AUDL Playoff Projections

I’ve always been a fan of Hollinger’s Playoff Odds for the NBA. They give you an idea of where teams are heading, and separate the contenders from the pretenders. It’s something I’ve always wished existed for the AUDL. It would separate the subjective from the objective. With that in mind, I’ve put together a formula that tries to accomplish this.

Methodology

The basic methodology behind the projections is that each team gets a score for their home games and away games. Each score is compared to their opponents. So if the Wildfire are playing the Radicals in Chicago, and the Wildfire have a home score of 4, and the Radicals have an away score of 2, the Radicals would be assigned a 66.7% chance of winning that game.

Scores are calculated by taking a teams average point differential and multiplying it by their winning percentage. If the team has a negative average point differential it is multiplied by their losing percentage. This is done for both home and away games. If a team is playing at home, two is added to its score. If a team is playing against a team on the second game of a two game road trip, one is added to its score. Then the average margin between the two teams is taken, divided by three, and added to the higher teams score. If the two teams haven’t played before, their average point differential against common opponents are compared, and the difference is added to the higher teams score.

If a team has a result below 1 (which is very common) then however much is needed to get that teams score to 1 is added to it. The score for the opposing team is increased accordingly.

So for example, on Saturday Raleigh is playing in Charlotte. Raleigh is 4-2 on the road, with an average goal differential of +2.17. So 2.17 is multipled by their winning percentage (.67%), resulting in a score for their away games of 1.44. Charlotte is 1-4 at home, with an average goal differential of -2.2. So -2.2 is multiplied by Charlotte’s losing percentage (80%), resulting in a home score of -1.76. Charlotte’s score is increased by 2 since they’re playing at home. Charlotte has lost to Raleigh twice, by 9 and 12. So 10.5 is divided by 3 and added to Raleigh’s score. That results in a Raleigh score of 4.94 and a Charlotte score of .24. Both scores are increased by .76 to bring Charlotte’s score to 1. Which brings Raleigh’s score to 5.7. 5.7/6.7 is .8507, so Raleigh is assigned an 85.07% chance of winning that game.

Obviously the methodology isn’t perfect. It emphasizes the home field differential in both the original home/away point differential, and then gives an additional 2 point bump to home teams. This helps balance out some teams that haven’t had a balanced home/away schedule so far this season. Some teams have played most of their easy road games, and tough home games, but haven’t yet played their easy home games or tough road games.

Imbalanced schedules with are a problem overall for the model, though this is somewhat compensated by the average point differential adjustment. As for the 2 point bump for home field, and taking the average point differential and dividing it by 3, these are not mathematically arrived at, but given the scale of the range of scores they both seemed appropriate.

I’m no mathematician. I’d be open to ideas for improving this. But it does take a formula and apply it the same way to every team, resulting in an objective way at looking at the projected records for each team. These aren’t my predictions, but they do largely affirm conventional wisdom (conventional wisdom being what Evan Lepler has said, and to some degree the AUDL Power Rankings) which I think is a good sign that the formula is somewhat reliable.

Below I’ll show the likelihood of each teams possible finishes, and discuss the implications a bit.

The West

Projected Playoff teams:

San Jose: 10-4

Seattle: 9-5

San Diego: 7-7

San Jose

  • Current Record: 9-3
  • 11-3: 29.2%
  • 10-4: 52.7%
  • 9-5: 18.2%

Seattle

  • Current Record: 6-3
  • 11-3: 6.9%
  • 10-4: 29.6%
  • 9-5: 40.1%
  • 8-6: 19.9%
  • 7-7: 3.4%

San Francisco

  • Current Record: 5-5
  • 9-5: 0.9%
  • 8-6: 10.6%
  • 7-7: 36.6%
  • 6-8: 39%
  • 5-9: 12.9%

San Diego

  • Current Record: 4-6
  • 8-6: 12.3%
  • 7-7: 48%
  • 6-8: 31.6%
  • 5-9: 7.6%

Vancouver

  • Current Record: 3-6
  • 7-7: 2.5%
  • 6-8: 15%
  • 5-9: 37.9%
  • 4-10: 36.8%
  • 3-11: 7.8%

Los Angeles

  • Current Record: 3-7
  • 7-7: 2.2%
  • 6-8: 20%
  • 5-9: 50.1%
  • 4-10: 24.4%
  • 3-11: 3.3%

The West is looking like two races right now, one for first, and one for third. Seattle may have just taken an overtime win against San Jose, but having played three fewer games than San Jose puts Seattle at a serious disadvantage, even after a big win. The West is brutal, there are no gimme games in this division. The model I put together gives no team a 90% chance of winning any single game the rest of the season. That’s not the case for any other division. So having to play a few more games makes it much tougher to see Seattle surpassing the Spiders for the #1 spot in the division.

