2018 Forum H2H League & Pigeon Cup
- Thread starter JayH
- Start date
If you have a bench player who had a great game like
Rimil Salloi or Russell replace one of your starters with a DC United player who has a bye week and the bench player score turns green and counts toward your score for the week.I like how all 15 players almost always matter in this year's game.
I'm trying to figure out why Wolf Eikrem has the highest projected point total for midfielders in this round. He averages 2 points per appearance, and is ranked only fifth on his own team on that basis. He only played one full game (Round 3). Last week he played 2 minutes. Is it the xFMLS/96 that's pulling him up? I'm not particularly concerned about him per se, just trying to analyze your system. Anything you're willing to say without disclosing your secret sauce would be appreciated.
Also on a pure technical matter, can you tell me how you're getting stats to feed automatically into the back end of your model? I presume you're not doing anything manually. Thanks.
Sorry if I was unclear: xFMLS/96 is the projection. I've included FMLS/96 (a player's actual fantasy output per 96 minutes) in the spreadsheet for reference, so that users can raise exactly the kind of question you just asked about Wolff, but my projections don't look at past fantasy scores at all. They use stats derived from expected goals and some passing numbers to produce a per-96 point prediction that's been tested for fit against 2017 FMLS actuals. Those projections are then weighted by matchup strength and home/away to produce the xFMLS/96 R8 you see in the right hand column of the spreadsheet. This represents how many fantasy points the model expects the player to score this week if he plays the whole game.
To get a pretty good idea of why the model likes Wolff so much more than fantasy managers do, you can visit ASA's interactive tables and sort by per-96 G-xG (Wolff is second from bottom among players > 180 mins.) and A-xA (ninth from bottom). In other words, he's regularly been involved in good chances that haven't been converted.
The whole premise of the model is that for most players, expected and actual goals/assists/etc. will converge over the course of the season, and that at this point expected output is more predictive than actuals. Because finishing is kind of random, while general involvement in the attack and the ability to find and create chances are indicative of play style and skill, it's likely that guys like Wolff whose current FMLS-xFMLS is really low will start putting up more fantasy points if they continue to get playing time.
The catch is that this works better for some players than for others. For example, at this point last year, two of the three lowest per-96 G-xG guys were Joao Plata and Sebastian Giovinco, who had gotten some great looks but had only scored one goal between them. xFMLS probably would have told you that you should be playing both guys. As it turned out, over the rest of the season, Plata managed 7 goals—not a bad output, but still last place in G-xG. Giovinco, on the other hand, wound up with 16 goals and one of the highest G-xG in the league—from unlucky to maybe too lucky.
So the model is a good starting point, but the challenge for fantasy owners using it is to figure out which guys are Platas and which ones are Giovincos. Whose low fantasy scores so far have just been unlucky? Whose indicate something (team play style, genuinely poor finishing skill, surrounded by scrubs, etc.) that will continue to make him a bad fantasy option?
In Wolff's case, I haven't followed Seattle closely enough to make that kind of informed judgment call (to be honest, I've got him in my autoroo just so I can feel smart if a $6.2m player goes off for 12 points—which is pretty dumb when I could have had Johnny Russell). The biggest problem with him, as you noted, is that he's only played a full game once this season. If Schmetzer keeps using him as a sub, he won't be a reliable fantasy option no matter how good he is. Can't match your xFMLS/96 if you're not playing 96 minutes.
I load the stats that feed xFMLS by pulling a few CSVs from ASA each week. For the actual FMLS data, see here:
https://www.reddit.com/r/FantasyMLS...ne_know_of_a_quick_way_to_grab_stats/dvadigl/
https://medium.com/@paulgambill/how...eadsheets-in-less-than-5-minutes-a3fede1a014a
But if you just want to browse FMLS data without scraping it into spreadsheets of your own, MLS Cheatsheet makes it easy to use:
http://www.mlscheatsheet.com/cheatsheet.php
Last edited:
Question about the wink/wink.
If all of those guys had 2 point games, wouldn't you need a significant amount of cash on hand to upgrade the scrubs in your starting lineup? That's what I don't really get about the strategy. Are you assuming you'd only have to upgrade 1 or 2 of them if it didn't work? How much cash do you have on hand?
I had viable options on hand (or an equivalent amount of funds to have useful alternatives.). For example I had enough money for players like Seitz, the Kaku and Sean Davis range, and useful defenders as well. You can either have them on hand as placeholders or just use scrubs and make decisions later. Since I don’t need those players now I turn them into (or keep) 4 dollar players and upgrade elsewhere if I want.
So basically you leave money on the table for the optionality? Makes sense.
I'm having trouble keeping up this year. I've been pretty good in the past, but this is all too much.
I just cant remember to check as frequently during the weekend as I want to.
I'm such a bonehead -- I tried it this week and had Opara and originally Sinovic on my bench. I was smart enough to see that Sinovic wasn't playing and swapped him out for Amor, but COMPLETELY missed that Opara wasn't playing.
I've had the same problem. If you're on Twitter, set up mobile notifications from @MLSLineupNews—that's helped me.
