So, Branch Rickey was the first to start using a 2-8 scale to handle this problem. Why 2-8 you ask? Well, if you assume that baseball players are normally distributed, then we can start with a 5 being the average player.

So, 5 is your average Joe. 50% of players will be better than he is, and 50% of the league will be worse than he is.

How do we assign the next two closest values, a 4 and a 6, well, since we aren't working with hard numbers and only the eyes of the scout, we need some metric that we can roughly apply with the eye to indicate skill. So, back we go to the normal distribution and leverage the standard deviation of a normal distribution which has been well established as follows:

Population covered by n standard deviations from the mean:

Code: Select all

```
Deviation from the mean % of population covered
1 68%
2 95%
3 99.7%
```

Code: Select all

```
Rating Standard Deviations
2 - 3
3 - 2
4 - 1
5 0
6 + 1
7 + 2
8 + 3
```

This is the beauty of the system, it really reveals just how many players in the league are basically average and indicates how difficult it is to judge the difference between a MLB average player and a guy who is a solid, but not spectacular player. 68% of all players in MLB could be average or slightly better, say an All-Star season or two during their careers and after that its up to you to judge their ML performance to decide who's worth keeping and who is not. But, much like real life, when you get to a "7", you're talking about top 5 maybe top 10 player in the league, which is a much smaller range than our 6, which should be somewhere between average to an occasional all star. As an interesting side note, many of you will be familiar with the scouting terms "plus" and "plus plus", which on the 2-8 scale indicates a 6 and a 7. This is perfect and very revealing to what's going on because a 6 is plus 1 standard deviations from the average.