Power Ratings - and The Deceptive Numbers That Are Created.

[G-Man!]

This is a Post that was reacting to PR's and Betting Market Spreads by L2KA.
The prior discussion ended without answers - to him being challenged.

"ERockMoney" asked for reasons for my quoted points.
This is why this thread was started as a separate discussion.

What I said to L2KA
Now after your PR ratings myth has collapsed - you are preaching a new angle with "Betting Market Derived Spreads" - Bullshit ..
Then you make a remark to me, like your some super smart capper and tell me to learn something?...LOL

Lines arent made by starting with assumed "Neutral site" games bud. Thats already a bad way to start your made-up numbers for value...
So your hypothesis is starting with no real standard, because youre starting point is already unbalanced.
You've already Dismissed the facts and created a "neutral site theory" which is unfavorable to both the better teams and worse teams.

Every year guys make up Power Rantings and starts with a "guessed number", at a presumed number, BEFORE the team has played any games - then claim to adjust as the season goes.

If your starting point is imagined - then NONE of the adjustments to follow matter, because they are tuned from a bad PR number from the start.
Now if YOU want to learn something - try figuring out what that means. I doubt you can.
Thats why you had no answer in your thread of 'Power Ratings' for all the categories I listed in it for you." END QUOTE!

Curious why you think there’s no standard to the starting point and that it’s unbalanced?

Tell what info/data was used for the first starting point umber? What created it?

Dismissed what facts and why would that be unfavorable to all teams?

Its either giving too much credit to the first Number for winners or not enough for losers or Vice-versa.

Why would adjustments to a number not matter, regardless of if the beginning point was positive/negative?

Because the base number created is automatically wrong which therefore cant be justified by adding or subtracting the updated adjustments. Its continually being compounded by the false starting number and updated by each team without a true number.

If the adjustments were sound, which can be a big if, wouldn’t that correct any issues from the starting point?

You said it correctly. "IF" the number was sound? Its not?
The answer to the first question tells it all. What data was used to create the first base number?


Points to answer.
1. Tell what info/data/perception was used for the first starting point base umber? What created it?

2. Why was the notion of using a "neutral site" as a starting point? What is the actual data that was used?

3. What is the Strength Of Schedule that created the base numbers? Lots of depth to this one.

4. What was the data used to validate first base number?.

5. What are the defensive PR numbers?

6. What are the offensive PR numbers?

7. What makes the very first base number right?

8. How does scoring effect the PR numbers?


9. What detirmens the adjustment times and points.

Many more to post - but lets see what these bring for input and replies.

 

[ERockMoney]

Thank you for the response.

I agree, the starting point is only as good as the data, analysis, work, etc., that are put into it. However, why would too much credit be given to the starting point for winners and not enough for losers, not following that point.

I don’t use these types of ratings personally, but I’ve never been overly skilled in their development and/or the ability to adjust properly from week to week, but plenty of people use them and are quite good at it, several on this very board, although most are long gone these days. Plenty of ways to skin a cat, so they say.

Also not following why the initial number is automatically wrong? Why would that be? In addition, if one is skilled at making the proper adjustments to the starting point, wouldn’t the original number self correct based on said adjustments even if the starting point was off, hence the reason for the adjustment?

I’m not sure if I missed part of this conversation, I only saw your post above, so maybe I’m missing some background here.

 

[Loves2kickass]

This is a Post that was reacting to PR's and Betting Market Spreads by L2KA.
The prior discussion ended without answers - to him being challenged.

"ERockMoney" asked for reasons for my quoted points.
This is why this thread was started as a separate discussion.

What I said to L2KA
Now after your PR ratings myth has collapsed - you are preaching a new angle with "Betting Market Derived Spreads" - Bullshit ..
Then you make a remark to me, like your some super smart capper and tell me to learn something?...LOL

Lines arent made by starting with assumed "Neutral site" games bud. Thats already a bad way to start your made-up numbers for value...
So your hypothesis is starting with no real standard, because youre starting point is already unbalanced.
You've already Dismissed the facts and created a "neutral site theory" which is unfavorable to both the better teams and worse teams.

