Subject: [youth] 29 January

depends on length of time. The longer they stay the less likely they're going to have a talent of exactly 3. Quite possible to have a talent of 3 if they're only there for 6-7 weeks popping 2 times. I wouldn't discount the possibility of having a talent of 3 for a much longer time but I would imagine it would be pretty rare for long periods of time. Those in shorter distances who have talent 3, might not have talent 3 but might have 3.2 or something but haven't been in the youth school long enough to show their true talent.

That's why talent will always be an estimate for my liking whether you have plus graphs or not :).

If someone has had a player with eight 3-week pops followed by a 4-week pop, that shows that his talent is somewhere between 3 and 3.125, which gives a nice upper bound on minimum talent. If someone has eight 3-week pops between 4-week pops, that shows his talent is somewhere between 3.111 and 3.125.

From a strictly scientific point of view, all you can say if you've had n 3-week pops and nothing else is that the talent is between 3 - 1/n and 3 + 1/n. The evidence strongly points to 3, and probably even 3.1, as a lower bound, but you can't actually prove it unless you can look at the code.

Me, I've never had a trainee with less than 4 weeks' talent, but I've only had my incredible youth coach for about a month (-:

Yup, no doubt it's rare. I've had only two 3-week juniors in 10 seasons of SK. And Ruini came soon after I started while Croft is a recent aquisition, so the frequency may be about one per 9 or 10 seasons!

Plus-graphs are notoriously unreliable because they can be faked without much difficulty. This was discussed at length in an international forum thread called 'Fake Talent Warning' by user 'Losing' which was posted on Dec 18th. I'm always a bit suspicious when a seller goes to the trouble of presenting a graph. Especially if it shows some 2 week pops! :)
(edited)

I had have only 3 bellow 4 (but I suspect that I had some before but the coach was not good enough)

I see you managed to pick up Pierre-Olivier, after all. Good on ya, mate! Now, let's see you get him into U21 in a season or two. Just don't let him come back to haunt me in a cup tie or a Q game! ;)

Good luck with him...

I don't think it's all that useful to apply stat. analysis to determine the exact value of a 'talent parameter' in a system where (3-1/n) can never be <3.0... With a relatively small number of data points, if you get a 3.0, take him and run. He's talented! :)

Btw, I pulled Ruini with an Incredible coach, so you could land a <4 anytime soon. If you don't, perhaps you too should complain to someone in Warsaw!

if you get a 3.0, take him and run. He's talented! :)

Heck, at this point I'd be satisfied with a couple of fours!

the exact value of a 'talent parameter' in a system where (3-1/n) can never be <3.0

My point was that, without looking at the actual code, we can never know what the lower bound on talent is. Scientifically, we can be increasingly sure that it's somewhere around, say, 3.1, but all it takes is a single firm data point to throw it all out. Too many people don't realise that science isn't about proof, it's about disproof (and falsifiability). Nobody can prove the Theory of Gravity, say, but after many millions of experiments have failed to disprove it, it's a useful working assumption (-: