They say numbers never lie. But they lie all the time. They lie when polls say that the Kardashians are less popular than many frightening diseases and Congress yet their twitter followers and TV money reach larger and larger sums. Numbers (like 42) lie when we don’t ask the right questions. Without the right question, a number is a tool without a wielder.
Baseball has always been about numbers. 755. 61. 73. 762. 2,131. Each of those numbers tells a story. Some might say that one or two of them even lie. But what none of them do is tell the whole story. They are useful instruments that punctuate the rich history of the game, but inside each of these numbers is a thousand more numbers that tell a thousand more stories. So it is with that in mind that I jump back into the debate about lineup protection. Some say there are no numbers to support such a thing exists. Some say they may just be looking at the wrong numbers.
"Tulo makes a lot of people better, especially me." - Carlos Gonzalez 4/2/2013
The Denver Post triumphantly notes that the Rockies are 178-154 when Cargo and Tulo are in the starting lineup, and the ROOT sports broadcast seems outwardly giddy over the reuniting of the tandem of Carlos Gonzalez and Troy Tulowitzki. After their performances so far, who wouldn't be?
|2013 - Troy Tulowitzki||.333||.300||1.000|
|2013 - Carlos Gonzalez||.400||.400||1.000|
It is interesting to me, however, that they put so much stock in the effect of Tulo’s return on Carlos Gonzalez without ever mentioning the swaths of stat gurus who argue that there is no such effect. I have also noticed that those who believe lineup protection (LP) is a myth also seem to accept their own opinion as a given fact and move on. This is a huge baseball debate where neither side is actually acknowledging the other. I on the other hand find myself firmly in neither camp and so I go searching for answers.
In part 1 of this examination, I posted some tables that showed a dramatic reduction in the number of fastballs Cargo saw without Tulo in the lineup. To trust or not to trust this stat is a microcosm of this whole debate. Is he seeing fewer fastballs because, without Tulo in the lineup, opposing pitchers are less fearful of the damage of walking him? Is it because they know he will chase worse pitches if he feels he has to be the only offense? Or is it because of a myriad of other factors including pitchers being generally more careful with a hitter who continues to prove his abilities? Is the number telling us the whole story?
Is everything that happens in baseball measurable?
Purple Row member Dutch1008 pointed out in the comments section after part 1 that the effects of lineup protection may be impossible to measure. For example, he mentions, the opposing pitcher may be focusing solely on his approach to each hitter until (perhaps) a full count. Once at that point a pitcher may decide to throw a different pitch to Cargo based on who’s on deck, or who isn't. Regrettably, finding stats for how many times a guy got a fastball or a strike in a 3-2 count with a certain person on deck (or not) and what the results were, can be an arduous task.
Even finding those numbers would leave you with a small sample size that would be largely inconclusive since the numbers would only apply to one hitter, wouldn't give you other factors (defense, ballpark) and could therefore be considered correlation and not causation. And even then, what if the pitcher wasted the 2-2 instead of the 3-2? Then we need a whole new set of stats.
The problem is that stats only tell us what happened, and not necessarily why. The most common reaction I got when I first waded into this topic was essentially one of belief that if the players think it’s a factor, it is one. That makes some sense to me. If the hitter and the pitcher are both mindful of it, and Carlos Gonzalez in particular has mentioned several times feeling more comfortable with Tulo in the lineup, then LP certainly has to be some factor right?
Most of the smart baseball people who disagree would probably say that these players are fooling themselves with placebos and beliefs in things they have always been told are true but that there is no evidence of. But the assumption here is that anything that is a factor on the game must show up in the advanced stats.
