gives a better measure than 12, and generally I agree, but I don’t think that means simply playing 13 games makes a team better. They do apparently believe that when it’s close.
Golf and car racing don’t have the same number of events per player. Car racing has racers in different positions, and that’s because it would be impossible to have 30+ racers in a round track and have everything be equal in position. In golf, a guy in theory could start 5 hours later than another and still compare them, even though the first guy might have played when it was colder and the ball didn’t travel as far and the second guy played when it was windier and perhaps even raining a bit. These two sports don’t truly have championship events.
Poker has more variables than all of those sports. To me, I think the more things you can find a way to effectively measure, the better your system is. Some things are easier to measure than others though. It’s “championship” event, if there is one, is the WSOP Main Event. It crowns a “world champion,” but in theory someone could enter only 1 poker tournament and win it. An NFL team couldn’t play one game and it be the Super Bowl.
BG has some good ideas on this, and a lot of league experience. He and I agree on way more than we disagree on. Hopefully he will chime in here.
On ratios, I’m a believer in having a consistent ratio between places. The more people you measure, the more I think that matters. There are plenty of ways you could set points and measure the number of players most of us are trying to measure.
Attendance
It seems to me there are 2 aspects of this. The first is how many games a player attends and the second is field size.
I say if one attends, they get something just for showing up. I’ll use 3 tables of 10 since that would be our max. I’d give every attendee 1 point. That one point is multiplied by field size. I start with 10 = 1.0. I’ve tried 3 things. Then they receive that number of points for attending. Field size divided by 10 is the multiplier. At 40-50 players, I’ve not run a scenario where this made a difference in our top players, but I don’t think that is the best way to do it. What I don’t like about 20/10 is for attending a 20-player tournament, you receive 2x as many points as a 10-player tournament, and 3x for 30. I’m not convinced that is right. As the field size increases, the differences become smaller and smaller.
BG introduced me some years ago to a percentage increase for each additional player in the field. At the time, he used a 4% increase for each additional player. He has also used 5%. Thus, if 10 = 1, for 11 players, the attendances points I’d give are either 1.04 or 1.05.
The reason I measure attendance with some kind of points is because it makes it possible to measure all players who attend. I personally am only looking to find our top 3-5 players, but I’d like that to be as accurate as possible.
I think 1.04 or 1.05 are both good numbers. However, this number must be considered in light of other scoring.
Knockouts
I think knockouts (KOs) is a skill, but it’s hard to measure. I’ll compare this to outfielders in baseball. Those who throw out more runners on the bases have better arms. But over a short run, that might not be the case. The percentage of runners thrown out per attempt is not a good measure because the better arms might be trying to throw out way more, but get fewer of them because a guy with a lesser arm isn’t trying the more difficult ones. There are so many variations of possibilities, it’s hard to truly measure beyond raw numbers.
It's hard to measure a quality KO in poker. There’s a huge difference between putting 10% of your chip stack at risk to put someone all in and putting 98%, where losing the hand would not knock you out, but would cripple you. It’s not the same KOing a guy who just got crippled in a bad beat vs. a guy who has frittered his chips away through sloppy play. But the KO counts the same.
I tracked this for a while with our league. We only had one year where altering the amount of credit given for KOs would have made a material difference in an award, and it wasn’t our top award. In that case, the scoring for that would only have made #3 the #2 in a category. So I found it difficult to measure where it made a difference.
My advice: I’d throw that out. As much as I played with it, I didn’t find that to be helpful in determining our best players. I did find it helpful to know that as a rule, better players have more KOs, but there are a lot of possible reasons for that. We have one player who rarely KOs anyone, but he consistently survives until late in the tournament. Survival is a skill too. His low number of KOs never altered his ending position, regardless of how much I counted KOs. But he also wasn’t at the very tops.
Rebuys
These are hard to measure as well. It seems to make sense that a rebuy counts as much as an original buy-in, but I don’t believe it does. A guy shouldn’t only get credit for where he finishes divided by buy-ins. I think if you are going to measure this, it shouldn’t be 1:1. Maybe a way to do it, I’m experimenting with this and don’t yet know how it will work. And COVID has caused us to cancel 5 games of our scheduled 14, so it will be a while before I really know. But I’ll share my thoughts.