The race for the final playoff spot looks like it is going to come down to San Diego and San Francisco. Los Angeles and Vancouver are still both technically alive, but the Aviators have a big hole to climb out of and the Riptide have a punishing schedule, with a southern California two game road trip along with a home and away split weekend series against Seattle.

The key right now is that the Growlers have a much more favorable schedule going forward. So even though San Francisco won the season series 2-1 and holds the tiebreaker, the odds are in San Diego’s favor. Of San Diego’s four remaining games three are at home. Their opponents average winning percentage is 31%. Of San Francisco’s four final games they have a two game road trip to Seattle and Vancouver, and their opponents average winning percentage is 61%.

By no means is San Francisco out of it though. If they can hold home field advantage against the Cascades and Spiders, they’ll get to seven wins which will most likely be enough to get them to the playoffs. And while those won’t be easy games, they’re actually the slight favorite for both. The formula gives them a 55.8% chance of beating Seattle and a 57.3% chance of beating San Jose. This playoff race won’t be decided until the final weekend of the regular season, and it’ll be interesting to watch it play out over the next month in the West.

The Midwest

Projected Playoff Teams:

Madison: 12-2

Pittsburgh: 12-2

Chicago: 9-4-1

Madison

  • Current Record: 9-1
  • 13-1: 33.1%
  • 12-2: 51.9%
  • 11-3: 14%
  • 10-4: 1%

Pittsburgh

  • Current Record: 7-2
  • 12-2: 59.4%
  • 11-3: 33.2%
  • 10-4: 6.7%
  • 9-5: .6%

Chicago

  • Current Record: 6-2-1
  • 10-3-1: 7%
  • 9-4-1: 49.3%
  • 8-5-1: 39.6%
  • 7-6-1: 3.7%

Indianapolis

  • Current Record: 7-3
  • 11-3: 2.8%
  • 10-4: 25.2%
  • 9-5: 59.3%
  • 8-6: 12.2%

Minnesota

  • Current Record: 2-7-1
  • 5-8-1: 6.2%
  • 4-9-1: 37%
  • 3-10-1: 53.6%
  • 2-11-1: 2.9%

Cincinnati

  • Current Record: 2-8
  • 5-9: 1%
  • 4-10: 13.1%
  • 3-11: 59.8%
  • 2-12: 26.3%

Detroit

  • Current Record: 0-10
  • 2-12: 1.1%
  • 1-13: 15.6%
  • 0-14: 83.3%

The main takeaway from the Midwest projections is that the formula likes Chicago to make the playoffs despite the obstacles ahead. The main obstacle of course is a double header this weekend against the Radicals. But the Wildfire are currently undefeated at home, and are the slight favorites (55.6%) in their home game against Madison. Given Chicago’s struggles on the road, and Madison’s dominance at Breese Stevens Field, the Radicals are overwhelming favorites (94.1%) to defend their home turf. Though even if the Radicals did drop a game to the Wildfire, they still retain the tiebreaker over the Thunderbirds, so if those two teams tied the Radicals would still get that first round bye.

After their win over the Thunderbirds last week there was talk of how Chicago only now trailed Pittsburgh by a half game, but that is deceiving. The Thunderbirds have already gone through the toughest part of their schedule, and it is extremely unlikely that Chicago passes them in the standings.

If things played out this way Indianapolis would be at 9-5 and on the outside looking in for the playoffs for the first time. The last two years they’ve lost to Madison in the first round. The AlleyCats need Madison to sweep Chicago this weekend, otherwise they’ll need to beat the Radicals (which they’ve never done) in Madison, or the Thunderbirds in Pittsburgh. Both of those teams are undefeated at home so far this year.

