What's new
Fantasy Football - Footballguys Forums

Welcome to Our Forums. Once you've registered and logged in, you're primed to talk football, among other topics, with the sharpest and most experienced fantasy players on the internet.

Playoff Seed alternative algorithm proposal. (1 Viewer)

jaybny2

Footballguy
I'm a long time FFPC/FPC player, and have been trying to put my finger on something for a while. I've finally figured it out. I believe there is a slight problem with current playoff seeding.  Please hear me out.. 

First I will state my assumption as to the reason and motivations behind the current playoff seed algorithm. Then I will show the problem with the current algorithm, and provide an alternative algorithm that more closely reflects the afformentioned assumptions on the motivation. Lastly, I will show the hidden gotcha in the current  (seemingly simple and intuitive) system. 

A) Current Algorithm for deciding which 4 teams make the playoffs: 

  1. Team with best h2h record
  2. Next team highest total points
  3. Next team with best h2h record
  4. Next team with highest total points  


B) Assumed motivations and reasoning: 

Way back in the day, playoff seeding was exactly like the NFL. Top 4 team records (after tie breakers), make the playoffs.  This would sometimes cause the team with the highest number of points to miss the playoffs due entirety to bad luck. 

  • Current Algorithm ensured that the top 2 in points would make the playoffs. This makes the game more fair and removes some of the luck factors in head-2-head records. 
Head-2-Head scoring is still a core part of season-long fantasy, and we acknowledge the skill of setting lineups based on your opponents strengths,  and sometimes "stacking" your lineups with the QB from the same NFL team as your  your opponents stud WR..  

After the top in points and the top in h2h record are in the playoffs,  we then want to reward wins and losses over points. (Someone that goes 8-3 with 1800 points, should get  the last playoff spot over someone that is 6-5 with 1810 points!) 

  • Current Algorithm give the nod to h2h over points by seeding best records before best points. Seed 1 is h2h before seed 2 of points, and the same for seeds 3 and 4.
Idea was to keep the fun and excitement of h2h but still reward teams that have the most points in the league. 

 

C) Problem by Example: 

Current seeding created some weird and un-intended  side effect, that in effect give much more weight to total-points than to h2h record. 

Example1 shows how this is intended to work - Team5 (4 seed) makes the playoffs with 2nd highest point total. 

Example1: 
Team1  9-2  1800  (1 seed)
Team2  8-3  2000 (2 seed)
Team3  7-4  1600 (3 seed)
Team4  6-5  1500
Team5  5-6  1900 (4 seed)
Team6  5-6  1650    



** Now look what happens when we add 1 win to Team2  ***

Example2 shows how this was not intended to work - Team3 with third best record, loses to Team6 with 4th in total points!

Example2: 
Team2  9-2  2000 (1 seed)
Team1  9-2  1800 (3 seed)
Team3  7-4  1600  
Team4  6-5  1500 
Team5  5-6  1900 (2 seed)
Team6  5-6  1650 (4 seed) 


By just looking at the standings, Example1 and Example2 are almost identical. The only thing different is that in Example1, Team2 had 8 wins, and in Example2 , Team2 has 9 wins! 

D) Solution

Here is an alternative that will generate the same playoff teams in Example2 as in Example1. 

New Algorithm for deciding which 4 teams make the playoffs**: 

  1. 1st place team in h2h standing 
  2. 1st place team in points standing
  3. 2nd place team in h2h standing 
  4. 2nd place team in points standing  
  5. 3rd place team in h2h standing 
  6. 3rd place team in points standing  
  7. 4th place team in h2h standing 
  8. 4th place team in points standing  
** After ranking the teams from 1-8, the first 4 UNIQUE  teams make the playoffs. 

New Example1 shows the same 4 playoffs teams as before. 

New Example1 standings: 
Team1  9-2  1800  (1st h2h , 3rd points)
Team2  8-3  2000  (2nd h2h, 1st points)
Team3  7-4  1600  (3rd h2h, 5th points)
Team4  6-5  1500  (4th h2h, 6th points)
Team5  5-6  1900  (5th h2h, 2nd points)
Team6  5-6  1650  (6th h2h, 4th points)

New Example1 playoffs:
Team1 (1st h2h)  -    playoff seed 1
Team2 (1st points) -  playoff seed 2 
Team2 (2nd h2h)   
Team5 (2nd points) - playoff seed 3 
Team3 (3rd h2h)    -  playoff seed 4
Team1 (3rd points)
Team4 (4th h2h) 
Team6 (4th points) 




New Example2 shows the same teams making playoffs as in Example1!! As intented! 

New Example2 standings: 
Team2  9-2  2000  (1st h2h , 1st points)
Team1  9-2  1800  (2nd h2h, 3rd points)
Team3  7-4  1600  (3rd h2h, 5th points)
Team4  6-5  1500  (4th h2h, 6th points)
Team5  5-6  1900  (5th h2h, 2nd points)
Team6  5-6  1650  (6th h2h, 4th points)

New Example1 playoffs:
Team2 (1st h2h)  -    playoff seed 1
Team2 (1st points)   
Team1 (2nd h2h) -    playoff seed 2
Team5 (2nd points) -  playoff seed 3 
Team3 (3rd h2h)    -  playoff seed 4
Team1 (3rd points)
Team4 (4th h2h) 
Team6 (4th points) 



E) Gotcha

The current algorithm seems fair and straight forward, but it is in fact giving much more weight  to total-points over head-2-head. 

