It starts with the current FNFT standings and generates random results for the rest of the season. Then it tallies which four teams would make the playoffs if the season went down that way. It repeats this process 10,000 times. A team's Playoff Probability as reported in the standings is the percentage of those 10,000 random iterations in which that team made the playoffs. In statistical speak, there are called Monte Carlo Projections.
Here are the standings right now:
| RED | W | L | Pct | GB | PL% |
|---|---|---|---|---|---|
| Bill | 9 | 3 | .750 | -- | 38% |
| Connor | 8 | 4 | .667 | 1 | 34% |
| Jake | 7 | 5 | .583 | 2 | 31% |
| Tim | 6 | 6 | .500 | 3 | 26% |
| John | 4 | 8 | .333 | 5 | 20% |
| GREEN | W | L | Pct | GB | PL% |
| Avery | 8 | 4 | .667 | -- | 37% |
| Joel | 8 | 4 | .667 | -- | 38% |
| Brett | 6 | 6 | .500 | 2 | 29% |
| Joe | 6 | 6 | .500 | 2 | 30% |
| BLUE | W | L | Pct | GB | PL% |
| Tief | 9 | 3 | .750 | -- | 50% |
| Jacob | 5 | 7 | .417 | 4 | 32% |
| Max | 1 | 11 | .083 | 8 | 17% |
| Gerry | 1 | 11 | .083 | 8 | 17% |
As you can see only one team - me - has so far established a 50% chance of making the playoffs. That means that in roughly half of the random iterations my team was one of the playoff teams and in the other half it wasn't.
Also note that these percentages are not precise. Despite having identical records, Avery/Joel and Brett/Joe's percentage differ by 1%. That's just rounding and/or a random fluctuation. If I ran the simulation again they might be the same or they might be flipped the other way. It takes about two minutes to run 10,000 simulations. That's about how much time I'm willing to invest in this each week. If I ran more simulations, the percentages would be slightly more accurate but it wouldn't mean anything to anything.
Perhaps obviously, the better your record, the better chance you have of making the playoffs. The further ahead your are of the other teams in your division, the better your chance. My chance is helped by the two teams in my division that are off to very poor starts. The probabilities in the other two divisions are much more closely bunched.
If you have a good week, your playoff chances will go up. If you have a bad week they will go down. At the end season, four teams will be at 100% and all of the rest will be at 0%.
Take this for what it's worth. In 2004 I trailed Avery by 14 games after Game 23. I wound up winning the division by 11 games. I'm certain this system would have given me no chance to win. But that's why we play the games.
No comments:
Post a Comment