Race for CCL Spot 2019

Based on the Language in the CONCACAF Document, Have We Qualified For the 2019 CCL? - Moot Poll Now

  • Yes - We Qualified

    Votes: 28 84.8%
  • No - We Did Not Qualify

    Votes: 5 15.2%

  • Total voters
    33
I decided to take a crack at projected cumulative based on the remaining h/a analysis I've been doing and came up with the following:
  • Atlanta 129
  • NYCFC 120
  • NJ 117
  • Portland 111
  • Dallas 109
  • Columbus 107
I know my analysis isn't perfect, but the fun thing is this projection will update as results happen (for example, a DC loss at home tonight would move NJ's predicted points to 116). Also, my analysis does not factor at all any injuries or current form, it only factors in each teams H/A form for this year.
I recommend you keep track of these projections, not just here, but record them as hard entries on your sheet so you can see how wrong or right they were. I think you have a decent 1.0 version of something, but it might need to get to 2.x or even 3.x to really be predictive. Right now, I just don't see Atlanta pulling away like that.

I'm not sure my predictions are too accurate, as they probably should only use 2018 ppg for the projection to end of season, as 2017 numbers really have no bearing on how these current teams will finish up.

I also agree with mgarbowski mgarbowski that ATL isn't going pull away and also average 2.5 ppg over then next 10. That's crazy!
 
  • Like
Reactions: SoupInNYC
I recommend you keep track of these projections, not just here, but record them as hard entries on your sheet so you can see how wrong or right they were. I think you have a decent 1.0 version of something, but it might need to get to 2.x or even 3.x to really be predictive. Right now, I just don't see Atlanta pulling away like that.
I agree, I don't think it's going to hold to this either. I think Atlanta will drop more than 4 points over their last 10 as well as (and I haven't shared this yet), I think DCU isn't going to do as good as 29 points in their last 14, and I think TFC will do better than 8 points in their last 11.

I haven't had a whole ton of time to think through how to tweak it to be better, but here is what I have for this season:
upload_2018-8-15_12-29-14.png
 
  • Like
Reactions: adam
I agree, I don't think it's going to hold to this either. I think Atlanta will drop more than 4 points over their last 10 as well as (and I haven't shared this yet), I think DCU isn't going to do as good as 29 points in their last 14, and I think TFC will do better than 8 points in their last 11.

I haven't had a whole ton of time to think through how to tweak it to be better, but here is what I have for this season:
View attachment 9048
I think the problem is that every (or at least most) additional factors added to a prediction model introduces smaller sample sizes.
Start with my basic PPG projected points chart that I post every week. That is a blunt, single factor model, taking PPG to date and multiplying by 34. It ignores schedule difficulty, Home/Away to date, remaining H/A, current form, and all the other things people add to predictors to fine tune them. I originally created it in 2015 to predict where the playoff line would fall, which is something we currently hardly care about here for NYCFC purposes. I think it works pretty well for that, because it uses every game played by every team, and all the schedule oddities, injuries, etc balance out. Plus, even though you focus on the current sixth place team to set that line, it incorporates all the teams, because you need to have 5 teams higher and x teams lower, and if any of them were better or worse enough the placements would shift and all of that goes into generating that projected line.
For predicting any given team's finish though, it has limitations that people try to fix with the added tweaks. Circling back to my thesis here, those additions increase sophistication and information at the cost of lower sample sizes. So my basic model has 24 data points for NYCFC. Yours adds another data point of just 12 games for NYC, and x games for the opponents, which in an extreme case is only 6 for DC as you've noted. I think what you need to do is figure out a weighted blend of a team's full record plus your smaller data points. Or maybe just pull things back to the mean when they are very far off. For example, halfway through the season NYC had a 2.75 Home PPG and 1.0 Away. Neither of those were likely to sustain and both have since moved in the expected direction.
 
Last edited:
For example, halfway through the season NYC had a 2.75 Home PPG and 1.0 Away. Neither of those were likely to sustain and both have since moved in the expected direction.
Putting a bit more meat on this. Right now league average Home PPG is 1.77 and Away is 1.01. That is a spread of 0.76. The NYCFC spread quoted as of the halfway mark was 1.75, more than double the average. Currently the NYCFC H/A spread is down to 1.42, which is probably still towards the end of the bell curve, but less of a ridiculous isolated anomaly.
 
Last edited:
  • Like
Reactions: SoupInNYC
Last week's results:

Wednesday 8/8:
LAFC 3 (6) @ HOU 3 (7) (USOC semi-final)

Saturday 8/11:
HOU 0 @ CLB 1
NJ 1 @ CHI 0
VAN 2 @ POR 1

Sunday 8/12:
NYC 3 @ TOR 2
DAL 1 @ SEA 2

With LAFC's PK loss to Houston in the USOC. The only Western teams who can open up another CCL spot with a MLS Cup win are now SKC & Houston. If Philly beats Houston in the USOC final, then it's only SKC. Furthermore, since Houston's chances of even qualifying for the playoffs are slim to none this season; we probably want to root for Philly in the final.

