(The reason that I suggest this goes in Off-Belay rather than the the Walls and Training section will be explained at the end...)
Having become a regular at Kendal Wall over the last year, I decided I wanted some way to track whether I was improving, staying the same of going downhill. So I started recording the lines I did each session; but then wanted to 'score' each session, so I would have a single number which I could compare from session to session. So I developed a simple algorithm to evaluate each day's results: firstly, I awarded 1 point for a 4, 2 points for a 4+, 3 points for a 5, 4 points for a 5+ and so on, up to 7 points up to a 6b; my current ceiling but it's obviously open ended. Then I weight that score by a factor according to the length of the line; any route on the main wall I award double points, any route on the King Kong section I weight by a factor of 1.5; all the rest of the lines just get the standard points.
Finally, I wanted to take into account if I tainted a route, so I knock off a point if I have a rest; I didn't make it any more complex than that, (just one point deducted even if I rest 2 or 3 times), and also don't count any routes that I don't get to the top of; they might have been legitimate training but it just becomes too complex.
So, a 5+ on Main Wall would get 8 points; a 6a+ with a rest on King Kong would get (6 x 1.5) - 1, i.e. 8 points also, and so on. I've got a spreadsheet and also an Access database to do the sums. My best day so far is 81; I had a session last week the day after a big walk and a bit under the weather and I got a miserly 42. (Yesterday was better with 66, I could have gone on but my partner had had enough. I haven't started correlating how well I do with different partners, that's obviously a possibility but perhaps best not!)
Now, my two points are this. 1) Is this a useful exercise, or just a bit of pseudo science, and too crude to be meaningful; if too crude could it be made more meaningful in a practical (i.e. simple) way?
2) (And the reason I've posted here) is that it IS an interesting exercise for anyone who relies on stats, and particularly stats recorded and processed on a computer - which is to say, all of us, and all of them. Everything from the basic algorithm, to the accuracy with which data is collected, to the accuracy which it is entered, is a potential weakness. (Interestingly, when I achieved a disappointing score the other day, my first reactions were that I had made a data entry error, or should tweak the algorithm; er no, giving surprising results is exactly the point of the exercise!) Big decisions are made and justified by systems with all the same potential weaknesses.
There will be people who will say 'Duh, obvious!' OK then; try it for yourself. Come up with a better algorithm, then try and collect data accurately, bearing in mind a single error of a single digit (recording a route on the Main Wall for instance) can vastly inflate achievement in a way that won't be obvious.) How many times does that happen in the real world?