Every year guys make up Power Rantings and starts with a "guessed number", at a presumed number, BEFORE the team has played any games - then claim to adjust as the season goes.

If your starting point is imagined - then NONE of the adjustments to follow matter, because they are tuned from a bad PR number from the start.
Now if YOU want to learn something - try figuring out what that means. I doubt you can.
Thats why you had no answer in your thread of 'Power Ratings' for all the categories I listed in it for you." END QUOTE!


Tell what info/data was used for the first starting point umber? What created it?

Its either giving too much credit to the first Number for winners or not enough for losers or Vice-versa.

Because the base number created is automatically wrong which therefore cant be justified by adding or subtracting the updated adjustments. Its continually being compounded by the false starting number and updated by each team without a true number.

You said it correctly. "IF" the number was sound? Its not?
The answer to the first question tells it all. What data was used to create the first base number?


Points to answer.
1. Tell what info/data/perception was used for the first starting point base umber? What created it?

2. Why was the notion of using a "neutral site" as a starting point? What is the actual data that was used?

3. What is the Strength Of Schedule that created the base numbers? Lots of depth to this one.

4. What was the data used to validate first base number?.

5. What are the defensive PR numbers?

6. What are the offensive PR numbers?

7. What makes the very first base number right?

8. How does scoring effect the PR numbers?


9. What detirmens the adjustment times and points.

Many more to post - but lets see what these bring for input and replies.

Maybe this will answer some of your questions.

Post in thread 'L2KA NFL' https://www.therx.com/threads/l2ka-nfl.1267551/post-14319527

 

[ERockMoney]

Also wanted to discuss the neutral site comment, I don’t use “power ratings”, but I do also start my analysis from a neutral site perspective. Figuring out true home value is one of the most difficult metrics to accurately portray IMO.

People used to use a flat 3, but that’s clearly inaccurate, I believe the value, on average, was 1.7 last season. However, while that’s an accurate number, it’s still an average of the entire league, so many teams were much less than that and others more than that. Just not enough home games and corresponding data within a reasonable window to project true home value. By the time a number seemingly becomes accurate, it’s likely already depicting a teams performance too heavily based in the past tense and depreciating/appreciating to some degree. Not always, those with limited home values, Vegas, Los Angeles, etc., seem to linger on, but positive home values are much more tricky.

There was a stretch the Seahawks were likely the best home team in the league, but that’s not the case anymore, still solid, but a definite decline. By the time a reasonable set of date pointed to the Hawks incredible home value, it had already started to recede, so you’d be overvaluing Seattle at home for a period of time. That’s the main reason I start from a neutral site perspective and look to home value last, since it’s such an unpredictable moving target.

 

[G-Man!]

Also wanted to discuss the neutral site comment, I don’t use “power ratings”, but I do also start my analysis from a neutral site perspective. Figuring out true home value is one of the most difficult metrics to accurately portray IMO.

Thanks EROCK for the reoply,

Home value is based on statistical differences between road and home games - providing that the teams played were the same. In other words - its reliable when the Home game played against Cincy is also compared with the road game against Cincy by the single opponent. i.e. Pitt vs Cincy play twice a year.
Any the home field advantage is only true, when it is played by the same team two times or more as mentioned above.
If Cle plays Arizona and the AZ team has no QB and CLE blows them out, this Score can't be accurate for a PR effect or adjustment. by using that game for HF advantage,
Its is "assumable" to be better because of the blowout score - but what happens if another team comes in and blows out CLE. One mismatch doest cancel out another in the two outcomes because the teams were completely different.
There is more - but UNTIL the PR is exposed by what it was actually created from - then its not any good.
Im waiting until someone posts HOW they got the First PR number for each team before the season started? Then I can explain where the faulty number was created. Once shown, it can be dissected.

If a PR number is good for a good team then what is the PR factors for a bad team, Using neutral number to start with allows the better team to appear better and the bad team to be worse.
The only way to scientifically use a neutral site starting point - is you must have a large sample of neutral site games with scores nd stats to be compared by the same teams played.