Speaking of Houston, they lost 1-0 at Columbus which doesn't really do much but keep the black & gold mathematically relevant for a few more weeks. NJ won 1-0 at Chicago, a game we were hoping they would drop points at. That ties them in wins (1st tie breaker) with Atlanta and they still have a match in hand. Portland finally lost at home this season to the same Vancouver team that just drew us at home. Portland still has 4 matches in hand on us and Atlanta, but they dropped points again and even if they won out those matches in hand, they'd only be able to draw with NJ if they win their match in hand. Isi gave us the win we needed at Toronto, a match we didn't expect to get 3 points at going in. By doing so, it shot us to the top of the table, even in matches with Atlanta. We also have the wins tie breaker as well. And lastly, Dallas lost at Seattle. They may be having a strong 2018 campaign, but down 16 points to us with 11 matches to play, it would take an incredible streak by them and collapse by NYC, ATL & NJ for them to have any chance as this CCL spot. Here's how the table looks after match week 24:

View attachment 9043

This is the 6th time that Atlanta & NYC have swapped 1st position in the last month.

This week's CCL matchups:

WEDNESDAY 8/15:
DCU @ POR 8:00

SATURDAY 8/18:
NYC @ PHI 7:00
NJ @ VAN 7:00
MIN @ DAL 8:00
POR @ SKC 8:30

SUNDAY 8/19:
CLB @ ATL 4:00

We start with Portland hosting DCU and Rooney's magic after their first home loss all year. Can they finally get back on track after dropping points 2 weeks in a row? Then a bid Saturday slate with NYC heading to USOC finalists PHI, NJ traveling cross country to Vancouver who picked up 4 points in 2 matches at NYC and POR, the 2 hardest places to play in the MLS. Dallas hosting a surging Minnesota team. Then Portland ending the night with their 2nd match of the week traveling to SKC, a very tough place to earn points. Sunday then has a big CCL matchup with Columbus traveling to Hotlanta. If ATL can get 3 points, CLB would be 13 points behind them with 9 matches to play, seemingly nullifying and CCL hopes from the aggregate table. Happy watching! And if you're a podcast, I just spelled all this out for you, please get it correct in your discussions ;)

First results of the week:

WEDNESDAY 8/15:
DCU 4 @ POR 1

SATURDAY 8/18:
NYC @ PHI 7:00
NJ @ VAN 7:00
MIN @ DAL 8:00
POR @ SKC 8:30

SUNDAY 8/19:
CLB @ ATL 4:00

Table now updated:

Screenshot 2018-08-16 10.41.24.png

Shwafta Shwafta noted that it's not looking good for Portland anymore. We all realize it's a 3 horse race. But we're still 4 matches or so before any teams are officially eliminated.
 
Yeah, at this rate even if portland win both of their next two games in hand, that puts then at 96 which is STILL 8 points, or 3 wins, behind us.
 
  • Like
Reactions: Kjbert and adam
First results of the week:

WEDNESDAY 8/15:
DCU 4 @ POR 1

SATURDAY 8/18:
NYC @ PHI 7:00
NJ @ VAN 7:00
MIN @ DAL 8:00
POR @ SKC 8:30

SUNDAY 8/19:
CLB @ ATL 4:00

Table now updated:

View attachment 9056

Shwafta Shwafta noted that it's not looking good for Portland anymore. We all realize it's a 3 horse race. But we're still 4 matches or so before any teams are officially eliminated.

Updated after Saturday's matches:

WEDNESDAY 8/15:
DCU 4 @ POR 1

SATURDAY 8/18:
NYC 0 @ PHI 0
NJ 2 @ VAN 2
MIN 0 @ DAL 2
POR 0 @ SKC 3

SUNDAY 8/19:
CLB @ ATL 4:00

Screenshot 2018-08-19 09.05.29.png
 
Don’t have ESPN+ to watch, so no idea of balance of play, but I like that Columbus has two guys on the bench that can do real damage at the end of the match to ATL’s back line: Mullins and Meram. Power and speed. Hopefully they play it right so those two have an opportunity.
 