So let me ask this. Are advanced statistics perfect? Do we have as many as we need? Might there be room to add more? If so, we must be open to the possibility that events occur in the game of baseball that affect the outcome, or show a player’s value, and do not show up obviously in the stats we already have available. Purple Row member Oldfoagie put it excellently in the comments section after part 1,
"One of my gripes with a lot of the new stats as opposed to the old ones, are how they are purposely designed to ignore situational hitting and pitching. To a lot of stat heads, situations are just statistical noise. To me they are the difference between an effective player ( success 30% of the time ) an an ineffective one ( 20% of the time )." -Oldfoagie
New And Improved Statistics?
Stats like BaBiP and FiP are both fun and informative. They also assume a great deal, both essentially boiling all of offense down to either home runs or dumb luck. Many would argue that some of the best skills in baseball include placing a solid opposite field single or getting a sacrifice fly with one out in a close game. Most stat-heads hate RBI but how else do we measure the impact and value of guys who understand the need for, and have the ability to, lift the ball into the outfield at the expense of potentially more solid contact knowing that it puts your team in a better position to win?
I had thought about this conundrum before and suggested to many (though I should give some credit again to Dutch1008 for pointing it out) that an RBI% stat could maybe be the perfect compromise. RBI% balances out other factors that can distort RBI the same way FG% vs PPG in basketball can help differentiate between efficient scorers and prolific scorers. If you measure the percentage of RBI chances converted you get a much clearer indication of a player's ability to create runs regardless of his luck in getting the chances to do so. A guy with only 40 RBI but 82% RBI% might be a better run producer than he appears, just unlucky.
And herein may lie the key to solving the dilemma of lineup protection. Rather than saying RBI is too flawed a stat, lets get ignore it now, we should be looking to improve and alter it. Rather than assuming proof of LP will emerge in current stats if it exists, we should be defining the parameters of what lineup protection would fully entail and see if any numbers back it up. It may be shown in a combination of stats. Maybe a new one will need to be invented, or an old one improved like with RBI%. It's not as simple as saying if there is such a thing as lineup protection, then player A should hit more home runs with player B in the lineup and walk more with him out of the lineup.
Unfortunately I feel the task of defining the parameters of lineup protection may be a bit beyond my abilities. I could certainly speculate (and am about to) but the truth is that I would want to talk to numerous major league hitters and pitchers and ask where they believe proof could be found. Is it in the pitchers approach? Fewer fastballs? Fewer first pitch strikes? Does it not become a factor until later in the count? With runners on base? Can lineup protection be a factor several spots down (i.e. are you more careful with pitches to the entire middle part of a particularly potent lineup)?
Then wouldn't the whole thing be much more of a factor for NL lineups? For 8 hole hitters in NL lineups? What about on the hitting side of things? Are you more aggressive knowing that the lineup around you is weaker and you may need to provide most of the offense? Less aggressive because you think they may just walk you? Do you find pitchers approach you differently depending on who's on deck? How so?
Understanding lineup protection, if it exists, would have to come from crunching many conversations like that with even more questions. Maybe everyone would have different answers and approaches making it nearly impossible to measure. Some players may walk more without their protection and others may strike out more because they start reaching. Some pitchers may be more aggressive thinking "if I can dominate this guy I own their whole lineup." Some may get way less aggressive thinking, "why let this guy beat me?" Are these scenarios immeasurable? How do you measure the impact of an idea that could affect everyone differently but still affect everyone?
From the pitchers, to the players, to the manager, to the guys and gals on ROOT sports, everyone seems to agree that the Rockies are a scary team when Carlos Gonzalez is standing in the batter's box and the pitcher gets even more nervous looking over at the on deck circle. I think I'm going to have to go with the guys in between the lines on this one. Are they really just fooling themselves?
There are too many assumptions about which of the already anointed stats lineup protection would show up in and ultimately the statistical argument ignores as many potential factors as RBI. It could be that lineup protection is a myth if it doesn't show up in WoBA or WRC+ just like it could be that hitting over 100 RBI means you are a good power hitter. Or both of these scenarios might not tell the whole story.
Yes I’m going with the players on this one. Maybe they are fools ...but numbers are liars.