One of the biggest misconceptions about “Moneyball” is that people were led to believe that the Oakland Athletics were the first team to come up with the idea of focusing on undervalued stats like on-base percentage. The belief, especially if you saw the movie, was that the Athletics just flipped a switch and changed their approach one year, and that led them to become a winning team.
The truth is that the methods the Athletics used were being used for years — not only by Oakland, but by other teams in baseball. What made the “Moneyball” story was the fact that the 2002 Athletics lost three top free agents, yet used the Moneyball approach to win 103 games, despite having one of the lowest payrolls in the league.
Ever since Moneyball came out, fans of small market teams have been looking for their teams to take the same approach. Sometimes that involved directly copying Moneyball and going with some of the stats mentioned in the book. Sometimes a team had success in a new area, or developed a clear trend when adding players (for example, adding nothing but ground ball pitchers). The usual criticism against any “New Moneyball” is that a lot of the new approaches are very obvious and very mainstream now, due to the focus on advanced statistics.
There’s no doubt that more people are focused on stats today than in 2002. But that’s not to say that people didn’t focus on stats in 2002. It’s just that the internet wasn’t as prominent. It was harder to get your voice out there if you believed that OBP was better than batting average, and that slugging was better than a home run total. Now you’ve got blogs, Twitter, and all sorts of social media, which means that it’s easier to find people who think xFIP is better than ERA or that UZR is better than looking at errors.
The thing about Moneyball is that it was never about any specific stat. It was always about finding stats that were undervalued by the rest of the league, and using those stats to gain a competitive advantage. Even if teams know about the stats and appreciate that they have value, it’s still possible that teams are undervaluing some stats. And that’s where the Pittsburgh Pirates come in with the “New Moneyball”.
I was reading Dave Cameron’s article today about “How the Pirates Built a Playoff Team“. It’s a great read that touches on a lot of the big key points that led to the Pirates building this playoff team. I’d write about every single point, but in a lot of cases I would just be echoing Cameron’s thoughts without adding anything further to the discussion. Two things did stand out to me. The first was on xFIP.
The idea that a pitcher can be primarily evaluated on his walks, strikeouts, and ground balls is still a pretty unpopular opinion with a lot of people. It feels wrong to essentially ignore all the other parts of baseball, especially when you see a guy give up tons of hits and home runs and then hear some nerd with a spreadsheet explain that those don’t actually mean that he’s a bad pitcher. A lot of teams — most teams, I’d say — still prefer to evaluate pitchers based on some kind of runs allowed basis.
The Pirates, though, bought into the idea that they could build a pitching staff with fixable underachievers. Here are the ERAs and xFIPs of three of their recent pitching acquisitions in the season before they joined the organization.
The crazy thing about the xFIP debate is that it’s exactly the type of situation where, if you called it “Moneyball”, people would disagree. xFIP is used so much that it can’t possibly be Moneyball, right? Despite the fact that everyone knows about xFIP, it is still under-appreciated. As Cameron points out, there are so many people who reject the idea that xFIP is a better evaluator than ERA.
We saw this in Pittsburgh with all of the above moves. No one wanted A.J. Burnett, despite the fact that the only thing that looked bad about him was his ERA. People rejected the idea of adding Francisco Liriano, who had the same ERA/xFIP difference. And the anger over Mark Melancon was almost entirely fueled on his ERA. In any of these cases, if you argued based on xFIP, your argument was dismissed by those using traditional numbers. After all, xFIP is just a made up stat to either:
A.) Allow Bob Nutting to go the cheap route and avoid paying for good players.
B.) Allow stat nerds who have never played baseball to re-create the game because they weren’t here first.
But it’s not just media and fans that discount the value of xFIP. Teams obviously discount the value as well. Cameron points out in the article that the Pirates couldn’t have expected these results from Liriano or Melancon simply by looking at xFIP. There’s more at play here, and I think a big factor is Ray Searage. But you could use xFIP to tell that Melancon and Liriano would be much better pitchers than their 2012 ERA showed. Still, let’s look at the prices paid for the guarantee of a good ERA, compared to the hope that xFIP will win out in the end:
Zack Greinke – 6 years, $147 M
Anibal Sanchez – 5 years, $80 M
Edwin Jackson – 4 years, $52 M
Those were the top three starting pitching contracts signed over the off-season. Greinke is a great pitcher who not only had an ERA, but an xFIP to back it up. He had a 3.48 ERA last year, and a 3.22 xFIP. Anibal Sanchez was in the same situation. He had a 3.86 ERA and a 3.60 xFIP. Edwin Jackson had a 4.03 ERA and a 3.79 xFIP.
The common theme above is that all of those players had a good ERA. They also had a good xFIP to back up the ERA. Now look at the two starters the Pirates have acquired.
A.J. Burnett – 2 years, $20 M
Francisco Liriano – 2 years, $14 M*
*This was Liriano’s original deal before the arm injury and the restructuring of his contract. I’m going with the original value since I think that has a better reflection of his market value.
