By Ted Sarvata (@TedSarvata)
Now that the Olympics are finally over, we can get back to NWSL action next weekend. Here’s the table, with Spirit, Thorns, Flash and Red Stars in playoff position and only the Breakers mathematically eliminated.
NWSL Elo tells a slightly different story. Spirit and Thorns are still 1-2, but 3-4 are Reign (6th on table) and FCKC (8th on table) respectively and our playoff teams Red Stars and Flash are in a virtual tie at 5-6. Sky Blue, the team seemingly most likely to claw its way into the playoffs sit in 7.
How is this possible, that a Reign with only 20 points an FCKC with only 16 points from 15 matches could be 3rd & 4th in NWSL Elo? Two factors, I think. One is that they won they played each other in the championship game last year and finished the league 1-2 in NWSL Elo, so started this year with a big gap on the rest of the league.
The second factor is this year’s unbalanced schedule. My awesome annotation skills (in pink) show the teams that play each other 4 times this season (everyone else plays each other 2 times). The Flash seem to be the main beneficiary of that schedule this year, having the pleasure of beating up on the Breakers so many times. The Reign have the misfortune of being 0-1-2 vs the Thorns so far this year, which is brutal on the table, but doesn’t hurt much in NWSL Elo. FCKC and Red Stars have only played twice so far, so maybe this theory doesn’t hold up for KC? I don’t really know.
One question I have looking at all of that above is why the huge gap to start the season for KC and Seattle. Were they really so much better than the league last year?
Well, yes. Turns out they were that much better, not just last year, but the last two years. This year, with the exception of the Breakers, everyone is much closer together. There is no one dominant team. What does this mean for the last quarter of the season? What if we use the NWSL Elo differences for each match to predict the outcome. For example, we could predict a draw any time the Elo rankings are close enough together that their team is at least a 60% favorite. Unfortunately, that’s pretty boring. Seattle and KC switch places. Red Star slip a bit. Boston continues to be bad at soccer.
What about forgetting about Elo and just predicting based on standings, with no draws? Just as boring, except our playoff teams end of doing a little better. Flash finish 5th in Elo and 4th on the table. Thorns and Spirit finish even on points and only 2 points apart in Elo.
If I was better and spreadsheets and programming and such, I could run 10,000 simulated seasons for the last 5 games and make some conclusions from that. Instead, I’ll just give you the
Week 16 odds based on NWSL Elo:
Pride 44% – Spirit 56%
Reign 58% – Thorns 42%
Flash 68% – Dash 32%
Sky Blue 56% – Red Stars 44%
FCKC 78% – Breakers 22%
See also: NWSL Elo Methodology