Thanks to "anonymous" who found a flaw in the Power Alley methodology, I can now unveil a new and improved version. It's more meaningful, yet more confusing too.
Flaw is perhaps too strong of a word, it did what it was supposed to - rank teams based upon actual record and pythagorean record - but the expected win total that resulted didn't make sense. The original formula did not take into account the actual number of current wins for each team, it simply took the Power Alley winning Percentage (1/3 real life Wpct + 2/3 Pyth Wpct) and multiplied it by 162.
The new and improved formula takes current wins, and adds expected wins to it. Expected wins are determined by: Games Left * Power Alley winning percentage.
Why more confusing? Take a look at the #2 and #3 teams below - the Oathbinder's and the Diamonds. The Oathbinder's are ranked higher but have fewer expected wins. This is possible because Power Alley favors Pythagorean win percentages over standard win percentages. The Oathbinder's are a "better" team because they do a better job of outscoring their opponents. However, based upon what has happened during the season, the Diamonds will likely end up with higher win total.
Also, this highlights how unlucky certain teams are. Look at the Rage, currently ranked 16th yet appear to be in the running for worst record in the game. They have had horrible luck as their 13th best Run Differential can attest too. It is not often a team has the fewest wins in the game yet have an RD closer to the middle of the pack.
If you have still reading, kudos to you, and the bottom line is - team rank remains completely unchanged, but expected win total is now more meaningful.
No comments:
Post a Comment