Reassessed Averages

[stevebrum 6,824]

Was it?

 

Two people attended, one said CVS mentioned Ford, the other said he didn't.

 

CVS and Adams always mention Ford and Poole!! However if wasn't mentioned that Ford instigated the 9 point rule.

 

Be honest, was these words on your mind in 2016.

Those words I wish for very single season so yes obviously I want transparency in the rules every single season that are set out clearly at the start of the year and no sudden we will change or amend this rule because we hadn't thought it.

I can only assume you are referring to Woffy coming back on his real average??

It was all within the rules wasn't it???

If they had regraded Woffy and all GP standard riders as a 9 at the start of 2016 I'd have applauded it.

For me it's always about what's best for our sport and league, not what's best for my club.

Fairness and clarity would make things so much better for the fans and BSPA IMHO.

[GWC 495]

 

Actually, it is very difficult.

 

For example Jason Doyle's UK average is 17% higher than his Poland average, which you would expect as Poland is a stronger league. Freddie Lindgren's UK average is 23% higher than his Poland average.

 

However, Chris Holder and Max Fricke's Poland averages are very similar to their UK averages, Holder's being within 1% and Fricke's 6%.

 

Therefore any calculations would have winners and losers and the smart teams will always be looking to sign the anomalies.

Indeed but as it confirms Doyle and Lindgren are the better riders (currently) so a premium will apply to them.

If you have a rider who is on the way down (Holder) and on the way up (Fricke) then does it matter a lot when building a team?

Thats always been the beauty of speedway hasnt it - riders improve and so does the team.

Going the other way - tough.

Avoiding average manipulation and petty squabbles amoungst erstwhile promotions is more important to the sport that is currently going downhill fast.

[MattK 3,447]

Indeed but as it confirms Doyle and Lindgren are the better riders (currently) so a premium will apply to them.

 

 

It doesn't confirm that. All it confirms is that they rode well in the UK. Lindgren's Polish average is very similar to Holder's Polish average (7.54 v 7.09)

 

 

If you have a rider who is on the way down (Holder) and on the way up (Fricke) then does it matter a lot when building a team?

 

The fact that Holder's averages in Poland and the UK are similar (7.09 v 7.16) only shows that Holder had a poor season in the UK, not that he is on the way down.

 

The point I am making is that there are no clear patterns when comparing Polish, Swedish and UK averages, so it is very difficult to come up with an algorithm which will accurately predict how a rider will perform in the UK based on how they have previously ridden in foreign leagues (which is what assessed averages are effectively trying to do).

[dontforgetthefueltapsbruv 10,706]

 

It doesn't confirm that. All it confirms is that they rode well in the UK. Lindgren's Polish average is very similar to Holder's Polish average (7.54 v 7.09)

 

 

 

The fact that Holder's averages in Poland and the UK are similar (7.09 v 7.16) only shows that Holder had a poor season in the UK, not that he is on the way down.

 

The point I am making is that there are no clear patterns when comparing Polish, Swedish and UK averages, so it is very difficult to come up with an algorithm which will accurately predict how a rider will perform in the UK based on how they have previously ridden in foreign leagues (which is what assessed averages are effectively trying to do).

Rank the averages by differential.

 

As with any sample removing a proportion of the top and bottom scores should even out any unusual performances and leave a sensible figure to use.

[hyderd 1,216]

Those words I wish for very single season so yes obviously I want transparency in the rules every single season that are set out clearly at the start of the year and no sudden we will change or amend this rule because we hadn't thought it.

I can only assume you are referring to Woffy coming back on his real average??

It was all within the rules wasn't it???

If they had regraded Woffy and all GP standard riders as a 9 at the start of 2016 I'd have applauded it.

For me it's always about what's best for our sport and league, not what's best for my club.

Fairness and clarity would make things so much better for the fans and BSPA IMHO.

Yes I am referring to Woffy 2016, at the time I didn't like what happened but accepted it as it was within the rules at the time,

If the BSPA had proposed a 9 pt assessed average for GP riders without a 2017 average at the same time they announced the one 8 point rider per team then fair enough, they didn't they let it slip after Zagar tweeted his interest of returning because of the new rules he had heard of ie, fixed race nights.

[MattK 3,447]

Rank the averages by differential.

 

As with any sample removing a proportion of the top and bottom scores should even out any unusual performances and leave a sensible figure to use.