Burnett had a 5.15 ERA in his final year with the Yankees, but a 3.86 xFIP. Liriano had a 5.34 ERA and a 4.14 xFIP last year.
Burnett’s xFIP was in the same range as the ERA from Sanchez, and better than the ERA from Jackson. Yet Sanchez got $6 M a year more than what the Pirates are paying Burnett, and Jackson got $3 M a year more. The Pirates got Burnett because he was struggling, and they got him for a low price because the Yankees paid $13 M of his remaining salary. They didn’t even get any quality prospects. That’s all because Burnett’s ERA was bad. Yet if you believe that xFIP is a future predictor of ERA, then the Pirates paid $10 M for a future production that would have cost $3-6 M more per year on the open market.
Liriano didn’t have as great of an xFIP, but he also cost less. He was half of what Jackson and Sanchez cost, but the difference between a 4.14 xFIP and a 3.60-3.79 isn’t that much. In a 200 inning season that’s the difference between 80-84 runs, and 92 runs. At worst, Liriano’s xFIP would have produced 12 more runs in an entire season compared to Sanchez’s xFIP. Twelve runs isn’t nothing. Ten runs is the standard for one win. But a win on the free agent market is usually $5 M, and the Tigers paid $11 M more per year for Sanchez, compared to what the Pirates were prepared to give Liriano. $11 M for one extra win is a lot of money.
There’s a reason Greinke, Sanchez, and Jackson got so much. They were in demand. That’s the biggest thing here. The Pirates had no competition in signing Liriano or trading for Burnett. Part of the latter was that Burnett had a no trade clause, and didn’t want to go to the West Coast where he did have suitors. But there were teams on the East Coast who needed starting pitching in 2012, such as the Yankees.
You also have to consider that no starting pitcher is immune to having a bad ERA. A.J. Burnett is an example of this. He signed a huge deal with the Yankees in 2009, and after one year he had a bad ERA. His xFIP went up in 2010 (due to a lack of strikeouts), but was back to normal levels in 2011. Nothing else changed. He was still a strikeout pitcher with a reasonable walk rate and a HR/FB ratio that was way too high and would regress. He was the same pitcher that he was when the Yankees signed him a few years earlier.
We’re seeing that this year with Edwin Jackson. He has a 4.74 ERA, but a 3.81 xFIP. The xFIP is two points higher than his 2012 totals, but the ERA is his worst since 2007. And it’s not like Jackson hasn’t had this problem before. In 2010 he had a 4.47 ERA and a 3.71 xFIP. The next year his ERA was 3.79 and his xFIP was 3.73. Yet all it took was two years with a good ERA to get him a huge deal on the open market, and his value has probably dropped considerably due to this one “bad” season, even though he’s the same pitcher he was when he signed the contract.
It almost makes you want to never pay starting pitchers big money, at least not until teams around the league place the proper value on xFIP. Then there’s the relievers…
Rafael Soriano – 2 years, $28 M
Jonathan Broxton – 3 years, $21 M
Jeremy Affeldt – 3 years, $18 M
Soriano had a 2.26 ERA and a 3.75 xFIP. Broxton had a 2.48 ERA and a 3.62 xFIP. Affeldt had a 2.70 ERA and a 3.34 xFIP.
And then there was Mark Melancon, who had a 6.20 ERA and a 3.45 xFIP. Melancon was one of four players acquired for one year of Joel Hanrahan — a reliever who would make $7 M, was coming off a 2.72 ERA/4.28 xFIP split, and went on to miss almost the entire 2013 season with an elbow injury. Oh yeah, Melancon also makes the league minimum this year.
If you’ve ever read anything I’ve written on relievers, then it shouldn’t be a surprise when I say that I’m appalled at the prices paid each year for “top relievers”. Soriano, Broxton, and Affeldt all had great ERAs last year, but horrible xFIP numbers. And what happened this year?
Player – 2012 ERA / 2012 xFIP / 2013 ERA / 2013 xFIP
Jeremy Affeldt – 2.70 / 3.34 / 3.74 / 4.48
Jonathan Broxton – 2.48 / 3.62 / 4.11 / 4.50
Rafael Soriano – 2.26 / 3.75 / 3.15 / 4.01
Mark Melancon – 6.20 / 3.45 / 1.41 / 2.05
In every case with the top free agents, the player regressed the following year, closer to their previous year’s xFIP. It’s not just about looking at that one number. There are other factors to consider. Affeldt saw his ERA drop this year, but his xFIP also dropped due to a decline in strikeouts and an increase in walks, which have been hidden by a lucky .245 BABIP and a 75% strand rate. Melancon went way beyond his xFIP, but his 2013 xFIP is also lower, suggesting that it’s not just luck. That’s due to a decrease in walks, an increase in strikeouts, and even with lucky strand rates (79.9%) and HR/FB rates (3.3%) he is projected to put up great numbers if he pitches like this going forward.
