That's the way it sounds on paper, in reality doing it that way around will just open the door to confirmation bias.
Think of it this way, you're looking for someone to pass the ball around like Alonso in midfield. If you go and look at lots of players, with your eyes, you'll immeadiately get a sense for whether they have it or not. You're not just looking at the completion %, or the sideways/forwards ratio, or long/short ratio. You're actually looking at the pass in all its context, the way they passed it, what was going on in the game when they passed it, were they looking at the players around them, what did it say about their character, was it an intelligent pass, was it a mindless kick aimed at another red shirt, was it appropriate to make that pass given the way the game was panning out, what does that say about their intelligence. All these sorts of things tell you about the player, and help you do what Rosco above is chatting about.
If you just use stats to find players with good passing statistics, you'll get 50 candidates. The real passer who you need will be listed 37 out of 50. You'll only send a scout to the top 10, all of whom are cunts. So what will happen? Your scout will uhhmm and ahh, and say ah well, maybe, uhm you know, yeah they're ok I guess, yes. Because he wants that sweet cash bonus that Macca was chatting about. It's a conflict of interest, and it will lead to confirmation bias.
That it basically why you should always do it the other way round, and go looking for number 37 out of 50 the hard way.