If CLE played all their home games vs losing record teams with negative scoring numbers, then the numbers are way off for ANY HF advantage UNLESS thats all they are going to play.
BUT even this is deceiving - because each losing team must be compared to the other same-division teams that were played in that division.
For the Arizona game to have any kind of factor in the HF PR numbers for CLE - they (AZ) must have also played the other 3 teams in that division like Pitt, Balt, and Cincy. In addition - they need to have nearly identical starters in all those games.
The other problem with those numbers is they are usually not all home games for the division. Thats the only way to get any real number. Cincy played SF in SF while Cle played SF in Cle. etc.

In College Bowl Games this is the biggest factor in finding winners. Its known as COMMONT OPPONENTS played by other common opponents. Directly or indirectly.. Thats a two-step capping factor that I have never seen anyone write about.
Its very time-consuming and one must know how to use it and understand it. During the season its nearly unavlaible because there ware so many opponents that dont play the same teams Only in conference competition are they helpful.
i.e. What you can look for if Alabama is playing Clemson in a bowl game is - how many common opponents did they both play OR were there any other winning BOWL-Levrel teams that they both played?
(Anyone reading this ^^^ should copy and paste and save that secret.)

IN the NFL games this is rarelly available unless two teams from the same division arent playing each other in the playoffs, who also played the same out-out-division opponents, with scores and stats to compare.

People used to use a flat 3, but that’s clearly inaccurate, I believe the value, on average, was 1.7 last season. However, while that’s an accurate number, it’s still an average of the entire league, so many teams were much less than that and others more than that. Just not enough home games and corresponding data within a reasonable window to project true home value.

Thats correct. But again its consisted of league averages and there's its not even a big enough number to give consideration to it. The 1.7 in NOT significant at all because like I stated above - it also was NOT created by the same teams played and cant be accurate because of the home and away games played, with out of division opponents one time per year, falsely affecting the number.

By the time a number seemingly becomes accurate, it’s likely already depicting a teams performance too heavily based in the past tense and depreciating/appreciating to some degree. Not always, those with limited home values, Vegas, Los Angeles, etc., seem to linger on, but positive home values are much more tricky.

There was a stretch the Seahawks were likely the best home team in the league, but that’s not the case anymore, still solid, but a definite decline. By the time a reasonable set of date pointed to the Hawks incredible home value, it had already started to recede, so you’d be overvaluing Seattle at home for a period of time.

They had consistency with a good offense and defense in those runs. Those advantages likely would have been because of true consistency on both sides of the ball.

That’s the main reason I start from a neutral site perspective and look to home value last, since it’s such an unpredictable moving target.

The better outcome is when you can compare the last 4 games to the last 4 game before them and again to the prior 4 games before those.
Its not confusing, but its true that the best caping practices have been by dissecting games by division opponents and out of division opponents. It is a moving target, but there is some advantage in doing the process in segments of games. and Not combining them all together for statistical data comparisons.
But again the neutral site method is unreliable for detailed reasons stated earlier.
My best guess is that if you created numbers without a neutral site hypothesis and more like what Ive shown - as dissecting teams more - you could find a reliable way to get a better edge on the spread.

I have seen for decades that PR's are not good and always fall short in the long run because the disadvantages are unrecognized by the one using unreliable factors - by generalizing the whole league into one assumption of averages to start with.

 

[G-Man!]

Maybe this will answer some of your questions.

Post in thread 'L2KA NFL' https://www.therx.com/threads/l2ka-nfl.1267551/post-14319527

NOT!
As usual you defected answering any direct questions and you have no PR system that can be validated.
Thats why you cant explain the double loss on Buffalo and Over 47.
Your "analytics" non-factor excuse is a joke.
Like I said! Show the factors that you used to start the season with a PR number?

 

[Loves2kickass]

NOT!
As usual you defected answering any direct questions and you have no PR system that can be validated.
Thats why you cant explain the double loss on Buffalo and Over 47.
Your "analytics" non-factor excuse is a joke.
Like I said! Show the factors that you used to start the season with a PR number?

