Logic of Ratings, Scales
 

It seems that the less granular ratings scales each correspond with the1-100 scale. This would only seem to make sense to me if the particular scale (say-, 1-10) chosen was the basis for sorting players, and more "precise" numbers were a way of making estimated comparisons of players within the particular "tiers" of the scale being used (here, 1,2.3...10). For example, if a scout rated four players (A,B-C-D) as (respectively) 7-7-8-8, that would indicate that he rated A and B in one tier (say, "Good") and C and D in another (say, "Plus-Plus") on his 10-point scale. No scout who wanted to keep his job would give his GM these ratings if they were based on his respective 1-100 scale ratings of 61-70-71-79; if he had that much confidence in his ability to separate players by one-hundredths of a scale, then he would present the ratings as such. On the other hand, if he first categorized the players on the 10-point scale within which he felt that he could articulate meaningful distinctions between players, THEN he might offer the 1-100 ratings as his best estimate of where the players fell within a given tier- with the understanding that the "tens" digit (thus, the "7" in "71"was the one of confidence. In this case, the GM would understand that B was rated in the same tier as A and C in the same tier as D- NOT that B was 8 points better than A and only one worse than C (because otherwise the GM would not want 1-10 ratings showing B as being more similar to A than C.)

The related issue is that if a scout rates a player X as more similar to Y than Z on one scale (say, 1-20), than he should not appear as more similar to Z than Y when using a different scale (say, 20-80), as this would merely be the result of where the scales' respective rating "breaks" fell on the 1-100 scale. If the scout had that much confidence in such granular distinctions, then he would present his reports in that form; he would not instead group players according to the vagaries cutoff-point differences between "cruder" rating systems.

Sorry for going on so long.