2018 Forum H2H League & Pigeon Cup
- Thread starter JayH
- Start date
On the negative side, I wavered between Maxi and Medina and chose Medina thinking Maxi's goal rampage couldn't keep going. 11 points ain't bad though.
On the plus side, I put Villa on the bench, then switched in Isi when the lineup was released and Villa didn't start. Now that Isi scored I subbed in a $4.0 nobody for one of my forward starters and used the extra $6 to upgrade other positions.
On the plus side, I put Villa on the bench, then switched in Isi when the lineup was released and Villa didn't start. Now that Isi scored I subbed in a $4.0 nobody for one of my forward starters and used the extra $6 to upgrade other positions.
Left out Columbus in the original tweet, here's the updated list:
Last edited:
Our league is really fucking strong. I’m ranked 150 and sit in 4th place. Damn.
yessir lost a hundo. why? who knows. Maybe at his price he needs to generate more then 5 points-per-week on a rolling two week average to avoid going down. All I know is i played the money game early and now i finish weeks with 8mm+ extra in the bank. It's pretty stupid. No starting 11 is out of reach for me. I don't even get mad when i lose 500 on medina. Like sure his last game was bad, but he did well on wednesday, why is he losing 500, who knows?
¯\_(ツ)_/¯yessir lost a hundo. why? who knows. Maybe at his price he needs to generate more then 5 points-per-week on a rolling two week average to avoid going down. All I know is i played the money game early and now i finish weeks with 8mm+ extra in the bank. It's pretty stupid. No starting 11 is out of reach for me. I don't even get mad when i lose 500 on medina. Like sure his last game was bad, but he did well on wednesday, why is he losing 500, who knows?
Thanks for confirming
I am pretty sure the values are based on the average points/game per season. At the end of Week 3 Medina was averaging 7+ points/game. Since then his average points have dropped to less than 6 and his value goes down +/- $2M going from a 7+ point/game player to a 5+ point/game player.yessir lost a hundo. why? who knows. Maybe at his price he needs to generate more then 5 points-per-week on a rolling two week average to avoid going down. All I know is i played the money game early and now i finish weeks with 8mm+ extra in the bank. It's pretty stupid. No starting 11 is out of reach for me. I don't even get mad when i lose 500 on medina. Like sure his last game was bad, but he did well on wednesday, why is he losing 500, who knows?
I am looking forward to the Fall season when (I believe) all the values of the best players will stay high and the roster funds drop back to $100M -- it will be challenging and the top players will be dropping significant value when they have a 2 point day. Everything will be much more competitive.
Big Game Week 8 on-deck with #1 Malinska FC v. #2 Berget's BadAss Beard
Big Game Week 8 on-deck with #1 Malinska FC v. #2 Berget's BadAss Beard
Erm...him, i must break him.
Got the xFMLS projections up and running in a Google Doc link that should stay the same for the rest of the season:
https://docs.google.com/spreadsheets/d/197PS0D4OnsG7OYeEeNHsFzMsMFtg62Hi7utQHCFwBmI/edit?usp=sharing
I've reworked the model after the previous xGC-heavy version got too attached to low-lying midfielders. This edition doesn't fit the 2017 data quite as well for lower-point players but it's good with high performers. After doing some work on the matchup rankings I'm happy with how this week's projections are looking. Other improvements: it's finally got positions listed for LionNYC, price data that helped generate some fun value charts (see the spreadsheet tabs), and clean sheet projections to help pick GK/DEF.
Somebody please take out Berget, he's crushing everybody in the non-H2H league.
https://docs.google.com/spreadsheets/d/197PS0D4OnsG7OYeEeNHsFzMsMFtg62Hi7utQHCFwBmI/edit?usp=sharing
I've reworked the model after the previous xGC-heavy version got too attached to low-lying midfielders. This edition doesn't fit the 2017 data quite as well for lower-point players but it's good with high performers. After doing some work on the matchup rankings I'm happy with how this week's projections are looking. Other improvements: it's finally got positions listed for LionNYC, price data that helped generate some fun value charts (see the spreadsheet tabs), and clean sheet projections to help pick GK/DEF.
Somebody please take out Berget, he's crushing everybody in the non-H2H league.
Am I understanding you correctly that you are ditching a more accurate model for a model that you think will do a better job at winning?
Sort of. The regression fit on the latest model isn't as good for the full data set of hundreds of FWD/MID, but it's better with players in the top quartile, who are the guys fantasy players actually care about. Most importantly, it's got fewer weird outliers and less of an identifiable bias toward a type of player who's not very useful for fantasy purposes.
I think most of my work on the model for the next few weeks will go toward improving matchups, which are the real difference maker in fantasy but can be hard to predict.
I was just busting on you and being regressive. You should be tweaking the model to make it more useful and I think you are on the right track. Its great work.
Tried it your way. Can’t do any worse... thanks for the charts
I literally had my finger hovering over save to switch Russell in but decided to keep salloi. I liked the 15 minutes of what I saw from him when he played against us. Now I am annoyed and it’s only the first game of the week. I hate fantasy again.
So what’s the trick where people bench starters and then get them in ge lineup? Cause I have a feeling that’ll happen to me this week
Similar threads
