Cherry Picking

In the final 14 appearances one season, he threw 74 innings, allowed only 50 hits and posted a sparkling ERA under 2. After the season fans were excited about what this pitcher could do the following year. Ignored in those 74 innings was 1) he walked a lot of guys, 2) was considerably lucky in that his BABIP and HR/FB was well below expected levels and 3) those 74 innings were only a small portion of the much larger body of work he had produced in his career. Still, fans saw the ERA and more than likely a decent FIP and were excited about the following season.

What should have been surprising to few people managed to catch many off guard. That same pitcher had been in decline for several years and the following year he didn't post the pretty looking ERA fans had thought possible. He didn't post an ERA or FIP as good as the previous eason combined. He didn't improve. He didn't apepar to be better in any way whatsoever. In fact, as is true with most people as they age, he continued to show decline.

All due respect to everyone, but this was something that even a monkey forecasted. How did so many miss it? It's easier to talk about why some did understand that what he did in the last part of 2010 wasn't indicative of a new pitcher. Some people saw the declining numbers and understood that the previous said appearances were helped in ways that nobody could expect to continue into the following season. There had been signs of decline evident for years. The end of season numbers were inflated by good luck that most likely could not be sustained. The best bet was on this pitcher regressing considerably. Actually, the best guess was that he'd regress all the way toward his career numbers and then worse.

It really couldn't have been any clearer. All the hope in the world wasn't going to change this pitcher's career trajectory. Sustained luck may have, but if luck could be sustained it wouldn't be luck. It would be a skill. This pitcher's career wasn't over by any means. He wasn't bad by any stretch of the imagination and it's not like we're talking about someone who was going to be out of the game in a week. There was just no way he was going to pitch as fans thought he did at the end of the previous season.

We knew this because you simply cannot take a 74 inning sample and use it as evidence of change when you have a much larger sample to work with. Which is more valuable: those 74 innings or the thousands before it? Imagine sitting around one day flipping a shiny nickely you find in your pocket. It's raining outside and you're allergic to water so you have nothing to do but flip this coin. You flip it several hundred times, marking down the result each time. You add it up and find that that half the time it lands on heads and the other half tails. Now imagine sorting through those hundreds of flips. You notice that there was a stretch of 4 consecutive heads. There was one of 5 consecutive tails. You notice this one stretch of 10 flips that resulted in 8 heads.

If you'd notice any of those stretches, would you have applied greater value to those flips than the hundreds before it? No.

Performance in baseball isn't exactly a coin flip. There's no talent in flipping a coin. It takes considerable talent to play baseball and talent is always changing. Pitchers learn new pitches and batters become more disciplined. Say you have a pitcher who threw exactly 200 innings for 6 straight years and the run environment each year is the same. His ERA was 3.25, 3.50, 3.45, 3.75, 3.65, and 3.95. One thing is clear here: the pitcher has posted higher ERAs as time has progressed. In other words, he's gotten worse.

Let's say that in the final third of that 3.95 ERA season (his most recent one) he posted a 2.15 ERA. Let's say he threw 75 innings over that span. What's more valuable: the 1200 innings of him getting worse each season or those 75 innings of him being better than he's ever been? Obviously the answer to this is those 1200 innings and it's not even close.

You say maybe he learned a new pitch. Maybe he did. You say he's throwing more strikes. He probably did since his ERA was so low, but I guarantee you I can find another 60 or 70 inning stretch over those 1200 innings in which his walk rate was nearly the same as it was to end the most recent season. Just as you could find stretches of that coin landing on heads or tails as if one is hotter than the other, you are going to find numerous stretches over 1200 innings that makes the pitcher look completely different than the totals. During some stretches he'll look like one of the worst pitchers in baseball. In others he'll look like he's punching his ticket to the Hall of Fame.

While it is possible that pitcher picked something up in those final 75 innings that made him a better pitcher, it's much more likely it's similar to that stretch of 10 coin flips in which 8 landed on heads.