I toyed around with an adjustment based on $5 increments. For each additional $5 on the rebuys, I’ve played with an adjustment of 1.5%, 2%, 3%, and 4%. Obviously you could do something similar for $10. We have a mix of mostly freeze outs but some rebuys. I applied these rules to past years. In our short seasons and relative low buy-ins ($35 currently for most games, but double that for 1 game), I found with a $60 buy-in at 4%, a guy who rebuys would have his final score adjusted by only 16.4% if you went by $5, or 8.5% if you went by $10.
Again, I don’t know for sure, but I think this ratio needs to be matched up with other ratios. I played with this concept on last year’s results. It didn’t change any slots near the top, which is really the only place I care about. I really don’t care if it changes from about it if it changes #10 and #11. At that point, I’m not measuring our best players.
One of your questions is should the 4 big events count twice as much. I don’t think so. I’d count them by some type of ratio above. I have 1 big event with a 2x buy-in that is also a re-entry event; 3 other rebuy events with the rest being freeze outs, and one event this year that will be a special amount – not quite 1.5x our buy-in. That is now looking like it will coincide with our big social event of the year, so I’m not sure yet how that will work. But, I have a formula to determine how to calculate the amount of buy-ins so that my biggest events will count for more, and rebuys will hurt a little, but not proportionally to the total buy-in. If next year is normal, ask me in about 1.5 years how I think it worked.
Placement/cashing
You put these together, but I’m going to bifurcate them.
If your finish points are identical to the payouts, you can put them together. For a lot of reasons, I don’t do that. I have a higher percentage of low payouts – either just getting their money back or a little more for the lowest spots. I do that because I’ve found it is easier to keep players who are consistent donators if they do sometimes cash. The fact that they cashed and won a little something is more important to them than their overall results. Thus, our payouts are top-loaded and at lower numbers of players, I pay out a higher percentage of the field. My lowest payout is determined by the number of players it takes to pay out one more spot where the last paid spot gets their money back. Some look at it as if they paid, but get anything back, they won. That’s not wrong; it’s just not how I do it.
I went from 10 at the tables to 9 last year, but I’ll use 10 so you can see the scoring.
Make the final table, and you get 1.6x the points as a guy who didn’t.
The top 6 spots each get 60% more, or 1.6x, the spot below.
11 = 1.0
10 = 1.6
9 = 1.6
8 = 1.6
7 = 1.6
6 = 2.560
5 = 4.096
4 = 6.554
3 = 10.486
2 = 16.777
1 = 26.844
In short, you get extra points for making the final table, but the final 6 get additional points. I do that regardless of the number of payouts, which varies from 3 to 6 for up to 27 players.
The problem with going by payouts instead of finishes is that it skews results between tournaments where there were only 3 payouts vs. one with 6 payouts. Thus, the final 6 are treated, in terms of scores, like we always payout 6 slots. However, when combined with attendance, one does better with a large field than a small field.
To further complicate this, I normally have 9 players at each table. Attendance at our 4 games this year has been 18, 14, 9, and 9. The last one was June, where I only allowed 7 per table for COVID. So do I count the 2 who didn’t make the final table like I normally do, or do I count them like I did our 3rd tournament where it was treated like all made the final table? I chose to do treat it like all made the final table based on attendance, but I prefer to score based on what actually happened.
Zombie’s System
I like some of the ideas Zombie has, even if I probably won’t take time to play with them. By the time that would be helpful to me, I’ll have to remember to email him and ask for a recap. I might be 2 years before I relook at this simply because this year is so weird.
What I do when possible is take other systems and apply it to our group for a year, then compare the results. I do agree with him that if you want to encourage rebuys, deducting points might discourage that. I think both of us have a big tournament that anyone can attend. We don’t have a championship game or a tournament of champions event.
I also think his finishes, and he uses 6 places like I do, are in the acceptable range. While the points use smaller percentages as they go up, a concept I don’t hate but don’t particularly like, I think it likely wouldn’t matter in terms of end results for the number of players I’m trying to evaluate.
I agree with his concept that playing to get in the money being an objective. One difference between our games is he I think has all rebuys, but we don’t. And I have some that are technically reentry events and not rebuy events.
