Charging the Mound: Labels

-----Original Message-----
From: jeff@rotowire.com
Sent: Tuesday, July 17, 2012 7:56pm
To: "Christopher Liss"
Subject: Charging - Labels

It's human nature to create classifications to explain what we don't understand. Life's easier in our minds if we can classify people or certain phenomena as part of a group, put a label on it, and move on, with that label explaining what we need to know. But all too frequently that explanation is inadequate at best, and often false. There are many issues in life where this is a real problem that rises to a much higher concern than fantasy baseball or baseball itself, but this is what we do, so it's still important.

Let's delve into some of those labels, and the players associated with them, to see if they (a) exist, (b) can be identified, and (c) can be exploited for our purposes.

1. STREAKY PLAYERS

Do streaky players exist? I know that streaks and slumps happen all the time - there's no disputing that. But are there some players that are prone to streaking and slumping than others? I've seen a lot of notes, on RotoWire and elsewhere, describing a player as a streaky player, or a feast-and-famine player. Is there something about that type of player that gives him more peaks and valleys than other "Steady Eddie" players? And how would we measure that sort of player?

My inspiration for this question is Pedro Alvarez. He homered Monday night against the Rockies, giving him 18 on the season. He's had a couple of surges this season where he's had the majority of his homers, but he's also had massive slumps where the strikeouts have come in big bunches. Is my perception that he's more streaky because he strikes out a lot and because he's a power hitter? Are strikeout pitchers more likely or less likely to be classified as streaky than pitch-to-contact pitchers?

Moreover, is there any way we can act on this? Is a streakier hitter inferior to a steady one? Is it possible to time the market if we can actually identify which players are streakier than others?

Who else fits that framework, if you believe such a player exists?

2. FIRST-HALF/SECOND-HALF PLAYERS

I'm less inclined to believe in this label than the other two here. There's no reason why a player should improve or decline based on the turn of a calendar. A player can wear down, or the weather in one half of the season can be a big factor, but I don't believe that players simply fit that framework. I cringe anytime I see a list of second-half surgers or faders, or when I get asked to provide said list. Adam LaRoche was supposed to be a slow starter, until he came flying out of the gate this spring. Dan Haren still gets listed as a guy that fades in the second half, conveniently ignoring the big second half he had when he got traded to the Angels in 2010.

3. CLUTCH PLAYERS

You and I have addressed this before, and while I think that some semblance of clutch or lack-of-clutch exists, I also think that gets played up too much. The lightning rod for this conversation in the past has been Adam Dunn, and in recent years Matt Holliday has caught his share of heat.

Forgetting for a second the sample size problems with this classification, let's tackle this from another angle. How many times in close-and-late situations does a batter face a specialist pitcher, brought in specifically to tackle that hitter's weakness? And how often do teams save their best relievers for that exact situation? Couldn't that explain a hitter's failings in close-and-late situations a lot more than as a character defect?

What other labels come to mind for you? Who else fits or breaks those labels?

-----Original Message-----
From: "Christopher Liss"
Sent: Wednesday, July 18, 2012 8:28pm
To: jeff@rotowire.com
Subject: Re: Charging - Labels

I think there are streaky players. Alfonso Soriano is another that comes to mind. I'm sure there's a way to measure it - bigger standard deviations in WAR per month would probably suffice. Those with more consistent month-to-month splits would be considered less streaky. Of course, streaks could be weekly, too. Or even game to game. A guy who goes 4-for-4 one day and 0-for-10 the next three every four games would look very consistent month to month or even week to week.

But whatever metric we use - assuming we figure out something that makes sense, then yes, we could act on that. You'd trade for streaky players if you needed to make a big move in the standings and for more consistent ones if you were comfortably ahead. But that begs the question a bit. Maybe the "Steady Eddie" goes off for two months, and then falls into the category of "streaky" under the metric. Still, I think some players are streaky, and I'd guess they depend more on physical talent than fundamentals. Soriano has made a solid major league career out of hitting balls out of the strike zone. The more fundamental guys - who have good plate discipline and do the right thing - might be more consistent. But again, first we'd have to define the terms and do some research.

