Posted by Brian Joseph on Saturday, June 20, 2009 at 7:55 pm

Sometimes I think there are baseball fans of the sabermetric sort that would rather watch FanGraphs’ Live Scoreboard than actually watch a game of baseball. This isn’t a knock on how people choose to enjoy the National Pastime, just an observation. Heck, having seen Adam Eaton pitch more times than I care to remember, there have been times I wish I wasn’t actually watching the game.

Those who love the numbers of the game are often refer to sabermetrics and almost treat as a way of life when discussing how they choose to enjoy the game. Wikipedia defines sabermetrics as the analysis of baseball through objective evidence, especially baseball statistics. While this is a simplified definition, I always found the definition ironic. The notion that sabermetrics is truly objective is silly when there are a number of ways to “objectively” look at a situation statistically depending on your subjectiveness toward the game. Take player value, for example. Some prefer VORP, others look at WAR and others consider Win Shares. Each serves a purpose and each way to evaluate players has its following and detractors. So, it is truly not objective.

Recently, a few discussions and happenings in baseball compelled me to write this. Most of them are statistical in nature but are also impacted by actually watching the game and realizing that while statistical analysis is vital to understanding the game, the study is interdependent on what is seen on the field. Let’s walk through four of those rather quickly:

Messing With The Johan

Last week, Johan Santana had the worst outing of his career. The Mets were pounded by the Yankees and most of the damage was done on Santana’s watch with Johan allowing nine earned runs in just three innings.

The impact? Santana went from a 2.39 ERA to a 3.29 ERA. He also went from 8-3 to 8-4 in the W-L column (even though we’ve already been told by many that W-L records are meaningless, right?).

At this point in the season, Santana would have to throw 31 scoreless innings to return his ERA to that impressive 2.39 number. So, even if Santana threw three consecutive complete game shutouts, his ERA would still not be as good as it was before he allowed nine earned runs in three innings. Does that make sense? Statistically, it does. Hence, if you evaluate Santana based on his ERA alone, even though Santana could go 3-1 with three complete game shutouts in a four game stretch, because of how poorly he performed in that one outing, we’d have to assume that he was a better pitcher ERA-wise four starts ago.

When actually looking at the game itself, there was a spot in the game that Santana’s reputation may have cost him one or two earned runs which were rather meaningless in a 15-0 loss but meaningful when looking at Santana’s ERA.

In the fourth inning, after Santana allowed a two-run homer, double and single, Derek Jeter stepped to the plate. With the score 6-0 at the time and two runners in scoring position, many lesser pitchers would have been removed at that point. Instead, Santana stayed in, Jeter singled which scored a run and Johan was removed. Both of those runners scored off of the Mets’ bullpen (and many more) and instead of a three inning outing with seven or eight earned runs, Santana allowed nine. In the grand scheme of things, on that day, did it really matter if he gave up seven, eight or nine earned runs? The outing was awful no matter how you look at it.

Since it is fresh in our minds, it is easy to consider these factors. However, when we look at Santana’s season three years from now, it is doubtful this outing will be considered when evaluating Santana’s overall performance. But it does have a huge impact. If Santana allowed eight earned runs, it would have taken 27 scoreless innings to restore his ERA to 2.39 and seven earned runs would have been even more forgiving at 23 scoreless innings to restore his ERA to 2.39. Again, on that day, did it really matter if Santana allowed seven earned runs or nine? Only statistically.

Wieters, Jesus and PECOTA

Another particularly useful function of sabermetrics is in the realm of projections. Often recognized as the best of the projecting bunch is Nate Silver’s PECOTA found at Baseball Prospectus.

Personally, I’m a fan of BP’s annual even though I don’t completely agree with all of the PECOTA projections (but who can say they do when there are thousands of projections made about the 2009 season alone?).

Sometimes, PECOTA is frighteningly close (see last year’s projections for Evan Longoria) and sometimes not so much (there’s a good chance that Zack Greinke will outperform his PECOTA projections in 2009 which have him at 12-10 with a 3.96 ERA.).

Then there’s the Matt Wieters 2009 PECOTA projection.

In case you lived under a rock, the projections on Wieters after what was considered the greatest minor league performance of the last four decades (according to the annual’s write-up on “Orange Jesus”) were off the charts! How off the charts? 31 homers, 102 RBI, 105 runs scored and a batting line of .311/.395/.544! No wonder some were surprised that Wieters didn’t levitate from the on-deck circle to the batter’s box when he made his Orioles debut.

Now, I’m not here to criticize PECOTA or BP or projections, in general. My question is how could people actually believe that such feats were feasible from a 23-year-old catcher who had not played above Double-A ball? Considering the fact that the projection also called for Wieters to notch 649 Major League plate appearances in 2009, it was easy for me to realize that while this projection should make me excited about what Wieters will eventually do in the Majors, I shouldn’t expect this much production out of Wieters.

Why is that? For one thing, only one catcher (Russell Martin) reached 650 plate appearances in 2008. Also, the Orioles spent some money on a capable backup catcher in Gregg Zaun. Plus, it was made pretty clear that Wieters would not open the season on the Major League roster due to the way Major League roster rules and free agency works.

