User Pass
Home Sign Up Contact Log In
Forum > Suggestions > Endorsements - What I am thinking
Page:
 
DavidD458
offline
Link
 
Originally posted by RandomBeast
why not only count stats earned while not in cruise control mode?


Solid idea.
 
mpatt75
offline
Link
 
I think this is something we can all live with Bort.

Just need to finalize the MVP/fame algorithms and we're set for you to write the code? How long will that take you, Sir?
 
kurieg
offline
Link
 
To calculate MVP stats:

1 - Throw out interconference games.
2 - Select the 9 highest average score differential teams. Only count stats against these teams. Throw one of them out randomly, and use the remaining 8 to compare stats with the 7 that don't make this list. Explanation -
Teams 1-9 have the best score differentials. Team 1 gains fame based on games vs. 2-9. Team 2 vs. 1 and 3-9. Etc.
Roll 1-9 and throw that team out. Team 10-16 are compared to the remaining 8 to gain fame.

Oh ya, don't bother calculating fame until the end of the season so you can weed out the terrible teams/gutjobs.
 
britdevine316
offline
Link
 
Originally posted by lump7502
i hate the ideal of using any kind of stats in the formula it just plains screws up more what is allready wrong with the game


Anything based on stats should be at the end of the season when every team has had a chance to play everyone. Also the inter conference game should be thrown out as that can really shade things.
 
Bort
Admin
offline
Link
 
Originally posted by mpatt75
I think this is something we can all live with Bort.

Just need to finalize the MVP/fame algorithms and we're set for you to write the code? How long will that take you, Sir?


The scripts to do the handouts are basically ready to go, with a couple of adjustments, so that's not a problem.

The main thing to work out would be the MVP formula.
 
Hikariu
offline
Link
 
Originally posted by Bort
Originally posted by Hikariu

Originally posted by Bort


Originally posted by Hikariu



8 Days is short for one national (2 tiers down) and two local (bottom tier) rewards.

I do like the spreading the love if the gear wasn't so bad. Pro teams being awarded a +2 seems rather silly, but at least it's something.

Why not go one tier down for national and 2 for local from where you player is currently playing? Then rotate on your 8 day scheme. Seems doable.


The National would be a full season deal, and are 2 tiers down. The locals are the ones that are 8 days. I guess I could see maybe doing 2 tiers/3 tiers instead of 2/4.


Why not 1/2. National is a full season. 2 or 3 local pieces rotate every 8 days? What does a piece of 2 tiers down look like exactly for a pro team?


+4 stats, likely +1 mid-range SA or +10/15%


Just to organize thoughts, is this correct ignoring the MVP awards.

A Pro team would get:

1 National Award - Full Season Long [+4 stats, likely +1 mid-range SA or +10/15%]

and

2 Local Awards - 8 Day Rotation [+2 stats, low range SA or +3/+5%]


 
tautology
offline
Link
 
Originally posted by Bort

Kk,

1. We'll run a full season on the test server and tweak our formula until it looks pretty good based on the full season's stats. We do a separate MVP list for Offense, Defense, and ST players, per conference.



Bort,

I am not sure that the test server is the right environment for testing this sort of thing.

real leagues have gutted teams, uber teams, teams that gut halfway through the season, and often quite uneven levels of competition in the two different conferences.

Unless you can simulate all of those elements on the test serves (seems like too much trouble) I think you need to find a way to test against sampling of real live leagues and evaluate the results.

 
DavidD458
offline
Link
 
Also, Bort, (it's off topic and I apologize in advance) as this thread grows and grows, I think a solid idea would be to pick 10-15 people (non-mods, just regular members) and put them in a private forum where things like these can be discussed without growing to 20 pages in a half hour. These get out of control pretty fast and good discussion is gone after the first half hour or so.

I'd love to be a part of it.
 
Bort
Admin
offline
Link
 
Originally posted by kurieg
To calculate MVP stats:

1 - Throw out interconference games.
2 - Select the 9 highest average score differential teams. Only count stats against these teams. Throw one of them out randomly, and use the remaining 8 to compare stats with the 7 that don't make this list. Explanation -
Teams 1-9 have the best score differentials. Team 1 gains fame based on games vs. 2-9. Team 2 vs. 1 and 3-9. Etc.
Roll 1-9 and throw that team out. Team 10-16 are compared to the remaining 8 to gain fame.

Oh ya, don't bother calculating fame until the end of the season so you can weed out the terrible teams/gutjobs.


Like judging figure skating?

Throw out the highs and lows, take the average?
 
MeatBomb
offline
Link
 
Originally posted by DavidD458
Originally posted by RandomBeast

why not only count stats earned while not in cruise control mode?


Solid idea.


That's a money shot right there.
 
Link
 
When losing your endorsements for getting traded or cut, will there be something for the players whose owners sell the team? It just seems like something that might need addressed?
 
MeatBomb
offline
Link
 
Originally posted by Bort
Originally posted by kurieg

To calculate MVP stats:

1 - Throw out interconference games.
2 - Select the 9 highest average score differential teams. Only count stats against these teams. Throw one of them out randomly, and use the remaining 8 to compare stats with the 7 that don't make this list. Explanation -
Teams 1-9 have the best score differentials. Team 1 gains fame based on games vs. 2-9. Team 2 vs. 1 and 3-9. Etc.
Roll 1-9 and throw that team out. Team 10-16 are compared to the remaining 8 to gain fame.

Oh ya, don't bother calculating fame until the end of the season so you can weed out the terrible teams/gutjobs.


Like judging figure skating?

Throw out the highs and lows, take the average?


Would be basically the same as comparing stats against teams in your division that made the playoffs.

 
kurieg
offline
Link
 
Originally posted by Bort
Originally posted by kurieg

To calculate MVP stats:

1 - Throw out interconference games.
2 - Select the 9 highest average score differential teams. Only count stats against these teams. Throw one of them out randomly, and use the remaining 8 to compare stats with the 7 that don't make this list. Explanation -
Teams 1-9 have the best score differentials. Team 1 gains fame based on games vs. 2-9. Team 2 vs. 1 and 3-9. Etc.
Roll 1-9 and throw that team out. Team 10-16 are compared to the remaining 8 to gain fame.

Oh ya, don't bother calculating fame until the end of the season so you can weed out the terrible teams/gutjobs.


Like judging figure skating?

Throw out the highs and lows, take the average?


Actually, I was just thinking about the highs. Only judge everyone against the season's best teams. That's what fans do, anyway.
 
ChuckTaylor
offline
Link
 
all this does is give teams a reason to keep their starters in to force stats to be MVP's
 
Bort
Admin
offline
Link
 
Originally posted by tautology
Originally posted by Bort


Kk,

1. We'll run a full season on the test server and tweak our formula until it looks pretty good based on the full season's stats. We do a separate MVP list for Offense, Defense, and ST players, per conference.



Bort,

I am not sure that the test server is the right environment for testing this sort of thing.

real leagues have gutted teams, uber teams, teams that gut halfway through the season, and often quite uneven levels of competition in the two different conferences.

Unless you can simulate all of those elements on the test serves (seems like too much trouble) I think you need to find a way to test against sampling of real live leagues and evaluate the results.



You're probably right, actually. I could do a MVP list based on last season stats per league from the live site, output in in a text file or something, and post results until it looks good?
 
Page:
 


You are not logged in. Please log in if you want to post a reply.