I'm also less inclined to believe in players partial to first or second halves as a matter of course. I suppose it's possible some guys need more at-bats before they get into a groove, or warmer weather, and others wear down mentally due to the travel schedule. But it's easy to make up explanations for what's happened after the fact and harder to come up with them before it happens.

As for clutch and unclutch, if your explanation as to why some hitters struggle late in close games is that they face specialists designed to exploit their weaknesses, then we have a pretty good idea of who you want to count on when the game is on the line: players with less pronounced weaknesses. If someone has a huge platoon split, then he's not clutch. If a hitter has almost no platoon split, can hit breaking balls as well as fastballs and hits to all fields, then he's less likely to encounter adverse circumstances when the game is on the line. But that still wouldn't explain why certain players improve under pressure. Maybe it's dumb luck, or maybe some are able to handle pressure better than others. Because everyone's under pressure, including the pitcher, that would give you an advantage over him.

In the end, every effect has a cause, and whether a player deserves a particular label depends on whether that cause is repeated in most circumstances, or whether it's particular to the circumstances that obtained over the span you're observing. For example, if a particular player tended to wear down in the Texas heat, and that was the cause of his second-half fades, you could call him a first-half player so long as he played in an outdoor park where it tends to get particularly hot and humid in the summer. But if he moved to San Diego, then the label would no longer apply.

Other labels that get thrown around: "Quad A Players" for guys who mash in the high minors but fail in their (usually short) stints at the majors, players with a "closer mentality," "Professional Hitters," guys who hit for average but not much power, "Innings Eaters" and "Utility Players."

-----Original Message-----
From: jeff@rotowire.com
Sent: Thursday, July 19, 2012 2:27am
To: "Christopher Liss"
Subject: Re: Charging - Labels

Whatever metric that we'd use would have to define both the length of strength and the amplitude. It's a lot like our previous discussions about breakouts - it's not just that a player is breaking out, but the magnitude of the breakout. I also have to imagine that the bigger the platoon split, the more likely it is that a player is streaky.

The other issue with explanations about first-half / second-half players is that the players aren't just static objects, they aren't cards with defined values. If a pitcher has worn down in the second half for a couple of seasons, chances are he'll take measures to counteract that.

I like the "Quad-A" label - and of course, I have to think of Bryan LaHair. I can't help think of the comparison to Chris Shelton, unfortunately. LaHair is better than that, but he also has the clear weakness (platoon split) that results in the high K-rate. But he also has had enough success to have graduated from the Quad-A perception to a "flawed major leaguer" instead. Which is another label. But I digress.

-----Original Message-----
From: "Christopher Liss"
Sent: Thursday, July 19, 2012 3:40am
To: jeff@rotowire.com
Subject: Re: Charging - Labels

I think that point is important - that players aren't static objects - and one that's often lost when you look at stats and try to discern patterns for too long. Someone who used to be streaky might no longer be so a couple years later. Someone who used to strikeout a batter an inning might pitch to contact as he loses velocity, ages and learns to pitch. The question then is threefold (length of streak, magnitude of streak and the tendency of streaky players to continue to be streaky generally). It might be that when you look at the data, only 50 percent of streaky players in Year x are still streaky in Year x+2. So we'd need to know that, too, before we had anything actionable.

I think that latter point is the biggest variable we're missing. If a player has a huge ground ball rate, or great line drive rate, or is deemed to be exceptionally streaky by our metrics, how likely are those tendencies to continue the following month, year and for his career? It seems people throw around all kinds of statistical analysis that's necessarily backward looking and treat it all as if it's equally predictive of how a player will be going forward. But I'm fairly sure each of these tendencies has its own persistence rate, some highly predictive and some almost random.