Maybe this will answer some of your questions.

Post in thread 'L2KA NFL' https://www.therx.com/threads/l2ka-nfl.1267551/post-14319527

Read it again there guy lol

 

[G-Man!]

Colorado 14
Washington State 56

Shhhhhhhhhhhh.

I see you love posing only a loss. But you never post all of yours?

Here is the "rest of the story this week". That portion that makes you haters turn into well-known assholes for lying.

G-Man! said:
Tuesday
Toledo -9.5 (-120) over Bowling Green. 3 units.....................L-360
Northern Illinois -4 (-120) over W.Michigan 3 units..............W+300

G-Man! said:
Wednesday
Ohio -9.5 (-130) over C.Mich. 3 units........................W+300
Buffalo +8.5 (-130) over Miami O. 3 units............L+390

G-Man! said:
Boston College +3 (-130) over Pitt. 3 units...............L-330

G-Man! said:
Friday.
UTSA -14 (-120) over USF. 3 units...............W+300
more later

G-Man! said:
Colorado +4.5 over washington St. 3 units.....................L-330

Week day games. 3-4. -510.00
W+900
L-1410

G-Man! said:
Saturday Plays.

G-Man! said:
Noon.
Louisville. PK over Miami Fla. 4 units. ........................W+400
Mia starter QB Van Dyke was benched last week vs FSU. for throwing away games with INTs. He's starting today because QB Williams was injured last week and is out for the season. Louisville defense is just as good as Miami's. look for a low scoring game..

Tulane-9.5 over FAU. 3 units........................................W+300

Penn state -19.5 over Rutgers. 3 units............................W+300

2PM
Appy State +10 (-120) over JMU. 3 units...........................W+300

3:30
Clemson-7 ooer N.C> 3 units..................................................W+300
Georgia -9.5 over Tennessee 3 units....................................W+300

4PM
Ohio State -27-5 over Minnesota. 4 units.............................W+400

8PM
Syracuse +7 over GT. 3 units...................................L-330

More soon
Click to expand...

G-Man! said:
Added
7:30PM
South Carolina-2 over KY. 3 units..................................W+300
Florida +12.5 over Missouri, 3 units.............................W+300

G-Man! said:
More
Oregon St. ML -120 over Washington 4 units...................L-480
Texas -6(-140) over Iowa state. 3 units........................W+300

Sat.
10-2. +2390.00
W+3200.00
L-810
Week 12. Combined 13-6.(68%) +1,860.00

 

[BoutMyMoney]

 

[Loves2kickass]

Nobody cares bro. Carry on

 

[G-Man!]

Nobody cares bro. Carry on

LOL. I never said anyone cared...BUT you sure DO! Those that do simply open the thread for the plays.

Youre the asshole..Oh wait!... and the other assholes... Your eAlias DocHolliday. Big LOU-SER Returns and MCM had been posting shit in my threads over and over and over...
We're all still waiting for how you came up with a PR's before the season started?

 

[jgm4661]

I can give you some insight into how a PR is created before a season. It follows the oddsmakers prediction of win totals for each team for the year. Dan Gordon wrote an entire chapter in his book " Beat the Sports Books' , the best chapter, on how to convert predicted win totals into a PR.

It used to use the 16 game results. An 8-8 team scores 23 pts/gm on avg. So a team predicted to win 8 games at YS you can assign a PR of 23
a 10-6 team , 26 pts, 11-5 , 28 etc. Reverse for teams w 6, 5 wins etc. Add 2.5 for HFA

 

[G-Man!]

I can give you some insight into how a PR is created before a season. It follows the oddsmakers prediction of win totals for each team for the year. Dan Gordon wrote an entire chapter in his book " Beat the Sports Books' , the best chapter, on how to convert predicted win totals into a PR.

It used to use the 16 game results.

Thanks for posting his read on them
Thats the first point of them being wrong.

It doesnt account for Trades, or drafted players or the way a team will perform this year. In addition- The 8-8 record could have been the loss of injured starters i.e QB out for 6 game etc.

