Time For A Rethink In Rolling Averages?

[Dekker 239]

try 32 with all matches included.

Edited September 9, 2012 by Dekker

[george.m 3,063]

Oh, it happened all right.

 

If I remember right the majority if not all (I stand to be corrected) Glasgow fans on BSF were, utterly appalled and embarrassed by what happened and said so on here. We didn't hold back and riders and promotion got both barrels. What is becoming tiresome is the selective memory of some on here (not suggesting you) and others who weren't members at the time assuming otherwise (not suggesting you). The Glasgow fans on here really don't need their noses rubbing in the dirt, we felt the hurt more than anyone else. Trust me! You're pushing at an open door.

 

If you weren't on here at the time, trust me we really weren't happy and made that quite clear. We paid good money and transport costs to get to Armadale too!

[bigeddiechek 1,094]

It's all systemic when it comes to averages. I agree that 24 is a sensible figure.

 

Although, I would like to point out, that average is, well, a pretty average function. We could be more sophisticated, or we could just bin them altogether.

[YAMYAMBANDIT 560]

try 32 with all matches included.

Question ?

What happens when a rider rides in all 3 leagues, Elite, Premier & National DEVELOPMENT League. Which rolling aversge do you apply, or as it is a rolling average, should all points scored in each League by added together and averaged out.

NL scores 15 points, PL scores 8, EL scores 4 totals 27, rolling average becomes a 9 ave.

[bigeddiechek 1,094]

You would take the average of each average, then the average of that, if you know what I mean?

[YAMYAMBANDIT 560]

You would take the average of each average, then the average of that, if you know what I mean?

The conclusion is then, everyone is average :drink:

[Tsunami 10,219]

You would take the average of each average, then the average of that, if you know what I mean?

 

In statistics, you should never average an average. Weighted average is more appropriate and there's less chance of distortion.

[SCB 0]

Question ?

What happens when a rider rides in all 3 leagues, Elite, Premier & National DEVELOPMENT League. Which rolling aversge do you apply, or as it is a rolling average, should all points scored in each League by added together and averaged out.

NL scores 15 points, PL scores 8, EL scores 4 totals 27, rolling average becomes a 9 ave.

Speedway fans amaze me. If a rider has a NL, PL and EL average then he would use his NL average for the NL, his PL average for the PL and his EL average for the EL. It's common sense and no need for anything else.

[crescent girl 1,903]

Some of the above is getting silly (on the BSF?) and distorting a serious discussion.

 

I'm in favour of rolling averages, and the 38-meeting history is fair, as it represents two full seasons.

 

Don't change it.

 

Who's next?

[bigeddiechek 1,094]

In statistics, you should never average an average. Weighted average is more appropriate and there's less chance of distortion.

 

I was attempting to out-stupid Taffy, but failed again...

[YAMYAMBANDIT 560]

I was attempting to out-stupid Taffy, but failed again...

 

Oh Scott. the irony of that statement :icon_smile_clown:

[bigeddiechek 1,094]

Oh Scott. the irony of that statement :icon_smile_clown:

 

Mission Impossible, Bob.

 

One thing everybody misses about rolling averages is that the longer they are used, and the more meetings a rider completes, the more accurate they become.

Surprised this has not been mentioned.

[Fromafar 10,332]

Mission Impossible, Bob.

 

One thing everybody misses about rolling averages is that the longer they are used, and the more meetings a rider completes, the more accurate they become.

Surprised this has not been mentioned.

Now Now Big Ed an average is an average don't change your tune now. Edited September 13, 2012 by Fromafar

[bigeddiechek 1,094]

Now Now Big Ed an average is an average don't change your tune now.

 

If you read my post, the average, as a function, is set in stone.

What I am saying is that the more measuremets you take, the more accurate an average is.

 

Now, this little example will demonstrate the point, and the one you didn't grasp a couple of weeks back:

 

Let's take two measurements: 5.8, 7.6.

Average = 6.7

 

Let's take 10 measurements: 5.8, 7.6, 5.8, 7.6, 5.8, 7.6, 5.8, 7.6.

Average = 7.2

 

So, as you can see, the number of measurements (or rides, in this case) is crucial.

 

 

This point is to prove that 97.98274928374% of all statistics are made up.

[Fromafar 10,332]

If you read my post, the average, as a function, is set in stone.

What I am saying is that the more measuremets you take, the more accurate an average is.

 

Now, this little example will demonstrate the point, and the one you didn't grasp a couple of weeks back:

 

Let's take two measurements: 5.8, 7.6.

Average = 6.7

 

Let's take 10 measurements: 5.8, 7.6, 5.8, 7.6, 5.8, 7.6, 5.8, 7.6.

Average = 7.2

 

So, as you can see, the number of measurements (or rides, in this case) is crucial.

 

 

This point is to prove that 97.98274928374% of all statistics are made up.

See you have a bit of time on yours hands today .The only the thing you said a couple of weeks ago was "an average is an average" you really should listen to the whole story before giving answers .A bit of Tap Dancing going on now

 

If you read my post, the average, as a function, is set in stone.

What I am saying is that the more measuremets you take, the more accurate an average is.

 

Now, this little example will demonstrate the point, and the one you didn't grasp a couple of weeks back:

 

Let's take two measurements: 5.8, 7.6.

Average = 6.7

 

Let's take 10 measurements: 5.8, 7.6, 5.8, 7.6, 5.8, 7.6, 5.8, 7.6.

Average = 7.2

 

So, as you can see, the number of measurements (or rides, in this case) is crucial.

 

 

This point is to prove that 97.98274928374% of all statistics are made up.

What do you teach again?