So far, Wieters has one homer in 61 plate appearances and a batting line of .259/.295/.414. Obviously, in this small sample, he is nowhere close to his PECOTA projections… based on last year’s 530 plate appearances in the minors. It is worthy to note that Wieters projections would obviously have been altered by his 163 plate appearances at Triple-A Norfolk and likely brought the earlier projections down a bit.

You can’t fault BP for putting out the PECOTA projection. The numbers are what they are. You take a fairly consistent projection system and these are the numbers revealed by it, what else can you do?

What you can fault is those who took the projections too literal. Inevitably, Wieters will likely not live up to those projections and some in the hyper-critical sabermetric cult… err… community will bash PECOTA for the miss. Another one of the projections out there on Wieters will be closer and that group will wear that projection as a badge of honor and use it as their calling card as to why their numbers cruncher is better than yours.

Whether it is CHONE, Marcel, ZiPS, Oliver or whatever projections are floating out there, there are always going to be black eyes that need to be covered up by the spot on projections that act as make-up for their blemishes.

While none of the Wieters projections were as grand as PECOTA’s call, they were all pretty clear that we should expect Wieters to be a special player. Now, we’ll have to see if the “Orange Jesus” monicker is sarcasm or honorable.

Butterflies and Brad Lidge

Finally, there’s the matter of Brad Lidge’s performance this year and his overall effect on the Phillies. Statistically speaking, before Lidge went on to the disabled list, there was speculation about whether or not the Phillies should stick with him in the closer role and either force him to cop to a phantom injury or switch roles with excellent set-up man Ryan Madson.

The theory, independent of many non-statistical variables, was that Madson was performing well statistically and Lidge was not, so swap the two or take Lidge out of the equation all together and the conclusion will be a better Phillies team.

The point became somewhat moot when Lidge hit the DL with an actual injury but the results show that statistics alone cannot tell us everything.

First of all, so far Madson has not been as effective as the closer as he was as the set-up man. In simple terms, the switch from Lidge to Madson has been, at best, a wash.

The butterfly effect of the move reached further than just the switch of Madson from set-up to closer. The move essentially affected every member of the bullpen as the roles throughout the bullpen had to be shuffled to accommodate the switch. Without getting caught up in the numbers, the bullpen in Lidge’s absence, as a whole, have been ineffective independent of Madson’s performances. While Lidge was pretty much awful as the closer until he hit the disabled list, the bullpen was very effective.

While the cause and effect of bullpen roles has been hotly debated for many years, in this specific instance, it obviously did not benefit the Phillies to have their relievers switching roles.

It’s Chaos I Tell You!

What it boils down to is that baseball is less tied to statistics and more tied to chaos theory. While statistics obviously play a major role in our understanding of the game, understanding the impact of chaotic behavior on the game is just as important.

Because of the complexities of the game, no statistic is unaffected by its environment.

Consider ballpark effect and its usage to attempt to level the playing field when studying player performance. It often comes into play when talking about players who have impressive offensive seasons in hitter-friendly parks. Matt Holliday is the first name that comes to mind but any home run hitter on the Phillies would suffice, too. However, to understand the deficiencies of park effect, one need only look at the way it is calculated and the fact that the numbers are interdependent by who is playing and pitching at that park in a given year.

Did I lose you? Let me simplify…

Let’s use Roy Halladay as an example. We all agree that Halladay is an elite pitcher, right? If not, stop reading and go watch video of him pitching and then come back to this. Welcome back! Now we are at agreement, right? Good.

Last year, Rogers Centre favored pitchers slightly. Those numbers are based on 81 games in Toronto of which Halladay was the starting pitcher for 15 of them. No other stadium was impacted as much by Halladay being a part of the equation as the Rogers Centre was. While the park effect is often considered to be a great way to measure the impact of a stadium on the game, it is impossible to truly measure such a thing without recognizing that it is inherently flawed by the players and pitchers that impact the numbers put up in said stadium. The park and the player are interdependent, no matter how you slice it.

From the way the infield is cut to the way an opposing pitcher tears up a mound to his liking to equipment choice to umpire to what that player had for breakfast, lunch or dinner could all have an effect on what happens on the field.

Simply put, it’s not all statistical. If you believe it is, you sometimes end up making silly arguments and defending them with such attitude that turns the casual fan off and keep many from seeing it as anything more than a “cult” rather than a tool to better understand the game.

It also leads people to make statements like “Placido Polanco, Mr. Underrated” and then basing their “statistical” evidence on RAR which is basically a rating. (And ironic that a Tiger would be good when it comes to RAR!)

I know, I know. I’m hard on those who love sabermetrics. My guess is while I love the numbers of the game, I will never be truly accepted in the sabermetric fraternity. But, at the end of the day, you can’t understand baseball just by looking at the numbers. The statistics of the game are too malleable to make an iron-clad complex argument without someone else manipulating the numbers slightly to fit their hypothesis. And no matter how snarky you are in your commentary or how sure you are in your conclusion, there’s another way to look at it.

It’s chaos theory, at it’s best. Too bad it sometimes brings out the worst.