 

That would get you closer to a fair multiplier, but there will always be anomalies.

[waytogo28 2,054]

MattK wrote " The fact that Holder's averages in Poland and the UK are similar (7.09 v 7.16) only shows that Holder had a poor season in the UK, not that he is on the way down."

 

That is one way of looking at it. but you have to have faith and hope that in 2018 his drooping average does not turn into an alarming slide.

[MattK 3,447]

MattK wrote " The fact that Holder's averages in Poland and the UK are similar (7.09 v 7.16) only shows that Holder had a poor season in the UK, not that he is on the way down."

 

That is one way of looking at it. but you have to have faith and hope that in 2018 his drooping average does not turn into an alarming slide.

 

No, that's not one way of looking at it. It is the ONLY way of looking at it if you use the averages he gained in Poland and Sweden. You are trying to introduce some kind of supposition into the equation, which is exactly what an algorithm is there to eliminate.

[arnieg 3,649]

No, that's not one way of looking at it. It is the ONLY way of looking at it if you use the averages he gained in Poland and Sweden. You are trying to introduce some kind of supposition into the equation, which is exactly what an algorithm is there to eliminate.

Straightforward really.

 

Step 1: Get a data set of riders with averages in both leagues.

 

Step 2: do linear regression to come up with a best fit.

 

Step 3: apply formula to riders whose averages you wish to assess.

[Skidder1 7,637]

4. check the differences in machinery used in each country and apply the appropriate multiplier/divider!!

[SCB 0]

 

It doesn't confirm that. All it confirms is that they rode well in the UK. Lindgren's Polish average is very similar to Holder's Polish average (7.54 v 7.09)

Id say it shows that a rider scoring 0 form 3 in one of his 11 meetings will heavily drag it down.

[GWC 495]

4. check the differences in machinery used in each country and apply the appropriate multiplier/divider!!

NO! Its 500 cc engines on big and little tracks ...

[B.V 72 1,054]

Straightforward really.

 

Step 1: Get a data set of riders with averages in both leagues.

 

Step 2: do linear regression to come up with a best fit.

 

Step 3: apply formula to riders whose averages you wish to assess.

Der - how can that be a better way than the tried and tested independent one mans own opinion of what every riders assessed average should be system.Even if he does change his mind every year.Move with the times man .

[g13webb 4,254]

I've always believed the riders averages create more arguments than any other area. Bearing in minds the frequency of changes to our racing program, it just proves how inaccurate these averages have become. Because of this, they become out of date the moment the season finishes.....

 

Sooner, rather than later the BSPA will realise the benefits of using the grading of all riders to assemble equal teams..... This could be done using current form and/comparative date from other leagues prior to the AGM, so everybody would be aware of all the possibilities available, when discussing the following season...

 

It would do away with, this '1over8s' ruling and allow manager to permutate their own ideas. The numerous calculations, the relentless arguments surrounding the strength of the league when riders last rode here, and the Ringers will become ghosts from the past. All it would needs to set it up is one meeting to grade all the riders, which each would retain for the whole year, whence they again become regarded... . Injuries and replacements become so simple. Replace a grade 3 with another grade 3,

 

This simple idea would solve so many issues and create a new understanding for moving the sport forward...

[dontforgetthefueltapsbruv 10,706]

I've always believed the riders averages create more arguments than any other area. Bearing in minds the frequency of changes to our racing program, it just proves how inaccurate these averages have become. Because of this, they become out of date the moment the season finishes.....

 

Sooner, rather than later the BSPA will realise the benefits of using the grading of all riders to assemble equal teams..... This could be done using current form and/comparative date from other leagues prior to the AGM, so everybody would be aware of all the possibilities available, when discussing the following season...

 

It would do away with, this '1over8s' ruling and allow manager to permutate their own ideas. The numerous calculations, the relentless arguments surrounding the strength of the league when riders last rode here, and the Ringers will become ghosts from the past. All it would needs to set it up is one meeting to grade all the riders, which each would retain for the whole year, whence they again become regarded... . Injuries and replacements become so simple. Replace a grade 3 with another grade 3,

 

This simple idea would solve so many issues and create a new understanding for moving the sport forward...

These gradings to be arrived from 'using current form/comparative data'

 

That sounds like averages to me.....