Ughhhhhh come on man. Russell put me behind the 8 ball and I was trying to play counter to sabo so I avoided piatti and put bush in for the small chance at a nil nil draw with piatti doing squat. 16 min into the second game of the week and it’s alk over for me. Ahhhhhh. Raaaaggge!!! Feeeeeeeeeelinnnggggsssss!
Update:
Ahhhhhhh raggggeeee continues. You probably think that the bush pk save makes me feel better but I just feel worse. I chased differential points too early. Piatti was the play; he’s the whole freakin team. Bush save was a freebie that would have put me back to neck and neck. Now I just feel like I squandered it all too soon.
Update:
Ahhhhhhh raggggeeee continues. You probably think that the bush pk save makes me feel better but I just feel worse. I chased differential points too early. Piatti was the play; he’s the whole freakin team. Bush save was a freebie that would have put me back to neck and neck. Now I just feel like I squandered it all too soon.
Last edited:
Streaming the LAFC game and their fans definitely lucked out wth a good team and stadium but wow are their announcers absolutely terrible. It’s amateur stuff
I figured you would be chasing me after the Russell/Salloi game. At least you didn't bet on Houston being able to shut out Toronto's USL team. A lot of the experts picked Cabrera and got -4. This is becoming an interesting week.
So I’m basically dead now, so I’ll spill my beans.
Edit:
Psych late bwp goal. Not giving up yet!! Hopefully you didn’t not see my plan of action, regardless of how much of a long shot it is
Last edited:
I've got Houston's Elis on the bench with 8 points. His team scores 5 freaking goals, he's their No. 1 scorer, and he manages 8 points. Right on the bubble whether to put him in. I'm inclined to do so (over Savarino or Zardes), but this should have been a simpler call.
Last edited:
We definitely have a tough matchup together this week. I’m averaging 12 points per player and you’re right at 10. What makes it tough on me is that most of it came from your defenders.
I thought you don’t have a choice with Elis. You’ve already BYE week 3 players and your goalie is a DNP. That’s 4 of your subs there. Mullins will automatically sub Elis in.
I didn’t have Mullins when I wrote the prior post. I decided to add Mullins, which puts Elis in, and I switched in Zlatan.
If you average 8 or 9 points with your 3 slots in hand you can catch me if Lodiero doesn't have a great game.
I’m quit amused that I have a chance, but I’m also annoyed that I chased differential points so early and it’s put me off, I’m playing a lineup I didn’t intend.
I took some gambles that worked: avoiding Houston defense and doubling down on Dallas. I will admit I probably would have had 1 or more guys from Houston looking at that weak Toronto lineup. I feel I might have still played bush backup keeper hoping Montreal had a good home form with intent on playing Dallas keeper if he did poorly.
But on the other hand I capped bwp when I likely wouldn’t have, and avoided piatti, Ellis and some combination of Orlando offense that I wanted initially. All in hopes for bust games. Piatti clearly the heartbreaker miss. For Ellis, i settled on zlatan hoping for some kind goal/assist or brace. Didn’t pan out and cost me -6. Orlando could be a miss or a wash pending today’s games as I will be replacing those slots with Nycfc or Portland players. Hoping for a shootout.
If left to my own devices I probably would have had a similar lineup to you but finishing with less points via the Russell/salloi swap. But since your my H2H opponent and I had to react and I just might create a patchwork lineup that edges you out. I put my chances at 20%.
I really need a bust from lodiero. I was halfway thinking of dropping him but even if he gets 10 as your cap vs my normal, i could get lucky with 30pts from my unique remaining 3. In the unlikely case he does bust (which is the only real case I should want to drop him) and I keep him in my lineup then I might sub salloi’s 8pts on and roll the dice with 2 players. In a normal 1-1 or 2-1 match, it’s much easier to get two 10s concentrated around star players then it is to guess which mid tier player outperforms for that 3rd tenner. If he goes off for 15+, then I really need a shootout in the final match, but if i drop him in this scenario I’m probably done for as I won’t be able to find 50 points between 4 players from one match. So I’m going to keep him in, and pray for a Seattle redcard or something. But even if he does go off, it makes the nycfc game all he more interesting, which I suppose is the whole point of fantasy.
Last edited:
PIGEON CUP 2018
This week, Week 9, is the seeding week for the Pigeon Cup starting Week 10. None of the 12 spots are set right now. At the end of the week the Groups will be set like I announced last month:
Group A -- 1, 6, 7, 12
Group B -- 2, 5, 8, 11
Group C -- 3, 4, 9, 10
3 games Spring season and then a long wait before 3 games Fall season to set the 8 finalists (top 2 each group and top 2 3rd place finishers). We are going to use the official standings (PD tiebreaker) for seedings. But the Pigeon Cup standings tiebreaker will be based on the number of points your team scores.
This week, Week 9, is the seeding week for the Pigeon Cup starting Week 10. None of the 12 spots are set right now. At the end of the week the Groups will be set like I announced last month:
Group A -- 1, 6, 7, 12
Group B -- 2, 5, 8, 11
Group C -- 3, 4, 9, 10
3 games Spring season and then a long wait before 3 games Fall season to set the 8 finalists (top 2 each group and top 2 3rd place finishers). We are going to use the official standings (PD tiebreaker) for seedings. But the Pigeon Cup standings tiebreaker will be based on the number of points your team scores.
Similar threads