Reports from various people are that he's doing something different and can sustain it, you say? Well, I'd bet you $1000 that you could say the same thing about those other hot stretches. In fact, I'll bet you that it happens almost every time someone is hot. A hitter is 28 for his last 56. What are you hearing? You're not hearing about how it's lucky. It's about he's using a new bat, gripping it differently, has a slightly different stance or any other number of reasons why. What you're not hearing is this: it's luck. You're not hearing from the scouts, announcers and coaches about how his BABIP is better, his line drive rate isn't sustainable, how he can continue to keep the ball in the park as often as he has. No, you're hearing about how he's locating that pitcher better, which leads to fewer home runs. You're hearing about how that fastball is runing in on the righties rather than drifting toward the center of the plate. The pitcher has a new grip. He's now throwing a mph faster with more movement. The hitter is seeing the ball better. He's able to drive the outside pitches forcing pitcher to come inside. The fielder has worked tirelessly on defense. He has a new glove. He's throwing more over the top. He's stepping directly at the 1st basemen. He 2nd basemen is applying the tag better. He's receiving the throw from the catcher more efficiently.

We hear these all the time and they don't mean a damn thing. You could take any of that and replace it with this: we don't know what the hell is happening so we're going to focus on something some people believe is happening but probably isn't.

But this one guy improved considerably. Yes. it happens. Sometimes it just comes out of nowhere. Sometimes the batter or pitcher does make changes or improves in ways unforeseen. Most of the time we hear the above excuses it's nothing more than something to talk about. It's nothing more than an explanation of what can't be explained.

In another season on a different team, another pitcher threw 120+ less than impressive innings before being released. He had an ugly ERA. His win-loss record was among the worst ever in that amount of playing time. As with our other player, some things were ignored: ridiculously high BABIP, an elevated HR/FB rate and one of the worst LOB% I can remember for a pitcher with more than 100 innings. Also ignored was the tremendous difference between his ERA and FIP. The FIP was nearly 3 runs less and as a result, he was worth 2 fWAR that season. League average is 2 WAR over a full season so based on FIP this guys was above average and yet he was released.

All the other teams let this guy go buy without much of a thought. It would take a progressive organization to see how unlucky this pitcher was. One that could look past the ugly ERA and focus more on why it was so high and why he was a better pitcher than the results. He would come at a very low cost too. It's the type of signing that progressive teams take advantage of: undervalued players.

While 29 organizations did fail to look past the ugliness of his stats, the Cubs did not. They signed him to a minor league contract, eventually added him to their active roster and were rewarded with a season that saw him produce more than 2.5 fWAR in under 130 innings.

That same season the Cubs picked up another guy who was released. This one was a defensive specialist. He played in only 23 games down the stretch on a contending Cubs team, but produced .8 fWAR. He was re-signed the following year and produced 1.5 fWAR. He was paid only $1 million.

That's two contracts the Cubs gave out during one season for undervalued players. They were players other teams passed on, which made them so cheap. Neither was a great player by any means. Neither was ever going to play in the mid-summer classic either. Both were undvervalued by other teams allowing the Cubs to capitalize on their value and they did so on a contending team.

We could have looked at these transactions as evidence the Cubs were a progressive front office, but that would have been absurd. For one thing, two contracts doesn't mean anything. Especially not when previous management signed signed many more different players. But using these two as an example as the Cubs being forward thinking is essentially using the same percentage of sample size as those who did when they thought Carlos Zambrano's 74 innings to end the 2010 season were a good sign of things to come.

The Cubs made a couple solid decisions in 2004 when they picked up Glendon Rusch and Neifi Perez. Both were quality contributors to the Cubs for little money. They also made many other quality decisions. In fact, if we looked at the ration of good decisions to bad decisions we'd find that past management was probably 1 to 1. Maybe even higher than that.

The point here is that looking at such small samples provides little to no predictive value. If someone noticed those two signings in 2004 they may have incorrectly thought the team was moving in a different direction, but they'd have been wrong. Those who thought Zambrano could even come close to maintaining the numbers he posted at the end of 2010 were wrong. And more than likely, Jeff Samardzija as a starter is going to fail miserably. Samardzija's overall career tells us far more than the 40 or so innings he finished the season with last season. Not to mention that those were in relief and now he's going to be pitching in a far more difficult role.