That means the 8-8 team with injuries is wrong to use in the assumed mixture of the teams.
We've seen several teams go from worst to first - which completely misuses the established number to start with.
Th lag time in correcting and adjusting any Rating- also is a problem with the current spread. The spread for any week isnt considered for all the other weeks because the teams are different each week. It takes us back to the true difference only when two teams are playing each other again in the same season with the same players.
In addition we have an excellent example of a 6-3 team being out scored on the season playing a team that outscored is opponents. If both team were "26 from being a 6-3 team someone is going to have no advantage in the spread. with the assumed PR number being the same for both teams.
Ive seen decades of PRs that are falsely adjusted each week to "match" the PR number to the difference of the point spread instead of the actual scoring difference of the two teams.

PR guys who posts that Number come up with a PR number exactly what the early lines were a week before the game comes up. That way they spread to be acute when they're not at all.
Their adjustment passes tis weekly with small moves that lag why behind the actual scoring por=tential.
So if the first number is wrong as it is in 90% of then - even as the adjustment is done weekly - the games were already lost and the estimated numbers are only a misleading number to the bettor who claims to use them.

An 8-8 team scores 23 pts/gm on avg. So a team predicted to win 8 games at YS you can assign a PR of 23
a 10-6 team , 26 pts, 11-5 , 28 etc. Reverse for teams w 6, 5 wins etc. Add 2.5 for HFA

So here is the true question to that synopsis.
A team with a 10-6 record is playing another team with a 10-6 record but one team scores 22pts and allows 21 points, while the other team scored 28 pts and allowed 16 points. Which team has been given the wrong (26PR) number?

 

[M.O.B.]

GMAN exposing the fake. If you guys have a system especially your system that cant be explained.....you're a poser. End of debate. Keep up the great analysis and info GMAN. What this guy does for the forum.....very valuable. Give him his flowers now.

 

[jgm4661]

The vegas oddsmakers set the team wins before the season starts is your starting pt. They aren't wrong. They use all the available data to set the # of wins at the time . The same gigantic computers and algorithms they use to set the lines each week. They have billions of $ to invest in info tech so they can post the most accurate #s based on all available information. So no they are not wrong.

You then set your PR at the start of the year. As the season progresses you adjust your PR up/down based on what u list. The linemakers are using rankings and advanced algorithms too similarly when they set the lines.

You may not understand or use ranking systems but there are complex analytics taking place behind the scenes which the linemakers are using but to each his own.

 

[G-Man!]

The vegas oddsmakers set the team wins before the season starts is your starting pt. They aren't wrong. They use all the available data to set the # of wins at the time . The same gigantic computers and algorithms they use to set the lines each week. They have billions of $ to invest in info tech so they can post the most accurate #s based on all available information. So no they are not wrong.

Vegas' Power Ratings are not wrong BECAUSE its their system. But to get a read on the won -loss number - they use public perception to set it - to get what the public believes is a fair number.

So YES you can sy they are right- but only in the sense as to what the public believes. They arent predicting wins and loss. They are using public opinion to create the number of wins -losses the public believes, etc. Thats to get betting action on those numbers and to make a profit.

If they know the public thinks the Steelers wont win over 8 games and they (Vegas) knows they will they will set the number at what ever it takes to get money on the wrong side that goes against their own projected number. That is privileged info we dont have. But Vegas does.
You are right about Vegas have the wide range of data. Thats why they win.. If we had the data and could use it to find the public betting habits, we could then bet the other side.

Its the same as setting weekly lines. But i THEY have the right information as to what a team will do in the scoring, but the public doesn't see it that way, then they set a line to push the public to the losing side of every game they know is going to be bet against their hidden advantage.

If I Told you that Team A was favored by 7 pts over team B and you never had the data to show which side was the bet - then you would guess. Vegas doesnt guess at all. When historical betting habits by the public masses show Vegas when're the money will be sided, they can post a line to make it appear to be a fair line. All the time its made to hold the public on the wrong side.

In a game where Vegas doesnt have a big advantage- thats when they post a line that draws even action -because all the betting history is theirs only and not shared with the public to know. On those games they are simply guaranteed a nearly even volume of money on that game.

