Alternate History Sports

You are not logged in. Would you like to login or register?



3/11/2021 6:15 pm  #171


Re: The AltHL Season 2020

It's halfway through the season, and it's power rankings time. After last season's fiasco (so very sorry, that method was bad), I spent an entire half of a season finding a new method. Simple version? How much you score matters.
The long description with all those nitty Gritty details is in the spoiler.

This system (JmeLO) is not unlike the ELO system, which I used unsuccessfully in the AltFL. ELO, though, is based off of expectations of wins vs. losses, while JmeLO is based on pure, unadulterated score.

Why doesn't ELO work?
Wins and losses are wonky in fantasy sports. Though the "better team" wins each game, a team with good luck but bad skills can gather a bevy more wins than a team that scores massively more. Initial rankings (which in the AltFL favored unreasonably the Bandits, my team, which is sketchy) in ELO also matter a lot, and are loathe to change. All teams in JmeLO start the season at the average 500 JmeLO.

What the heck is a JmeLO then?
JmeLO is an ELO-like ranking system. Rankings are in the thousands, with a JmeLO of 500 being average. The compositional unit of JmeLO rankings is bees, for lack of a better term. A big portion of ELO-adjacent rankings is converting bees into expectations, comparing expectations into reality, and turning that comparison into a way to update bees. ELO gives each team a win% based on the ELO difference between itself and its opponent and updates the scores of each competitor based on the difference between the expected win% and the result (100% for win, 0% for loss). JmeLO takes the bees of each team, converts to expected deviation from the mean (at a rate of 100 JmeLO = 1 STDEV), takes the difference between that expectation and the number of standard deviations that team performed above the mean score that week, multiplies that difference by 25, and adds it to the previous JmeLO. Thus teams that consistently score 1 STDEV above the mean will tend toward 600 JmeLO and those that score consistently 1 STDEV below the mean will tend toward 400 JmeLO.

Why is it JmeLO, though?
It was originally JHALO — JHA for JamHeronArk and the last two letters of ELO. Then, in a brainstorm session, QCS suggested JmeLO as in "JmeLO Ball," and it was introduced in the vein of FiveThirtyEight's Carmelo player rankings. The name stuck. Unfortunately.

Where can I find the spreadsheet?
Here.

Steal the spice trade?
That's not a question but the Dutch did it anyway.

Now, the rankings.

Last edited by JamHeronArk (3/11/2021 6:19 pm)


 

3/12/2021 1:32 am  #172


Re: The AltHL Season 2020

JamHeronArk wrote:

Why is it JmeLO, though?
It was originally JHALO — JHA for JamHeronArk and the last two letters of ELO. Then, in a brainstorm session, QCS suggested JmeLO as in "JmeLO Ball," and it was introduced in the vein of FiveThirtyEight's Carmelo player rankings. The name stuck. Unfortunately.

Maybe we need a JaVAR rankings; increasingly unhinged and bombastic claims about why each team is the best.

 

3/12/2021 9:46 am  #173


Re: The AltHL Season 2020

Look at the Arrows, #2 in these rankings, with a lousy record.



2x Alt Champion :: AltLB Champion Oklahoma City Bison - 2022 :: AltFL Champion New York Emperors - 2022
 

3/12/2021 1:26 pm  #174


Re: The AltHL Season 2020

Dan O'Mac wrote:

Look at the Arrows, #2 in these rankings, with a lousy record.

 
I’ve been out for JAM on this take. Have a win over them, a 7-1 record and more points (for now) yet I haven’t been able to hop over them for the last month.



 

3/12/2021 2:39 pm  #175


Re: The AltHL Season 2020

*Beats Calgary and Winnipeg* 
*Ranked below them in the very suss power rankings*

Checks out.



AHS Admin. Creator of the THLPUCHWHA: Redux and Retroliga.
 

3/12/2021 3:57 pm  #176


Re: The AltHL Season 2020

Thehealthiestscratch wrote:

Dan O'Mac wrote:

Look at the Arrows, #2 in these rankings, with a lousy record.

 
I’ve been out for JAM on this take. Have a win over them, a 7-1 record and more points (for now) yet I haven’t been able to hop over them for the last month.

Last week, your bladey boys were first. The teams are the same across the rows, no matter how a team ranked in previous weeks. The sheet is sorted by the most recent rankings. You've dropped out of the top four once, while the Arrows started low, and have been scoring enough points recently to pull them to within three points of you on the season. It is only fair that their rankings reflect their performance.

Steelman wrote:

*Beats Calgary and Winnipeg* 
*Ranked below them in the very suss power rankings*

Checks out.

*Beats Winnipeg week 3*
*Is ranked above Winnipeg after that week*
*Loses two straight*
*Drops in rankings because they aren't doing as well*
*Beats Calgary week 6*
*Is ranked above Calgary until the next week where they give a mediocre performance while Calgary kicks their hockey up a notch*
*Is ranked fifth in suss power rankings in a league where the top seven has had more turmoil than the Balkans*
Steel: *Angery noises*

I will defend my rankings. They value consistent improvement in relation to the pack and essentially show how scary a team is to go up against week-by-week. The Arrows at two is mostly because they took many close losses to teams who had their best week specifically against Winnipeg but haven't been outside the top five in weekly scoring since week 2 (where they were sixth). In that span, they went 3-3. Consistency is key, otherwise, rankings will swap with the ups and downs of play, as the Wranglers and the Kingfishers have been dancing around each other for weeks. If you want a record-based ranking, there's always Fantrax.

sportsfan7 wrote:

Maybe we need a JaVAR rankings; increasingly unhinged and bombastic claims about why each team is the best.

This is essentially what we had for the AltFL. Yeah, the old ways were bad. Those rankings were biased. I must admit that now.


 

3/12/2021 5:17 pm  #177


Re: The AltHL Season 2020

It’s not fair because you aren’t factoring in the Scratch bias that I am demanding. I can help though. Just do a simple if/then and reference another cell. If that cell = “POR”, then it is +200 JmeLO.



 

3/14/2021 8:37 am  #178


Re: The AltHL Season 2020

Thehealthiestscratch wrote:

It’s not fair because you aren’t factoring in the Scratch bias that I am demanding. I can help though. Just do a simple if/then and reference another cell. If that cell = “POR”, then it is +200 JmeLO.

'=IF(B5="POR",G5+25*(((D5-D$3)/D$2)-((G5-500)/100))+200,G5+25*(((D5-D$3)/D$2)-((G5-500)/100)))'


 

3/14/2021 7:32 pm  #179


Re: The AltHL Season 2020

So... what's the rationale for suddenly adding two IR slots?



AHS Admin. Creator of the THLPUCHWHA: Redux and Retroliga.
 

3/14/2021 8:21 pm  #180


Re: The AltHL Season 2020

Steelman wrote:

So... what's the rationale for suddenly adding two IR slots?

I realized it didn't give us them at the beginning of the season. We should have had them all along. I also noticed the teams around the league could probably use them. I know it changes things a bit mid-stream but I figured it was better to just correct it now. 


 
     Thread Starter
 

Board footera

 

Powered by Boardhost. Create a Free Forum