It's that time of year again, where we start looking at the remaining schedule and figure out what it takes for each team to win the Coastal. Or at least that time of year when I've finally gotten off my ass and crunched the numbers.
At this point in the season, we are down to 11 conference games remaining involving Coastal Division teams, meaning we have 2^11 = 2048 different possible paths the season can take. Here is a list of each team still alive in the Coastal race, how many of the 2048 paths lead them to Charlotte, and what their odds are (based on the S&P+ rankings):
| Team | # Outcomes | Odds |
|---|---|---|
| Pitt | 1221 | 48.6% |
| UVA | 320 | 21.0% |
| VT | 356 | 14.4% |
| Miami | 86 | 11.6% |
| GT | 63 | 4.35% |
| GT, Pitt, or VT | 2 | 0.0100% |
Note: the 3-way tie listed in the table above will be broken based on the highest-ranked team out of the group in the "Team Rating Score metric provided by SportSource Analytics", as per tiebreaker #7.
Yes, 5 teams are still alive in the Coastal race. Yes, the 2 teams with the best chance to win have never won the Coastal before. Yes, Duke is eliminated, despite the fact that they tie for the lead in 72 possible outcomes (and lose by tiebreakers to each of the 5 teams still alive in the race). And while Virginia Tech is no longer the favorite, we are still in the race. However, to stay in the race we absolutely have to beat Pitt next week, as we won't get any help from tiebreakers in the unlikely chance we still get to 5-3 and tie with Pitt and potentially Duke, GT, and/or UVA.
Comments
Love it. Thanks for crunching the numbers and presenting them in such a readable fashion.
How many of this Pitt chances disappear when they lose next weekend? #TheSunWillComeUpTomorrow
A Pitt loss next week reduces their "outcomes" to 293 (out of 1024), with a 14.6% chance of winning based on today's S&P+ rankings. A VT win keeps VT's "outcomes" at 356 (since our chances rely on beating Pitt and UVA) but increases our odds to 28.5%, just behind UVA's 33.0%.
THIS is the high quality user (or staff) generated content that makes me love TKP. Great work!
Appreciate all the work. Question. Why is UVA with 320 outcomes more likely than VT with 356 outcomes to win the title?
Because UVA's outcomes are, on average, more likely than ours. The odds are based on the likelihood of a particular set of 11 game results occurring. For starters, slightly over half of our outcomes include Pitt beating Miami, a game where the S&P+ has Miami favored by 14.8, while the same is true for only about 20% of UVA's.
I am truly impressed with the level of neediness involved in putting this together.
Well that's a hell of a thing to say /s
My bad. I obviously meant nerdiness. But auto correct got me.
I've done it every year since 2011. It started as something to do one early November day when I was bored at work. At this point it's tradition. And I'm finishing up a Python script that should automate the process for future seasons, and make it possible to do the Atlantic division and even other conferences, as well as running it earlier in the season.
Is there a site with an api that lets you automate the data collection?
To date I've always just used an Excel spreadsheet since it looks like I'm working, but now that I'm my own boss I decided to teach myself Python by writing a script. So far I have a simple data file with the ACC games and results, but once I get the script working properly the next step will be finding an easy-to-read scores page and automating a scraper. Then I'll probably expand to other conferences, though the script would have to be adjusted since the tiebreaker steps will be different.
I actually been planning to make a full computer ranking system, so I had to search for a data set with full box scores. I found a really good one, but it's SQL-based. Really good set though if you're interested, I was planning to start building something out this off-season.
I just got the alert, what's wrong?
Figured I should add a short blurb about what's at stake for each team this weekend:
The only way VT can be eliminated from the Coastal race this weekend is with a loss.
GT is eliminated with either a loss to Miami or a Pitt win over us (which, as above, would also eliminate us).
Miami is only eliminated with a loss to GT.
Pitt and UVA cannot be eliminated this week, due to Pitt having 1 fewer loss than anyone else in the division and UVA not playing an ACC game this week.
I'm reading this as stating that if we win, we can't be eliminated from the Coastal race. But if we lose, are there still possible outcomes where we can win the Coastal? Is this truly an elimination game for VT?
It is an elimination game. We have to beat both Pitt and UVA to have any chance of winning the Coastal Division.
I'm not sure who to tell you to root for in Georgia Tech/Miami, though. It's an elimination game for both, and there's one tie-breaker with Georgia Tech + us and Pitt that can get messy depending on the interpretation of tie-breaker #4 (and I've seen it interpreted multiple ways online) but
Surely you mean tiebreaker #3, as #4 doesn't come into play with only 2 cross-division games unless there's a freak scenario with cancelled division games or whatever.
Yeah, I mean 3. I don't know why I keep trying to call it 4.
My prediction:
1) I think we win the next two games but barely. Maybe Willis has a couple of comeback 4th quarter victories over Pitt and Miami.
2) After laying the smackdown on Liberty, UVa races to an early lead against Georgia Tech and the Jackets can't catch up. Wahoos win, a score like 24-10.
3) The Wahoos come into Lane with their chest puffed out. It will be marketed as the most important game in the Bronco era as the division crown is on the line. Maybe they're even favored to beat the good guys by Vegas. Fuente and Co. will be much maligned by Hokie Nation as the previous two wins didn't inspire much confidence.
4) 38-0. Again. Bro.
I will go three rails in one night if all of this happens.
I might be right there beside you
If that happens I'll join you for three rails and buy yours!
I subscribe to this scenario.
Make it so!
Virginia Tech's chances of winning the Coastal are now 0.00%, as are Georgia Tech's.
Pitt is by far the favorite at 84.4%.
UVA still has a shot at 8.8%.
Miami is still clinging to hope at 6.8%.
The thought of UVA winning the Coastal bothers me a lot more than VT not winning it.
Figured it's worth another update... The Coastal is down to either Pitt or UVA.
UVA needs to beat both Techs and have Pitt lose to both Wake Forest and Miami to win. If Pitt wins another game, or if UVA loses another game, Pitt wins the Coastal.
Obviously, this means Pitt should be the overwhelming favorite. And according to my numbers, Pitt has a 91.2% chance of winning, versus an 8.8% chance for UVA, pending tomorrow's update to the S&P+ rankings.