The South

Projected Playoff Teams:

Atlanta: 11-3

Raleigh: 11-3

Atlanta

  • Current Record: 8-2
  • 12-2: 7.5%
  • 11-3: 63.3%
  • 10-4: 26.2%
  • 9-5: 2.9%

Raleigh

  • Current Record: 8-3
  • 11-3: 77.7%
  • 10-4: 20.9%
  • 9-5: 1.4%

Jacksonville

  • Current Record: 7-4
  • 10-4: 8.7%
  • 9-5: 76.6%
  • 8-6: 14.1%

Nashville

  • Current Record: 1-8
  • 5-9: 0.6%
  • 4-10: 11.5%
  • 3-11: 74.8%
  • 2-12: 12.6%
  • 1-13: 0.6%

Charlotte

  • Current Record: 1-8
  • 4-10: 0.9%
  • 3-11: 8.6%
  • 2-12: 36%
  • 1-13: 54.4%

Jacksonville is on life support in the playoff race in the South Division. Given that home field advantage has been much more pronounced in the South than any other division though, the Cannons still have a chance. The formula has the Cannons as heavy favorites (88.9%) this weekend in Jacksonville against the Hustle, and then they have a home game later in the season against Nashville, which should definitely be a victory.

To make the playoffs though, Jacksonville will have to overcome the odds and beat the Hustle in Atlanta. The Hustle are undefeated at home so far this season, the only home team in the South that remains so, and are strongly favored (89.6%) to beat Jacksonville that fourth of July weekend game. Jacksonville still controls their own destiny. If the Cannons beat Atlanta in both games they’ll have the tiebreaker over the Hustle, which would allow them to qualify for the playoffs.

Meanwhile Raleigh has the tiebreaker over Jacksonville, making it near impossible for the Cannons to leapfrog the Flyers. Things would get really interesting if the Flyers lost to the Express in Charlotte this weekend. Then, if the Cannons did take both games from Atlanta, and each team finished 10-4, there would be a three way tie. Raleigh holds the tie breaker over Jacksonville, Atlanta over Raleigh, and Jacksonville would hold the tiebreaker over Atlanta in this scenario. It’s unlikely to happen, but it’s not completely outside the realm of possibility.

The reason it’s possible is because the Express play well at home. They’ve always got a punchers chance in any game there, despite their 1-4 home record. Charlotte has a 14.9% chance to win that game at home against Raleigh, and a 21.3% chance to get a win at home against Atlanta.

But as surprisingly strong as Charlotte has been at home, the Express have been just awful on the road. At home they’re 1-4 with an average goal differential of -2.2. On the road they are 0-4 with an average goal differential of -11.25. This is why the formula has Charlotte as a significant underdog when the Express travel to Nashville to take on the Nightwatch. It’s not totally fair to Charlotte, given that they haven’t yet played in Nashville, which would most likely improve their average goal differential on the road. But it doesn’t take much digging to see that Charlotte has been close in every single home game, and usually trailing significantly when playing on the road.

The East

Projected Playoff Teams:

Toronto: 13-1

New York: 11-3

Montreal: 9-5

Toronto

  • Current Record: 8-1
  • 13-1: 53.5%
  • 12-2: 38%
  • 11-3: 7.8%

New York

  • Current Record: 6-3
  • 11-3: 58%
  • 10-4: 33.6%
  • 9-5: 7.6%

Montreal

  • Current Record: 6-3
  • 11-3: 2.3%
  • 10-4: 23.2%
  • 9-5: 46.9%
  • 8-6: 23%
  • 7-7: 4.3%

Ottawa

  • Current Record: 6-4
  • 10-4: 1%
  • 9-5: 17.3%
  • 8-6: 46.6%
  • 7-7: 29.8%
  • 6-8: 5.2%

Washington DC

  • Current Record: 4-4
  • 9-5: 1%
  • 8-6: 8.9%
  • 7-7: 32.9%
  • 6-8: 46.9%
  • 5-9: 9.8%

Rochester

  • Current Record: 1-6
  • 3-11: 14.4%
  • 2-12: 39.1%
  • 1-13: 43.3%

Philadelphia

  • Current Record: 0-10
  • 3-11: 1.5%
  • 2-12: 18.3%
  • 1-13: 70.1%
  • 0-14: 9.5%

At 8-1 Toronto has essentially sewn up the first spot in the East, given that the Rush own the tiebreaker over the Empire. And what may be surprising is that New York looks like it will walk into second place. The Empire have to play DC twice, Philadelphia twice, and Rochester once. So it’s no surprise to see them as having a 91.6% chance to finish with at least 10 wins.