Past week's results:

WEDNESDAY 8/15:
DCU 4 @ POR 1

SATURDAY 8/18:
NYC 0 @ PHI 0
NJ 2 @ VAN 2
MIN 0 @ DAL 2
POR 0 @ SKC 3

SUNDAY 8/19:
CLB 1 @ ATL 3

Standings:

Screenshot 2018-08-20 15.29.30.png

And we've swapped spots with Atlanta again! The away loss is nothing out of the norm for our Pigeons, but it's once again shifted the table around. Moreover, Columbus' loss has dropped them into the Dallas/Portland category. It would take nothing short of a miracle for those 3 teams to jump the top 3 teams. NJ picking up a point didn't help us much either. Their max points are only 3 behind us and our max is only 2 behind Atlanta. The match on Wednesday could actually ties with NJ in max points & wins if they were to beat us at YS. This race is super tight and we really need all 3 points at home on Wednesday night to both keep pace with Atlanta and keep NJ at bay. Especially with NJ's DGW to catch up with us and Atlanta on games played.

Possible min/max points by the Sunday night for the 3 top teams with all 3 having 8 matches remaining:

ATL 106/109
NYC 104/107
NJ 98/104

This week's matchups:

Wednesday 8/22:
NJ @ NYC 7:00

Thursday 8/23:
CLB @ CHI 7:00
DAL @ HOU 9:00

Friday 8/24:
ATL @ ORL 8:00

Sunday 8/26:
DCU @ NJ 7:00
SEA @ POR 9:30
 
Past week's results:

WEDNESDAY 8/15:
DCU 4 @ POR 1

SATURDAY 8/18:
NYC 0 @ PHI 0
NJ 2 @ VAN 2
MIN 0 @ DAL 2
POR 0 @ SKC 3

SUNDAY 8/19:
CLB 1 @ ATL 3

Standings:

View attachment 9075

And we've swapped spots with Atlanta again! The away loss is nothing out of the norm for our Pigeons, but it's once again shifted the table around. Moreover, Columbus' loss has dropped them into the Dallas/Portland category. It would take nothing short of a miracle for those 3 teams to jump the top 3 teams. NJ picking up a point didn't help us much either. Their max points are only 3 behind us and our max is only 2 behind Atlanta. The match on Wednesday could actually ties with NJ in max points & wins if they were to beat us at YS. This race is super tight and we really need all 3 points at home on Wednesday night to both keep pace with Atlanta and keep NJ at bay. Especially with NJ's DGW to catch up with us and Atlanta on games played.

Possible min/max points by the Sunday night for the 3 top teams with all 3 having 8 matches remaining:

ATL 106/109
NYC 104/107
NJ 98/104

This week's matchups:

Wednesday 8/22:
NJ @ NYC 7:00

Thursday 8/23:
CLB @ CHI 7:00
DAL @ HOU 9:00

Friday 8/24:
ATL @ ORL 8:00

Sunday 8/26:
DCU @ NJ 7:00
SEA @ POR 9:30

Up to date scores and standings, 2 more matches tonight.

Wednesday 8/22:
NJ 1 @ NYC 1

Thursday 8/23:
CLB @ CHI 7:00
DAL @ HOU 9:00

Friday 8/24:
ATL @ ORL 8:00

Sunday 8/26:
DCU @ NJ 7:00
SEA @ POR 9:30

Screenshot 2018-08-23 15.57.36.png
 
Starting?
Well up until now we still had some semblance of control of our destiny, even if there was ATL ahead of us by a bit, but now RB is definitely looking like they're going to overtake us if we don't right this ship.
 
Well up until now we still had some semblance of control of our destiny, even if there was ATL ahead of us by a bit, but now RB is definitely looking like they're going to overtake us if we don't right this ship.

Why do you say that? Even if they win their match, we have a 3 point lead with 8 matches to go. But now we need some help from Atlanta to regain the lead.
 
Why do you say that? Even if they win their match, we have a 3 point lead with 8 matches to go. But now we need some help from Atlanta to regain the lead.
3 point lead with 8 matches to go isn't exactly the safest of cushions, especially if we keep dropping points. That doesn't mean we won't be able to uphold it, just means that I'm much more nervous about it now than I've been before. For all I know we can go on an 8-game win streak to close out the season.
 
  • Like
Reactions: adam
3 point lead with 8 matches to go isn't exactly the safest of cushions, especially if we keep dropping points. That doesn't mean we won't be able to uphold it, just means that I'm much more nervous about it now than I've been before. For all I know we can go on an 8-game win streak to close out the season.

I hear you. But with our remaining schedule we should be favored except for the next match away vs Columbus.

NJ & ATL still have to play each other. And some tougher opponents too. I’m not conceding anything.
 
I hear you. But with our remaining schedule we should be favored except for the next match away vs Columbus.

NJ & ATL still have to play each other. And some tougher opponents too. I’m not conceding anything.
You bring up a good point, NJ vs ATL, which right now I guess right now we are looking for a tie. I can't remember off-hand the schedules, but I do remember that all three teams have relatively easy schedules to close out the season except for that. So unless there's some major slip-up by any of the teams (and ATL are the last of us three to yet have a slip up this season..) I think it'll end somewhere around what we have right now.