As berselius has pointed out, even as a reliever Jeff Samardzija's expectations are below replacement level. It's hard to understand why a team would give a guy like this more innings. This is the kind of guy who you limit the number of innings he's given. Imagine if Koyie Hill was named the starting catcher at any point in the last few seasons. That's pretty much the equivalent of intentionally giving Samardzija more innings than he should be pitching.

I know. Maybe he really has improved. Who knows? Stranger things have happened, but why not see if he can come anywhere close to replicating the second half of the season in relief before giving this man more innings? Odds are he won't come close to doing so and you can pass on the idea without looking idiotic. That's what the Cubs are likely to look like too. Below are the percentile forecasts for Samardzija using the average projection available on Fangraphs. Almost all the innings are as a reliever so you'd need to add about 1 run per 9 to get a reliever to starter conversion. Thanks a lot to SG from RLWY for helping me out with the percentile forecasts.

% G GS W L IP H R ER HR BB HBP SO RA ERA FIP WAR
80% 69 6 7 2 102 83 45 40 7 45 2 99 3.98 3.57 3.39 1.1
65% 66 6 6 3 97 85 47 42 9 47 3 90 4.31 3.89 3.85 0.4
Baseline 63 6 5 4 93 85 48 43 10 48 4 81 4.65 4.21 4.31 -0.1
35% 50 5 3 4 74 72 41 37 9 41 4 61 4.99 4.53 4.77 -0.5
20% 38 3 1 4 56 57 33 30 8 33 4 43 5.32 4.85 5.24 -0.7

Those are better than the average projection Berselius found by about .3 WAR though I'd argue if he's a -.4 reliever he wouldn't come close to pitching enough to be worth that little. That's a lot better than the percentile forecasts that PECOTA offers for him as a starting pitcher. Over 159 innings the baseline projection is -.4 WAR. As a starter it's easy to see someone being that bad. His 10% projection is -1.3 WAR and his 90% projection is only .7 WAR. BPro thinks he has to pass the 80th percentile just to be a replacement level starter.

His 90% projection ERA is only 4.28, which has an FRA (Fair Run Average) over 4.6.

I've been blogging about the Cubs since 2004. I've followed just about every decision they've made since then and I can safely say that of all the decisions this organization has made over those years, Samardzija to the rotation is probably the strangest decision yet. It's also quite likely to be the worst decision they've made. People say that maybe there is something that's changed. There damn well better be. If he's actually the same below replacement level relief pitcher and they just moved him into the rotation becasue of a small sample size and a sparkling ERA one has to wonder whether or not the Cubs took a step forward as an organization as we've assumed since the hiring of Theo Epstein.

Some have said, "what's the risk"? Well, for one thing, if he is the same pitcher and the smart money is on just that, the organization looks as clueless as ever. Second, you have to move some pitcher to the bullpen. If you insist on F7 being in the rotation Chris Volstad should be the guy sent to the pen. Or Iowa. Randy Wells average projection is better than Jeff Samardzija's 90th percentile forecast. Moving a pitcher to the bullpen only diminishes that player's value. What's another team going to be thinking if they look at Wells or Volstad? That he lost a job to Jeff fucking Samardzija? Yeah, good luck getting anything in return if you wanted to trade either one of them.

It should be pointed out that most of everything above was written yesterday with some additions this morning. That Samardzija had a bad outing that now makes the decision to move him into the rotation even worse was inevitable. The strongest argument in favor of Samardzija now is this: he ain't walked anyone. In 14 freaking innings. It should also be pointed out that pitchers have gone 14 innings without giving up a hit and nobody has ever used that as anything other than an example of how well he's pitched over those 14 innings. Not a single person has ever thought a pitcher like that was going to continue to allow no hits. For some reason though, people want to believe Samardzija has learned to control his pitches like David Wells did.

Quantcast