Variance is a common term in statistics

Statistical variance gives a measure of how the data distributes itself about the mean or expected value. From this distribution it can be worked out where an unknown data point can be expected.
 

Used here for the variance in stage position for the drivers for instance, it indicates how far off the average position the driver has performed. The ideal would be to have a variance of 0, where all your stage positions are equal – for instance 235 stage wins. However, that might be slightly far-fetched.

In reality, if Sebastien Loeb would not have retired in Mexico and Italy, you would see stage positions ranging from 1st to 12th and hence a variance of 3.5.

Another example, let us say 5 stages are driven. If the first four stages were: 4th, 4th, 4th and 4th the variance would be 0. The driver is of course pushing hard for a stage win and unluckily ends up in a ditch on the fifth stage and decides to start under SupeRally system the next day. That gives him a stage position of 55 for that stage and the variance shoots through the roof ending up in 487.7.
 

In conclusion, the closer your stage positions are to your average position, the smaller the variance will be.
 

We are just a few days away from the last rally of the 2012 WRC season, the RACC Rally de España, and we have had a look at last years event. 

Unlike this year, last years Rally Spain was the second to last rally of the year, and Sebastien Loeb took the victory home with a safe margin.

Some drivers, as with other things in life, lasted longer than others. Especially Petter Solberg had a rather short event, with hanging dust sending him off the road permanently already on SS1. Ken Block also had some struggles on that stage, but rejoined the others under super rally for day two only to retire on SS13.

Sebastien Ogier was in for an ok position after some struggles especially on the first day, but SS17 was the end of it for him, together with Matthew Wilson.

Below you have the stage positions throughout the rally for the finishing top 10 drivers.

As most of you know, the flatter and lower the curve is, the better the driver performed. Look at the Sebastien Loeb's red line as an example.

(Click on picture to enlarge)

So based on last years performance, who would you place your money on in this years Rally Spain?

The previous graph only showed us the drivers results for each stage. If we are to make some probabilities out of this, we need to rearrange things a bit.

Perhaps the following graphs might help you out? It shows, based on  last years Rally Spain, the probability of the top ten drivers finishing each stage at top 20.

The narrower the curve is, the more consistent the driver is. For instance, Loeb (red line) has a very narrow curve with a highest probabability just over 2. place. He had 1st position as his best, and 4th as his worst.

Hänninen (yellow line) on the other hand has a very wide curve. This is a graphical display of  his variation from a 5th position as best, to 25th as worst.

(Click on picture to enlarge)

To conclude, numbers are one thing - statistics another. There is a book called "How to Lie with Statistics" which points out that you can not  trust numbers blindly, but at least you will now have a probable outcome of this years Rally Spain - given that the same happens again, right?