I agree with StatTracker’s thoughts, though I’d do it a little differently than he suggests on the finish points. His appears to be a linear system, and I don’t like those. He does make a great point that ratios between things like attendance, KOs, and points are important. I’ll give my thoughts at the end.
Final Thoughts
I ask people what they are trying to measure and why they are trying to measure it. I think you have laid that out. That makes answering your questions much easier to me.
To me, attendance counts for something. In our case 1 x total attendance.
I don’t count KOs, but have counted them as much as 20% of the total score.
Finish points, +60% for each additional slot, and +60% for making the final table. That rewards making the final table (FT). However, the smaller the number of players, the less valuable it is to make the FT. It is not possible for someone in our system though to attend 14 games and never get in the final 6 and beat a guy who came once and won.
Rebuys count the same way attendance does, but based on either $5 or $10 increments. I’m using $5. So a $60 buy in and a rebuy for a total of $120 would basically count 1.5% more for each additional $5, or 12 x 1.04 per time. However, I won’t do that calculation until the end of the year. That’s important for the example below.
Here would be a formula based on my thoughts:
(A) Attendance = 1.0 x Total Attendance Factor
(FP) Finish points
Formula for FP -- A (Attendance x Attendance Factor) x Place Points (see chart above)
(BI) Buy-ins and Rebuys = 1 or 10%, but based on increments of $5 at 4% increase for each additional $5.
A x FP / BI = Pts
If I added KOs, it would probably be the same as attendance and/or buy-ins.
Here is an example our Main Event, 18 players, and a re-entry event. At the end of the rebuy period, players had the option of trading in their existing chip stack for a new one by re-entering and effectively starting over. It was weird that we had no true re-entries. In a re-entry, a player doesn’t have to decide to re-enter immediately, and they draw for a seat. They are unlikely to get their same seat back. For a player turning in their chips and getting a fresh stack, they keep their seat. One player chose that option, so that made 19 buy-ins. She actually chopped for 1st place. I’ll use the players by a letter, and their final score for the event. To the far right in bold, italics, and underlined is what the payout would have been without a chop. There were 63 BB left in the game, and in 5 minutes, that would have been down to 38 BB. The spread between BB left at the chop was approximately 23/20/20. At that point, it was a luckfest. Players chopped because of the differences in scheduled payouts.
Note: In our Main Event, the BI was not calculated. I anticipate calculating the BI at the end of the year for all events, and will see whether that makes a material difference in our top players. Therefore scores for the Main Event don’t reflect that factor.
1st A $430 cash, $290 profit – 25.670 pts. (reentered, so actually paid $140; all others paid $70) -- $520
1st B $350 cash, $280 profit – 25.670 pts. -- $390
1st C $350 cash, $280 profit – 25.670 pts. -- $220
Formula for 1-3 – A x FP = Pts
(38.207 + 23.879 +14.925)/3 or 25.670 for each of the 3 chopped places
4th D $120 cash, $50 profit – 9.328 pts.
5th E $80 cash, $10 profit – 5.830 pts.
6th F $0 cash – 3.644 pts.
7th-9th G/H/I $0 cash – 2.277 pts.
10th-18th All others -- $0 cash – 1.423 pts.
B and C had slightly different chip amounts at the chop, but it was less than 1 BB. In fact, if you averaged them, they had the same rounded number of BBs. A had more than either by about 15%. Because she had bought in 2x, B & C agreed to effectively give her the second $70 back, and then reward her with $10 more for having the largest chip stack at the chop. The payouts amounts between her and B and C were much bigger than the difference between chip stacks. Basically two players who over time have proven to be better players than her agreed to this rather than risk going home with less.
I think the agreed payouts make great sense because all profited at better than third place money, but not quite at 2nd place money. Had the lady who rebought finished 3rd, she would have had an $80 profit instead of a $290 profit. The other two could have had swings of up to $300, and it seemed unlikely to last more than another 20-40 minutes.
I think the point chop makes sense. To me, when players do this, they are stopping the war by armistice, not annihilation or surrender. All three can claim to be the champion; all got a nice profit for the just under 5 hours of play for the buy-in.