Probably because it is based on the admittedly dubious assumption that all teams in the league, including CHV, have the same home-field advantage. I think that it is much more plausible in the aggregate than for a single game.
Since projections are basically pointless at this point in the season, here's my model's probabilities for each of the remaining games. HomeAwayHome Win %Draw %Away Win %RSL CHV 67.8 18.8 13.4PHI SKC 41.7 27.7 30.5TOR MON 43.3 26.0 30.7SJE FCD 48.4 25.3 26.3CHV POR 32.6 26.7 40.8DCU HOU 35.1 28.2 36.6CLB NER 45.6 25.7 28.8NYR CHI 57.2 22.4 20.4VAN COL 48.8 24.9 26.3SEA LAG 44.6 26.0 29.4 Not much to say here. Pretty much all of them are right about where you'd expect them to be.
Here's what mine look like after the LA-SJ game. Not very different at all, just including it for completeness' sake.
Oh, I totally get where you're coming from. Coming up with probabilities on that game is basically an exercise in futility. Recent form says LA wins easily, but I think everyone still kind of expects Seattle to wake from the dead before this is all said and done with. FWIW, my model is telling me the spread is 36%/33%/31% - not a whole heck of a lot different.
I'm not sure about waking from the dead, but I'll give them a bit of a benefit of the doubt since they'll have a large (probably) pro-Seattle crowd there.
Here are the final 2013 MLS ELO+ team and strength-of-schedule ratings based on all teams starting the season at 100 (average). Code: Rank Team ELO+ Rank Team Sched 1 NYR 118 1 CHV 104.4 2 POR 117 2 DAL 102.5 3 SKC 116 3 DCU 102.3 4 RSL 116 4 SEA 101.7 5 SJE 112 5 VAN 101.5 6 LAG 109 6 COL 101.3 7 COL 107 7 LAG 101.1 8 HOU 106 8 SJE 100.8 9 NER 105 9 RSL 100.5 10 VAN 104 10 CLB 100.4 11 SEA 102 11 POR 100.4 12 CHI 102 12 MTL 99.6 13 PHI 98 13 TOR 99.6 14 MTL 96 14 PHI 98.4 15 DAL 93 15 HOU 98.2 16 CLB 89 16 CHI 98.0 17 TOR 80 17 NER 97.3 18 CHV 72 18 NYR 96.0 19 DCU 58 19 SKC 96.0 Every team played 34 games (K=5.88), and home teams had a winning percentage of 0.635 (H=27). Here are the final (iterated) 2013 MLS ELO+ team ratings and corresponding points predictions, obtained by setting the initial rating for each team equal to its final rating, rather than 100, and then summing each team's expected number of points across the entire season. The actual point totals are given for comparison. Code: Rank Team ELO+ Rank Team Proj Actual 1 POR 128 1 NYR 59.3 59 2 NYR 125 2 POR 58.6 57 3 RSL 121 3 SKC 57.8 58 4 SJE 120 4 RSL 55.2 56 5 SKC 119 5 SEA 51.6 52 6 COL 116 6 LAG 51.2 53 7 LAG 114 7 HOU 50.8 51 8 SEA 112 8 COL 50.6 51 9 VAN 110 9 SJE 50.3 51 10 NER 106 10 NER 50.1 51 11 HOU 105 11 VAN 48.4 48 12 CHI 105 12 CHI 47.9 49 13 DAL 95 13 MTL 46.8 49 14 PHI 90 14 PHI 46.0 46 15 MTL 88 15 DAL 45.1 44 16 CLB 85 16 CLB 39.0 41 17 TOR 64 17 TOR 29.4 29 18 CHV 62 18 CHV 25.6 26 19 DCU 35 19 DCU 16.1 16
Alright, here's the end of season expected performance: TeamEAPActualLuck POR61.357-4.3 SKC58.958-0.9 NYR58.259+0.8 RSL57.656-1.6 LAG57.053-4.0 NER54.351-3.3 VAN52.248-4.2 COL51.451-0.4 MON47.549+1.5 SEA46.452+5.6 HOU46.351+4.7 PHI44.344-0.3 FCD44.244-0.2 CLB43.741-2.7 CHI43.549+5.5 SJE41.051+10.0 TOR34.430-4.4 CHV23.626+2.4 DCU21.516-5.5 Basically, the model is based upon GF:GA, so a team like San Jose which somehow got 51 points while having a negative goal differential is perceived as being a poor team. Portland's 5 goal torching of Chivas in the last weekend moved them from 58.3 up to 61.3, which is a fairly major change. The "Luck" column is (Actual - EAP), and is a representation of how many points were gained/lost over what the team would be expected to do against an average opponent. Interesting here is that this shows Chivas as overperforming with 26 points, while both DC and Toronto underperformed.
Surprise surprise, New York finish at the top of the ratings with SKC is close tow. The two surprises here are San Jose and Vancouver each finishing so high (7th and 8th) despite getting knocked out of the playoffs. Montreal has slid all the way down to 16th, and finish the lowest of all the playoff teams. Obviously, the End of Season Points Predicted are kind of meaningless here. I also used my game outcome predictions to predict playoff advancement by round, here are the results:
[quote="vividox, post: 28948440, member: 66900"The two surprises here are San Jose and Vancouver each finishing so high (7th and 8th) despite getting knocked out of the playoffs. [/quote] San Jose shouldn't be a huge surprise. While they finished 10th in the Supporter's Shield, they were tied on points and wins for 7th (losing on goal difference). Vancouver makes less sense.
Vancouver doesn't surprise me at all, because they actually did well at winning/staying in games. After all, they did finish with a positive goal differential, unlike San Jose. I'd say San Jose finished rather high because they did well in the second half of the season.