The same is true with Power ratings to start the season. They are using historical data with or without algorithims and they set a number to get the public to the wrong side as much as possible.

If you bet the Steelers to win under 8 games and your bet is confirmed... and then 2 weeks later before the season starts they made a trade and got TomBrady before the season starts...What happened to your chance of winning the bet if you originally bet UNDER 8 Wins?
Points spreads and Power rating are all created the same way. Its all about Historical betting habits and the current team potential that Vegas has created the line with.

You then set your PR at the start of the year. As the season progresses you adjust your PR up/down based on what u list. The linemakers are using rankings and advanced algorithms too similarly when they set the lines.

You may not understand or use ranking systems but there are complex analytics taking place behind the scenes which the linemakers are using but to each his own.

I do understand them - but I also know they re no good in predicting spread outcomes. The adjustment phase by the capper, is too late and way behind the curve on the team on game day.

Ive said it previously...Guys who do Power Ratings are not doing anything but adjusting the PR to the spread posted and not for the current potential of the teams ability to score.
That how this whole subject started. One person said points-scored, have nothing to do with ratings!?! LOL...

 

[ERockMoney]

GMAN exposing the fake. If you guys have a system especially your system that cant be explained.....you're a poser. End of debate. Keep up the great analysis and info GMAN. What this guy does for the forum.....very valuable. Give him his flowers

GMAN exposing the fake. If you guys have a system especially your system that cant be explained.....you're a poser. End of debate. Keep up the great analysis and info GMAN. What this guy does for the forum.....very valuable. Give him his flowers

GMAN exposing the fake. If you guys have a system especially your system that cant be explained.....you're a poser. End of debate. Keep up the great analysis and info GMAN. What this guy does for the forum.....very valuable. Give him his flowers now.

I’m guessing you haven’t read much of the thread or your reading comprehension is lacking. Seemed like good discussion in here, leave the fanboy shit out of it.

 

[G-Man!]

The vegas oddsmakers set the team wins before the season starts is your starting pt. They aren't wrong. They use all the available data to set the # of wins at the time . The same gigantic computers and algorithms they use to set the lines each week. They have billions of $ to invest in info tech so they can post the most accurate #s based on all available information. So no they are not wrong.

You then set your PR at the start of the year. As the season progresses you adjust your PR up/down based on what u list. The linemakers are using rankings and advanced algorithms too similarly when they set the lines.

You may not understand or use ranking systems but there are complex analytics taking place behind the scenes which the linemakers are using but to each his own.

I just want ri address the paragraph in bold^.
These are Power ratings in the Playbook newsletter this week.

Look at the PR's for Cle and Cincy
The win-loss factors was so far off now and like I mentioned so many times here- the lag time to update the numbers is way behind.
Its posted in the Newsletter thread here in the NFL forum.

Playbook November 21-27.pdf

drive.google.com

 

[jgm4661]

I just want ri address the paragraph in bold^.
These are Power ratings in the Playbook newsletter this week.

Look at the PR's for Cle and Cincy
The win-loss factors was so far off now and like I mentioned so many times here- the lag time to update the numbers is way behind.
Its posted in the Newsletter thread here in the NFL forum.

Playbook November 21-27.pdf

drive.google.com

There is no lag time. You adjust your PRs daily. Just like the lines adjust every day if there is new information. But yes you are correct you don't follow others PRs that were set at the start of the week if lots of new information has taken place.

 

[G-Man!]

There is no lag time. You adjust your PRs daily. Just like the lines adjust every day if there is new information. But yes you are correct you don't follow others PRs that were set at the start of the week if lots of new information has taken place.

AHHH. But there is a lag time.
You can debate all you want - but all ratings are done after games are played and if you bothered to look at the example mentioned. The Bengals were ranked over Cleveland when they we not winning or scoring and allowing more points.
Remember - like others have said - the first number is based on expected wins? Which is also wrong.

Please give me the PR changes during the week when a game hasn't been played. What kind of 'information' example can you offer and what the effect would be on the PR number change?