The projection systems are a representation of how much the inputs matter. I mean, it's common sense.
- more data is better than less
- more recent data is better than older data
The ideal prospect has great recent data and a track record of good data over time
The worst prospects suck now and sucked before
The tough calls are the guys with a solid track record but bad recent performance, or vice versa
Getting the "tough calls" right separates the Rays front office from the Rockies/DBacks/Pirates/Tigers/Orioles.
What the data is specifically also matters. When it comes to recognizing breakouts or declines, teams will want to look at indicators and not really results. They care about bat speed, velocity, exit velocity (max and average), spin rates, contact rates... they don't care that much about results. Projection systems are incorporating some of this data as it seeps into the public sphere. Half a season of massively increased contact and max EV might mean more than four prior years of mediocrity.
No matter how you slice up Samad Taylor, he's a "tough call"
And by the way, there's no need to dress down what Mejia did in the minors with language like that. The kid hit for average and power, without striking out, and with very good age vs. level figures. He was a statistical stud. These are the indicators that matter the most in the minors: age vs. level, ISO, BB and K rates, contact rates (which are obviously heavily correlated to K rates), and place on the defensive spectrum. The worst things you could have said about Mejia as a prospect were: the BB rate was just average, the ISO was only above average, and catchers have a precarious place on the defensive spectrum. He had a precious ability to make contact and get hits at a very young age, relative to his peers.