Look at Team1 and Team2 in Example2, they both have a 9-2 record. So why is Team2 the number 1 seed in the h2h standing? Its the tie breaker! The Tie breaker is Total Points already! By using total points as the h2h tie breaker, we just negated any advantage to choosing h2h in 1 and 3 seed. In fact, since there are ties in almost every league in h2h and never any ties in total points, we are in effect choosing the 4th highest in points over the 3rd best h2h record! 

Currently, there is a lopsided reward for points over wins, so much so that the correct strategy is to only draft and roster to maximize total points. If this is intended, then I would suggest we eliminate Head-2-Head all together and just use total points. However, if this was not intended, then I believe this new seeding will give us the intended results. 

-------------------------------------

Would love to know your thoughts and can explain further. I'm suggesting we make this change for 2017.

 
I think you did a great job of explaining it and I 100% agree that it would be the most fair method I have seen for playoff seeding.  I even think this would win out over victory points.  

Here are the problems I see.

1.  High stakes fantasy football players loathe change, and avoid it like the plague.  FFPC took a lot of flack over changing to their own webpage for league management when they got away from RTSports.  That has turned out to be the best move for the players, and hopefully it has worked out for FFPC as well.  At the beginning, it was chaos, and I am pretty sure they even lost a few players over it.

2.  Because of #1 (players hating change), Alex can be apprehensive about making changes.  I would call him stubborn, and he has every reason to be.  He gets suggestions and ideas about changes all offseason, and there is a point where I am sure he just gets fed up with the complaining.  That is not to say he (they) are not open to changes, but something as drastic as you suggest would be incredibly tough for them to explain to their core players.  I don't believe it would go over well, despite being much more fair and balanced.

3.  They have acknowledged that total points is weighted more heavily overall.  I think they have accepted it and moved on from other options.  Their dynasty format uses the victory points model, and they could have gone to that in the FFPC and FBG years ago, but they have chosen to stick with the current system.  I am sure that EVERY year someone suggests a new way to seed, but they stick with it because they don't like to mess with the format that has made them the biggest and best HSFF site. 

4.  I don't think they feel like they have the perfect system, but it is THEIR system, and they are likely going to stick to it.  

5.  No matter how good you (and others) may think your idea is, someone will find a reason to complain and a flaw in your system.  One player who would have been in this year but didn't get in because of the new system would whine that it wasn't fair... blah blah blah.   Right now, its just the way its set up, and it works pretty well.

These are just my initial thoughts, and I can tell you that I would love to see that seeding format enacted, but I don't see it happening.  You should definitely suggest it, but I wouldn't expect it to happen.  I would play in your league!!  Good luck with your idea.

 
I'm a long time FFPC/FPC player, and have been trying to put my finger on something for a while. I've finally figured it out. I believe there is a slight problem with current playoff seeding.  Please hear me out.. 

First I will state my assumption as to the reason and motivations behind the current playoff seed algorithm. Then I will show the problem with the current algorithm, and provide an alternative algorithm that more closely reflects the afformentioned assumptions on the motivation. Lastly, I will show the hidden gotcha in the current  (seemingly simple and intuitive) system. 

A) Current Algorithm for deciding which 4 teams make the playoffs: 

  1. Team with best h2h record
  2. Next team highest total points
  3. Next team with best h2h record
  4. Next team with highest total points  


B) Assumed motivations and reasoning: 

Way back in the day, playoff seeding was exactly like the NFL. Top 4 team records (after tie breakers), make the playoffs.  This would sometimes cause the team with the highest number of points to miss the playoffs due entirety to bad luck. 

  • Current Algorithm ensured that the top 2 in points would make the playoffs. This makes the game more fair and removes some of the luck factors in head-2-head records. 
Head-2-Head scoring is still a core part of season-long fantasy, and we acknowledge the skill of setting lineups based on your opponents strengths,  and sometimes "stacking" your lineups with the QB from the same NFL team as your  your opponents stud WR..  

After the top in points and the top in h2h record are in the playoffs,  we then want to reward wins and losses over points. (Someone that goes 8-3 with 1800 points, should get  the last playoff spot over someone that is 6-5 with 1810 points!) 

  • Current Algorithm give the nod to h2h over points by seeding best records before best points. Seed 1 is h2h before seed 2 of points, and the same for seeds 3 and 4.
Idea was to keep the fun and excitement of h2h but still reward teams that have the most points in the league. 

 

C) Problem by Example: 

Current seeding created some weird and un-intended  side effect, that in effect give much more weight to total-points than to h2h record. 

Example1 shows how this is intended to work - Team5 (4 seed) makes the playoffs with 2nd highest point total. 