The drama in the East will be for the third playoff spot. Montreal is still the favorite, but Ottawa holds the tiebreaker over the Royal. That’s what could make this interesting. Ottawa has a tough schedule, with two games against Toronto. But if the Outlaws can beat the Royal in Montreal again, and take care of DC when the Breeze come to town they’ll already be looking at 8 wins. Splitting that last weekend series in Toronto would most likely be enough to get that third playoff spot. If not, they’ll be hoping that somehow Rochester or DC can pull off an upset against Montreal. Both have already come very close to doing so this season, so it’s a possibility.

At 4-4 DC is on the periphery of the playoff race. Frankly it’s impressive that they’re still in it. The Breeze have demonstrated that they’re a class above Philadelphia and Rochester, and have even snuck in a win against Ottawa while nearly getting one against Montreal. With a 9-5 record likely being needed to advance, or at least an 8-6 record, things aren’t hopeful for DC. But the mere fact that they have a real chance to go 7-7 after losing more talent than any other team in the league in the offseason is a testament to the coaches and players still on that team.

Playoffs-First Round

These projected standings would give us these three first round matchups:

San Diego at Seattle

Montreal at New York

Chicago at Pittsburgh

New York would be favorites against Montreal, with a 78% chance of winning according to the formula. That’s good, but not yet safe. Seattle would be favorites against San Diego, with an 86% chance of winning. Though San Diego did win in Seattle earlier this season. Pittsburgh would have the best chance of advancing of all the first round home teams, given Chicago’s struggles on the road. The Thunderbirds would have a 90.8% chance of winning.

Obviously all of the percentile projections are fluid. As more data comes in the more they would change. And obviously there are huge holes in taking this approach. It doesn’t factor that the worst part of San Diego’s season was the first five games. Since then the Growlers have been on fire, going 4-1. And it doesn’t factor in that Seattle has added significant talent for the stretch run.  And it could be that it favors the home team too heavily at the moment, but I’m not convinced that’s the case. Remember, in 2013 and 2014 the only home playoff teams to lose a game have been in the final four. They’ve never lost in the first two rounds.

Playoffs-Second Round

Raleigh at Atlanta

Seattle at San Jose

Pittsburgh at Madison

New York at Toronto

Again, the higher seeded (and home) teams would all be favored here. Atlanta would have a 77% chance to win against Raleigh. San Jose would have a 68% chance to win against Seattle. Madison would have an 83% chance to win against Pittsburgh. And Toronto would have a 91% chance to win against New York.

Playoffs-THE FINAL FOUR

Whatever value this formula and these projections have for intra-divisional play, it would be completely useless for inter-divisional play.

Takeaways

If I were to make changes to this formula (but it actually took a while to put all this together, it wasn’t easy) I would probably bump the home bonus from plus two points down to plus one point. There would be games where I’d regret doing it, but I think on the whole it might be an improvement. But I do think it’s a necessary component, at least in season before all the teams have finished playing their games. I also might use the average margin against common opponents as a bonus for games even when the teams have already played each other.

The other thing I found was that statistics without context can be extremely deceiving. For example, did you know that Midwest teams playing their second game of a back to back on the road are 7-4-1 this season? With an average goal differential of +4.17? That’s amazing, right? Conventional wisdom is that teams should be performing worse in that scenario. But if you look closer you might notice that Detroit has hosted four teams playing their second game of a back to back, and the Mechanix have yet to go on a two game road trip yet this season. In fact, Cincinnati, Detroit, and Minnesota have hosted a combined 9 games against teams playing their second game of a back to back road trip. Madison, Pittsburgh, and Indianapolis have only hosted 1 such game between the three of them.

And for what it’s worth I don’t care what my formula tells me, I’ve still got Raleigh coming out of the South.

Advertisements

One comment

  1. […] AUDL reading! From Nathan Jesson at Semi-Pro Ultimate, AUDL Playoff Projections now that ten weeks are completed in the 2015 season. Using Hollinger’s Playoff Odds for the […]

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: