A look at odds of predicting Rams' schedule
By Nick Wagoner
http://espn.go.com/blog/st-louis-rams/post/_/id/7546/a-look-at-odds-of-predicting-rams-schedule
Early Monday evening, the St. Louis Rams announced a contest for fans to take their best shot at accurately predicting the team's 2014 schedule.
If someone can hit on all 17 weeks, they win $100,000. Sounds like a nice prize, but the reality is that meeting the team's requirements to win the contest turn a long shot into a near impossibility.
Simply predicting the opponent in each week would be a difficult enough task, but the rules of the contest also stipulate that the person must correctly predict the bye week and, making matters most difficult, the day of the week in which the game will take place.
Full disclosure here, I am the furthest thing you will find from a mathematician. I went to Journalism school and was always partial to English and social studies rather than math and science. So after spending all of about six seconds to ponder whether I could accurately come up with some sort of odds, I decided to ask someone who actually understands these sort of things. (For the record, my quick calculations put the odds at something like one in eleventy billion)
I turned the project over to the brilliant, patient and unbelievably helpful folks at ESPN Stats & Information as well as our analytics department. With so many moving parts to a contest such as this, calculating exact odds are nearly impossible because it's hard to factor in certain things such as the average participant's knowledge of how NFL scheduling usually works.
By Tuesday evening, I received a response and, as expected, winning this contest is essentially impossible. According to the work of Alok D. Pattani of our analytics team, the odds of getting just the 16 opponents plus the bye week comes in right around one in 188 trillion.
That numbers comes with only one major assumption on his part, which is that the bye week would come between weeks 4 and 12, which is common practice in the NFL. It does not account for other reasonable assumptions such as the unlikelihood the Rams would play a division opponent on consecutive weeks.
So if that was the only restriction for the contest, the odds would obviously be even longer but a bit more feasible than what it actually is. When you factor in the day of the week, it creates a lot more chaos.
In weeks 1-15, you can play a game on Thursday, Sunday or Monday. Teams are allowed a maximum of five total primetime appearances with one guaranteed Thursday night game. That is all accounted for, but for the purpose of this exercise, it would have taken too long to throw out some of the less likely constraints like playing three straight Monday night games, for example.
Pattani came up with 486,753 possible day-of-week combinations, which again doesn't even account for trying to guess the opponents. After factoring in everything, the final estimation of the odds of nailing every facet of the contest is about one in 92 quintillion.
That number came about from multiplying the ~188 trillion possible orderings of games plus bye week, by the ~500,000 game date combinations.
To be clear, though, those numbers are not exact. As Pattani points out, there are a lot of logical assumptions that can be made about how a schedule is put together that would improve the chances of getting the schedule correct. Even if you factor in a general knowledge of how the schedule works and how the Rams are perceived by the league for things like primetime games, predicting the schedule with 100 percent accuracy almost certainly won't happen. And if it did, based on these odds, the prize should be much larger than $100,000.