Example1: 
Team1  9-2  1800  (1 seed)
Team2  8-3  2000 (2 seed)
Team3  7-4  1600 (3 seed)
Team4  6-5  1500
Team5  5-6  1900 (4 seed)
Team6  5-6  1650    



** Now look what happens when we add 1 win to Team2  ***

Example2 shows how this was not intended to work - Team3 with third best record, loses to Team6 with 4th in total points!

Example2: 
Team2  9-2  2000 (1 seed)
Team1  9-2  1800 (3 seed)
Team3  7-4  1600  
Team4  6-5  1500 
Team5  5-6  1900 (2 seed)
Team6  5-6  1650 (4 seed) 


By just looking at the standings, Example1 and Example2 are almost identical. The only thing different is that in Example1, Team2 had 8 wins, and in Example2 , Team2 has 9 wins! 

D) Solution

Here is an alternative that will generate the same playoff teams in Example2 as in Example1. 

New Algorithm for deciding which 4 teams make the playoffs**: 

  1. 1st place team in h2h standing 
  2. 1st place team in points standing
  3. 2nd place team in h2h standing 
  4. 2nd place team in points standing  
  5. 3rd place team in h2h standing 
  6. 3rd place team in points standing  
  7. 4th place team in h2h standing 
  8. 4th place team in points standing  
** After ranking the teams from 1-8, the first 4 UNIQUE  teams make the playoffs. 

New Example1 shows the same 4 playoffs teams as before. 

New Example1 standings: 
Team1  9-2  1800  (1st h2h , 3rd points)
Team2  8-3  2000  (2nd h2h, 1st points)
Team3  7-4  1600  (3rd h2h, 5th points)
Team4  6-5  1500  (4th h2h, 6th points)
Team5  5-6  1900  (5th h2h, 2nd points)
Team6  5-6  1650  (6th h2h, 4th points)

New Example1 playoffs:
Team1 (1st h2h)  -    playoff seed 1
Team2 (1st points) -  playoff seed 2 
Team2 (2nd h2h)   
Team5 (2nd points) - playoff seed 3 
Team3 (3rd h2h)    -  playoff seed 4
Team1 (3rd points)
Team4 (4th h2h) 
Team6 (4th points) 




New Example2 shows the same teams making playoffs as in Example1!! As intented! 

New Example2 standings: 
Team2  9-2  2000  (1st h2h , 1st points)
Team1  9-2  1800  (2nd h2h, 3rd points)
Team3  7-4  1600  (3rd h2h, 5th points)
Team4  6-5  1500  (4th h2h, 6th points)
Team5  5-6  1900  (5th h2h, 2nd points)
Team6  5-6  1650  (6th h2h, 4th points)

New Example1 playoffs:
Team2 (1st h2h)  -    playoff seed 1
Team2 (1st points)   
Team1 (2nd h2h) -    playoff seed 2
Team5 (2nd points) -  playoff seed 3 
Team3 (3rd h2h)    -  playoff seed 4
Team1 (3rd points)
Team4 (4th h2h) 
Team6 (4th points) 



E) Gotcha

The current algorithm seems fair and straight forward, but it is in fact giving much more weight  to total-points over head-2-head. 

Look at Team1 and Team2 in Example2, they both have a 9-2 record. So why is Team2 the number 1 seed in the h2h standing? Its the tie breaker! The Tie breaker is Total Points already! By using total points as the h2h tie breaker, we just negated any advantage to choosing h2h in 1 and 3 seed. In fact, since there are ties in almost every league in h2h and never any ties in total points, we are in effect choosing the 4th highest in points over the 3rd best h2h record! 

Currently, there is a lopsided reward for points over wins, so much so that the correct strategy is to only draft and roster to maximize total points. If this is intended, then I would suggest we eliminate Head-2-Head all together and just use total points. However, if this was not intended, then I believe this new seeding will give us the intended results. 

-------------------------------------

Would love to know your thoughts and can explain further. I'm suggesting we make this change for 2017.
Really well thought out post.  Too bad the guy with a forum dedicated to supporting his business couldn't take two minutes to respond.

 
Really well thought out post.  Too bad the guy with a forum dedicated to supporting his business couldn't take two minutes to respond.
I read it and appreciate Jay's time and his input. It's well thought out, I agree. If I was starting a new contest, I would consider it. The Footballguys Players Champion has been running since 2010 (it's sister contest FFPC Main Event since 2008), both using the current format in place. It's not perfect - no league format is - but literally thousands of players are enjoying the current format and we are not going to consider scrapping and replacing it, and attempt to reeducate these thousands of players to the merits of another system.

 
Last edited by a moderator:
I read it and appreciate Jay's time and his input. It's well thought out, I agree. If I was starting a new contest, I would consider it. The Footballguys Players Champion has been running since 2010 (it's sister contest FFPC Main Event since 2008), both using the current format in place. It's not perfect - no league format is - but literally thousands of players are enjoying the current format and we are not going to consider scrapping and replacing it, and attempt to reeducate these thousands of players to the merits of another system.
:lmao:   :lmao:   :lmao:

 

Users who are viewing this thread

Top