Jump to content
British Speedway Forum
Sign in to follow this  
False dawn

Reducing starting position bias in the Speedway Grand Prix

Recommended Posts

During a search for something else speedway related, I tripped over this. As a bit of a nerd and having studied the stats on starting gates (wins and points scored) I was interested to see that someone has tried to apply some science to this subject.

Oh well, as I write this, I realise I might be the only other person in the world interested enough to even read this article :sad:

Click here

  • Like 2

Share this post


Link to post
Share on other sites

It's certainly an indepth study and seems to take in all the factors involved, except rider starting ability. A fast starter will, most times, have an advantage over a poor starter, but that can't be changed. Would be interesting if the authorities adopted Proposed Plan B for a season to compare the statistics.

  • Like 2

Share this post


Link to post
Share on other sites

The problem with using empirical data is that you need to tease out the differences between variances that are due to random chance and those attributable to actual biases. For example if Roger Rabbit who somehow fluked his way into the GP series and came last in every round had draw numbers 2,3 & 4 in most rounds then to what extent is his poor performance caused by being in those positions vs. the extent to which the scores from those numbers are adversely effected by the frequency with which Rabbit appeared in those positions. (see for example http://blog.philbirnbaum.com/2011/08/tango-method-of-regression-to-mean-kind.html in the context of baseball)

The  question is does 152 GPs give us a large enough sample to remove the affect of random chance? It seems quite possible that all but the most egregious of variances will be due to chance.

Some things do stand out from this analysis, notably the huge advantage offered by gate 1 in heat 1 and that you don't want to be in yellow after a track grade. I wonder to what extent biases in gates are damped by overtakes - table 11 suggests a strong correlation between number of passes and the variances shown in that table with Krsko and Prague having much higher variances than Malilla. Interesting to see the input from Phil Morris too. 

 

Anyway I do love a good discussion on Steiner systems (I first encountered them when doing my Oxbridge entrance exam, and immediately saw I was dealing with the 20 heat individual speedway race format).

  • Like 2

Share this post


Link to post
Share on other sites
38 minutes ago, arnieg said:

The problem with using empirical data is that you need to tease out the differences between variances that are due to random chance and those attributable to actual biases. For example if Roger Rabbit who somehow fluked his way into the GP series and came last in every round had draw numbers 2,3 & 4 in most rounds then to what extent is his poor performance caused by being in those positions vs. the extent to which the scores from those numbers are adversely effected by the frequency with which Rabbit appeared in those positions. (see for example http://blog.philbirnbaum.com/2011/08/tango-method-of-regression-to-mean-kind.html in the context of baseball)

The  question is does 152 GPs give us a large enough sample to remove the affect of random chance? It seems quite possible that all but the most egregious of variances will be due to chance.

Some things do stand out from this analysis, notably the huge advantage offered by gate 1 in heat 1 and that you don't want to be in yellow after a track grade. I wonder to what extent biases in gates are damped by overtakes - table 11 suggests a strong correlation between number of passes and the variances shown in that table with Krsko and Prague having much higher variances than Malilla. Interesting to see the input from Phil Morris too. 

 

Anyway I do love a good discussion on Steiner systems (I first encountered them when doing my Oxbridge entrance exam, and immediately saw I was dealing with the 20 heat individual speedway race format).

This post is one of the best I have ever seen on here! You sound just like me!

I have often got into arguments with people over statistics (particularly averages and probability), and they seem to be offended by me questioning whether samples are large enough. I got into a really heated argument on a speedway group when I suggested that averages could never be considered accurate when they will constantly change over an infinite number of cases. Also, statistics are only a reflection of what HAS happened, and not what is LIKELY to happen.

I must admit that I have never heard of Steiner systems (I'm smart, but not an intellectual), but I will definitely look into them, because I am nerdy like that!

 

  • Like 3

Share this post


Link to post
Share on other sites
49 minutes ago, chunky said:

This post is one of the best I have ever seen on here! You sound just like me!

I have often got into arguments with people over statistics (particularly averages and probability), and they seem to be offended by me questioning whether samples are large enough. I got into a really heated argument on a speedway group when I suggested that averages could never be considered accurate when they will constantly change over an infinite number of cases. Also, statistics are only a reflection of what HAS happened, and not what is LIKELY to happen.

I must admit that I have never heard of Steiner systems (I'm smart, but not an intellectual), but I will definitely look into them, because I am nerdy like that!

 

Slightly off tangent but as much as I enjoy analysing rider's averages I've always believed there is much more to a rider's ability than just the statistical....that just tells part of the story. I always use John Davis and Gordon Kennett as parallel examples in that their averages for many years compared favourably with each other's but the way they went about accumulating their points was often very different. Two great talents but Gordon was the better all round rider in my opinion (and I did see a lot of them during theire fomative years at Cowley and elsewhere)

Share this post


Link to post
Share on other sites
24 minutes ago, steve roberts said:

Slightly off tangent but as much as I enjoy analysing rider's averages I've always believed there is much more to a rider's ability than just the statistical....that just tells part of the story. I always use John Davis and Gordon Kennett as parallel examples in that their averages for many years compared favourably with each other's but the way they went about accumulating their points was often very different. Two great talents but Gordon was the better all round rider in my opinion (and I did see a lot of them during theire fomative years at Cowley and elsewhere)

I've always said that statistics do not lie - but they don't always tell the true story behind them.

