In the morning, I walk out of my apartment toward my car, and a man is getting into the car to my left. That’s the pure sight/interpretation part of it. It is what is actually happening at that moment. I, however, can make more of an observation than this. Knowing this car belongs to someone, is that the person it belongs to? Does that person live here? Have I seen them before? How is he getting into the car? All of these questions ask me to apply meaning to what I’m seeing and make a judgment based on that. Whatever I decide to, though, won’t be completely backed up by evidence. He could be trying to jimmy the door open, and while that should tell me that he’s a thief, it may also mean he locked the keys in his car and doesn’t want to call and pay for a locksmith. Analysis, even with the best evidence, may be a bit misguided without knowing the entire situation and what went into it.
Later that day, I go to a baseball game. While warming up, the stadium radar gun keeps saying 94, and when I watched him warming up in the bullpen, he had a nice curveball and change-up. He’s the team’s best prospect. By the end of the game, his line doesn’t look so good—5.2 IP, 8 H, 4 ER, 1 BB, and 7 K. My eyes tell me that he didn’t pitch so well. 8 hits and 4 runs isn’t good. Now, my sight tells me that the pitcher threw the ball, the batter hit the pitched ball, and the batter got a hit. The fact that this happened repeatedly tells me the pitcher wasn’t effective. But was he? Two of the hits were decently hit balls that a average defender could have made, and two of the fly balls would have been caught by guys with any speed at all. Wait … read that sentence again. Decently? Average? Guys with speed? Those are kind of vague, don’t you think? What do those mean? We think we know, but what’s the border between weakly/decently/sharply, below-average/average/above-average, and slow/normal/fast? It’s hard for our eyes to tell unless there’s a direct comparison alongside the player, which there almost never is. Then, how do we incorporate that into the pitcher’s performance? Is he getting hit hard? Is the defense bad? Is the defense positioned poorly? Is the batter just fortunate that his swing was the fraction of inch in the right direction when he hit the ball? It’s hard for our eyes to see and interpret what’s in front of us without also having to add such analysis.
None of this is to completely devalue what we see. There’s credence to seeing before you believe. A jump in strikeout rate? Does the guy have a new pitch? Is something breaking more? Has he changed his approach? Did he add deception? Some of these things cannot be seen by statistics. Our eyes can see things, but they are best used when focused on one particular thing. The more we try to see, the less we can fully grasp. I would hate to be a scout who’s there to see a pitcher and two batters who are on the opposite team as the pitcher. When the batters get up, who do you watch? Can you watch both? If you’re watching the mechanics of the hitter, can you see who is responsible for the outcome? Sometimes, it’s easy, but a lot of times, it isn’t. Not only do you have to do this for that game and every game you see, but then you have to compare players. The problem is that you can’t see every game out there, and unless you have seen it, you can doubt the results.
This is the crux of the problem. You can’t see everything, but you have an opinion/analysis of everything. It’s based on observation, and it’s not blatantly wrong. But it’s not empirically correct, either. Stats try to fill in the gaps, and they try to isolate the variables for you. But just like our sight, they aren’t perfect. Stats are given names and meaning, but they are really only formulas. The formula is what we actually see, the brain turning the color and light into an image. Changing it from a formula to a statistic creates meaning, and sometimes that meaning is misguided. So how do we manage this issue? By continuing to use both. Scouts and fans see things, and they should verify those with stats. If they look at stats, they should verify them by watching and observing. If there’s a disconnect, then you have to ask harder questions. What’s the context? Are we seeing a big enough sample in our observations or statistics? The question has never been stats or sight. It’s always been how to properly integrate the two. And no one has been able to perfectly do it. Dayton Moore couldn’t pick wOBA out of a line-up, but he can scout with the best of them. The Rays can run their team incredibly efficiently, and they’ll still miss on guys like Buster Posey. The problem, however, is that sight is hard to argue against. You saw it. Your brain was responsible for the interpretation of what you saw. And sometimes, it’s hard to admit that you might have been wrong. Statistics never have to because they can’t speak.