To me talent can be seen as height, movement, pace, great accuracy. However, that is only part of the answer as it would have to be backed up by a history of being effective.
I dont think that aspect should be taken out, just backed by evidence of doing well and of consistency.
I think talent can easily be assessed but must be married to real world evidence to have any relevance.
That's what FC statistics are for, no?
Anyway, here's a stab at a selection formula for batting. In this formula:
S is the selection value
F0 = FC average in the current season.
F1 = FC average 1 season ago
F2 = FC average 2 seasons ago
Any seasons spent in Div 2, the FC averaages are automatically halved.
A = Availability percentage for that season (so anyone fit for every game get 100%).
Any season must include 10 games to be deemed sufficient. That also means that current season only becomes relevant once someone has played 10 games in it.
Like I said previously, averages from 1 & 2 seasons ago are discounted to give greater relevance to more relevant performances. So here it is:
S = (F0 x A) + (F1 x 0.75 x A) + (F2 x 0.5 x A)
I haven't yet worked out to quantify previous success or failure in previous test appearances.
Obviously the above only works for batting performances: bowling averages would beed to be inverserly discounted, I suppose. Alternatively, we could use weighted averages - say current season x 2, previous season x 1.5 and 2 seasons ago x 1.
EDIT
I've just realised that the previous comments were primarily about selecting bowlers. Ah well, I'll try and work out the weighted average version for them if I get a quiet moment this afternoon.