  • Like 1
  • Haha 1

Share this post


Link to post
Share on other sites
13 minutes ago, chunky said:

I've always said that statistics do not lie - but they don't always tell the true story behind them.

Good analysis.

Share this post


Link to post
Share on other sites

The best thing about this analysis is that it tries to show that a fairer race formula may be possible.
The only remaining problems are the number of variables and the sample size.

Simple then. Well no, of course not. But whatever changes we may make will need to be tested over many meetings, possibly several years, to determine if the new system is effective / fairer.

Share this post


Link to post
Share on other sites

The simple solution to this statistical nonsense is to construct starting grids comprised of a hard surface which riders cannot excavate with their boots such as tarmac or concrete (as in the good old days) and instead of having the start line at right angles to the track, make the line slightly forward from inside to out so that the rider in grid 4 starts slightly ahead of the other grids!! 

Share this post


Link to post
Share on other sites
1 hour ago, Vic Meldrew said:

The simple solution to this statistical nonsense is to construct starting grids comprised of a hard surface which riders cannot excavate with their boots such as tarmac or concrete (as in the good old days) and instead of having the start line at right angles to the track, make the line slightly forward from inside to out so that the rider in grid 4 starts slightly ahead of the other grids!! 

...seem to recall this being a suggestion put forword some years ago? May have been from Olsen?

Share this post


Link to post
Share on other sites
18 hours ago, steve roberts said:

Slightly off tangent but as much as I enjoy analysing rider's averages I've always believed there is much more to a rider's ability than just the statistical....

Rider averages will never be an accurate reflection of true ability because - at least in Britain - each riding position meets the opposition riders a different number of times and under different circumstances. Reserves in particular tend to have easier races in comparison to heat leaders, and the No. 2 position in the classic heat formula generally had easier opposition than the rest of the top 5, although the downside was that Heat 8 was always the obvious place to do a (double) tactical substitution.

But as with cricket, there's all sorts of factors such as when you're put into bat or expected to bowl. Case in point is Freddie Flintoff who's batting and bowling averages were never exceptional by classic all-rounder standards, yet every team in the world would have been happy for him to play for them. :approve:

 

  • Like 3

Share this post


Link to post
Share on other sites
14 hours ago, Vic Meldrew said:

The simple solution to this statistical nonsense is to construct starting grids comprised of a hard surface which riders cannot excavate with their boots such as tarmac or concrete (as in the good old days) and instead of having the start line at right angles to the track, make the line slightly forward from inside to out so that the rider in grid 4 starts slightly ahead of the other grids!! 

Problem is that Gate 4 can sometimes be the best gate on certain tracks and in certain conditions!

Share this post


Link to post
Share on other sites
12 hours ago, steve roberts said:

...seem to recall this being a suggestion put forword some years ago? May have been from Olsen?

Didn't he also have an idea to put a chicane in the bends... :rolleyes:

Share this post


Link to post
Share on other sites
7 minutes ago, Humphrey Appleby said:

Didn't he also have an idea to put a chicane in the bends... :rolleyes:

...or loop the loops!

Share this post


Link to post
Share on other sites
2 hours ago, Humphrey Appleby said:

Rider averages will never be an accurate reflection of true ability because - at least in Britain - each riding position meets the opposition riders a different number of times and under different circumstances. Reserves in particular tend to have easier races in comparison to heat leaders, and the No. 2 position in the classic heat formula generally had easier opposition than the rest of the top 5, although the downside was that Heat 8 was always the obvious place to do a (double) tactical substitution.

But as with cricket, there's all sorts of factors such as when you're put into bat or expected to bowl. Case in point is Freddie Flintoff who's batting and bowling averages were never exceptional by classic all-rounder standards, yet every team in the world would have been happy for him to play for them. :approve:

 

And I knew a rider through a sponsor who liked to finish his season on 5.9 as a utopian figure..

He was always regarded as a "solid six point rider" so felt any promoter looking for such (weren't they all?), would see the "potential" in his scoring for them

He never was wanting for a team place but used to usually "struggle" a wee bit, (after a good start and middle), towards the end of the seasons when/if  his teams couldn't win anything..

A promoter also once told me he would never sign anyone who had been a particular standard through the year, and then had a great last month of the season, as he wasnt convinced it was exactly a step change in the capability of the rider, more a reason of getting some "gift" points of others..

There are lies, damn lies, and statistics.. :D

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
Sign in to follow this  

×

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. Privacy Policy