I believe that one should always cut back minutes when a rating goes red.  Views vary on how low to go.

I set each skill at a practice level that will avoid significant loss of rating - no more than one point.

I then give priority to blue skills in core areas.

Then I allocate minutes to a blend of other core areas and some blue skills in noncore areas.

Posted by spasticity on 2/26/2014 3:15:00 PM (view original):
Yeah I think it's very likely a curve such that, technically, diminishing returns kicks in almost immediately.  Most people probably think of the phrase "diminishing returns" as the point where that effect gets really noticeable and prohibitive.  I would say that point is 25 personally.

Keep in mind there are multiple variables at work as well -- the minutes in the practice category, which may or may not be linear, and the bell-curve of growth rates depending where on the 1-100 line the atribute itself is...putting 20 minutes of practice time on a blue attribute that starts at 35 is going to get more gain than that same 20 minutes in a blue attribute starting at 75 because improvement comes quicker near 50. That's going to skew perceptions of whether or not there's a linear rate of gain. (when comparing across multiple players, differences in work ethic can also come into play)

zbrent's comment above does make me rethink my position though, because we do know there's a "balance point" for each attribute where the minutes put in are simply maintaining the attribute...it isn't going to increase, but it also isn't going to decrease. For instance, if 7 minutes is the balance point in BH (ignoring for now the excellent analysis that has been done showing that for many players it is actually lower than that), then if each minute is linear 20 minutes (13 over the balance point) should result in more like 4x the rate of growth as 10 minutes (3 over the balance point). I don't think any of us have seen that sort of difference, so there has to be some regression in the impact of minutes.

^ Good point.

I think that's where work ethic really works into the equation. At least in my experience it seems like the players with more work ethic need less minutes to prevent losing ratings. 

Posted by rednu on 2/26/2014 3:57:00 PM (view original):

Keep in mind there are multiple variables at work as well -- the minutes in the practice category, which may or may not be linear, and the bell-curve of growth rates depending where on the 1-100 line the atribute itself is...putting 20 minutes of practice time on a blue attribute that starts at 35 is going to get more gain than that same 20 minutes in a blue attribute starting at 75 because improvement comes quicker near 50. That's going to skew perceptions of whether or not there's a linear rate of gain. (when comparing across multiple players, differences in work ethic can also come into play)

zbrent's comment above does make me rethink my position though, because we do know there's a "balance point" for each attribute where the minutes put in are simply maintaining the attribute...it isn't going to increase, but it also isn't going to decrease. For instance, if 7 minutes is the balance point in BH (ignoring for now the excellent analysis that has been done showing that for many players it is actually lower than that), then if each minute is linear 20 minutes (13 over the balance point) should result in more like 4x the rate of growth as 10 minutes (3 over the balance point). I don't think any of us have seen that sort of difference, so there has to be some regression in the impact of minutes.