If you subscribe to the common belief that keeping a player's attributes within a specific ratio is optimum for success, then using a bit of math to answer this type of question is quite helpful. Step 1 is figuring out who should be playing right now. For this we want to use something I call the Effective Overall Rating (EOR) which is the total of the important stats for a given position modified by the desired attribute ratio. It's calculated as MINIMUM ( Attribute 1 / Ratio 1, Attribute 2 / Ratio 2, ...) * (Ratio 1 + Ratio 2 + ...). So if you want your goalies to be trained as a 2:1:1, then player A's EOR is MIN (101/2 , 40/1 , 37/1) * (2+1+1) -> MIN(50.5,40,37)*4 -> 37*4 = 148. Doing the same thing for the remaining 3 players gives us A(148), B(132), C(112), D(136), therefore I would use A for now and have D as backup.
Step 2 is figuring out who has a future on the team and who should be released. For this I calculate the Effective Quality (EQ) as described in this wonderful article:
glanvalleyeaglets.wordpr ... . The EQ is essentially the weighted harmonic mean of the player's important qualities and is a true measure of their rate of advancement where a simple average is not. Something the article doesn't mention but is very important is to adjust the EQ based on the player's CL. In a post somewhere on the forums, the manager Stan reported that at CL 5/6, a player trains 10% slower then he would at 6/6, and at 4/6 it's a 20% reduction. Thus the EQ's for each player (as modified by their CL) are A(48.97), B(43.18), C(52.16), D(56.00). What this tells us is that D will advance the fastest and B the slowest - at least until the end of the season when some players may drop in CL and thus some recalculations may need to be done.
So let's list out each player by EOR and EQ for comparison:
A - 148 - 48.97
B - 132 - 43.18
C - 112 - 52.16
D - 136 - 56.00
A should be the starter though he will eventually be passed up. D should be our backup, but will move to starter before too long, especially considering his age should keep his CL high for a few seasons. C could possibly become a backup someday since his EQ is higher than A's, but he will never be a starter. Both his EOR and his EQ are lower than D, so he will never catch up (barring some extreme injuries to D). B has no future on the team. His current EOR is too low to play, and his EQ is too low to allow him to ever catch up to the others - he will only fall farther behind with time. So get rid of B, and if you really only want 2 goalies on the team, then C would be the next in line for a pink slip, though you should probably keep him around till he passes A, and then you could send A packing.
Something to note is that comparing players by EQ assumes you're going to keep them fairly close to a perfect attribute ratio. For example A is going to train slower then his EQ would seem to indicate since he would have to train in his two lowest qualities for quite awhile to get back into the desired ratio, assuming you want to train him to a 2:1:1.