Jump to content

Nacho Simulation Football League (Season 20 - Taco Bowl XX POSTED!)

TheKillerNacho
 Share

Recommended Posts

Tk3

Tk3

Posted
Posted
6 minutes ago, pwny said:

Yeah, don't use that lol. Base pythagorean wins does not work for predicting this league. Point differentials in this league are too tight to allow accurate tabulation. Even your S10 team would only have 11.45 pythagorean wins.

That's precisely why I created PWN scores. PWN scores are literally just Pythagorean Wins for NSFL, where I tabulated data and adjusted the scale to match what point differentials in the league actually show. 

If you used the same scale I do (replace all of those 2.37s with 5.3) you'd end up with:
UZmcaf2.png

Which is probably closer to what you're looking for.

Interesting. Makes sense.

Did you have a rationale for 5.3? Or did you kind of goal-seek it to get the deltas to a low level?

Link to comment
Share on other sites
pwny

pwny

Posted
Posted
3 minutes ago, Tk3 said:

Interesting. Makes sense.

Did you have a rationale for 5.3? Or did you kind of goal-seek it to get the deltas to a low level?

The old formula I used, I had it set to 6.95. That one, I was trying to just set it so that the best team of all time got a 15, and the worst was ~1, because I felt that made sense when at the time I had only intended to use it to rank the truly great and bad teams. But after tabulating data for all 19 seasons, I realized those didn't work, as 18% teams were projected at least 2 games away from where they would up, and I was projecting too many teams as 13+win teams, and too many below 3, despite those being incredibly rare in the sim. So with that issue there, it just made it so the great teams were just automatically viewed as unlucky, and the really bad as lucky in their win totals. 

So the revision, I tried to get everything in line, and then goal sought to reasonable tolerances. Now, only 8% of teams are 2 games or more lucky/unlucky, which I fine reasonable. And after checking different formulas, I felt 5.29 gave me just every so slightly too low deviation from the 8 win average, and 5.31 just ever so slightly too much what I felt was resonable for the luck variation, so I settled on the average at 5.3. 

 

But I wouldn't say that I would necessarily settle exactly on 5.3 for what you're doing. You might want to play with it and drop it a little lower if you want to put more emphasis on the value of wins already accumulated, or may even want to make the number larger if you want the valuation of projection to play a larger role. 

Link to comment
Share on other sites
pwny

pwny

Posted
Posted
26 minutes ago, Malfatron said:

Ah, i thought it was because you made them

Happy accident that I was able to manipulate it to have two meanings. 

Link to comment
Share on other sites
pwny

pwny

Posted
Posted (edited)
8 minutes ago, El Ramster said:

Damm. When you hate a brother so much you add a L. 

My bad. I have it set so it auto-adds a loss to everyone (Losses are calculated by games played-Wins so I have less data to input every time) and then I update the Win totals to get the loss totals correct. Was on a work phone call when I did this and was somewhat distracted. I'm sure that's not the only mistake I made lol, but hopefully I just forgot yours because it was the last one to update.

 

But I went and fixed it, and the image I posted after that has been changed as well.

Edited by pwny
Link to comment
Share on other sites
Guest
This topic is now closed to further replies.
 Share
Go to topic listing


×
×
  • Create New...