While Affeldt is making $6 M a year, Broxton is making $7 M a year, and Soriano is making $14 M a year (or about the same as what the Pirates are paying Burnett, Liriano, and Melancon combined this year…for at most 70 innings!), Melancon is making the league minimum. And if he follows the Joel Hanrahan/Matt Capps/Heath Bell timeline for good relievers who don’t have a lot of saves but might be eventual closers, he’ll probably make $5 M over the next two years in arbitration. That means one year of any of those relievers is going to be more expensive than three years of Melancon.
Why do teams do this?
There was another “Moneyball” aspect that Cameron wrote about in regards to Russell Martin. He pointed out that Martin has value blocking and framing pitches. Those are two emerging stats for catchers that get publicity — just like xFIP — but aren’t taken seriously yet. To quote Cameron on Martin’s value:
Based on the DRS calculations, Martin has saved the Pirates nine runs through base stealing prevention, adding basically a win of value just by gunning down opposing runners so often. And that’s just one part of catcher defense. Martin has also been terrific at the others parts as well.
One of the stats we have here on the site but we don’t talk about much is RPP, which is runs saved or lost due to passed pitches, or said more commonly, blocking balls in the dirt. Based on the location of the pitches he’s received, the expectation is that Martin would have had 65 pitches get by him for wild pitches or passed balls; in reality, he’s only had 54 get by him, a difference of +11 pitches blocked, which translates into four runs saved. The #1 catcher in pitch blocking, Yadier Molina, is at +5.5 runs, so Martin isn’t far off the big league lead in this area either.
And then there’s pitch framing. Jeff Sullivan has written about Martin’s framing abilities on multiple occasions, noting even early in the season that Martin’s receiving skills could have a significant positive impact on his team’s pitchers. While pitch framing calculations are still in their early stages, Matthew Carruth’s catcher report at StatCorner estimates that Martin has saved the Pirates 18 runs through turning balls into strikes, the fourth highest total in baseball.
The bold parts are for my emphasis. It doesn’t take advanced statistics to know that 9+4+18 = 31 runs saved this season. If you go by ten runs per win, that means Martin’s defense alone has been worth three wins this year. FanGraphs has Martin as a 4.1 WAR player this year, and defense has been an overwhelming factor. Again, if you consider that the price for one win above replacement is $5 M, then Martin only needs to be a two win player to provide value on his two year, $17 M deal. Martin was a two win player in each of the last three seasons, and has been a two win player every year of his career, except 2009 when he had 1.5 wins. Still, 1.5 wins amounts to $7.5 M, which is close to the annual value of Martin’s contract.
The Pirates got Martin because the Yankees didn’t want to pay a few extra million per year. Martin’s value was down because of his .211 average last year. Even if you use advanced numbers like OPS (which is the gateway drug to advanced stats), his .713 OPS wasn’t horrible. But even if Martin did absolutely nothing at the plate, his defense alone would have provided enough value to cover his contract. The Yankees could have easily out-bid the Pirates for Martin, and still would have collected value on the deal. As it stands right now, Martin has provided over $20 M in value in year one of a two year contract that pays $17 M total. And that’s largely because of his defense, as his OPS of .713 is exactly the same as it was in 2012 when his offensive numbers were seen as disappointing.
Defense is extremely under-valued, not just by fans and media, but again by the teams. If it received the proper value, then there’s no way Russell Martin ever makes it to the Pirates in free agency. Clint Barmes has been horrible at the plate the last two years, and everyone thinks the $10.5 M for two years was a total waste. But he has provided $8.6 M in value, largely because of his defense, and that’s even with his offense taking away some of that defensive value. If the Pirates brought Barmes back next year in the same role that he’s played over the last two months, and gave him $2-3 M a year, they’d get value on that deal from his defense alone. But the general feeling would be that $2-3 M for Barmes is too much because he doesn’t hit, even though he’s been worth $4.2 M this year as an all-defense shortstop who has spent over half the year as a backup.
I’ve pointed out the stats, but it’s not entirely on the stats. It takes a change in approach and a commitment to a plan to execute some of these strategies. Travis Sawchik of the Trib had a great article on the 2013 Pirates and their defense. He talked about everything that went into the defensive shifts, from Clint Hurdle getting on board, to convincing the pitchers and fielders to accept the unconventional approach, and then getting pitchers to rely on two-seam fastballs at a career high rate to play into the infield defense shifts. The end result is that the Pirates rank fifth in defensive efficiency this year, and that has been a huge reason why they’ve been contenders.
All of this seems so simple. The Pirates aren’t doing anything that is unheard of. They’re just doing something that teams don’t normally do. They’re not the only team employing extreme defensive shifts. They rank fifth behind the Orioles, Rays, Brewers, and Cubs. That’s three small market teams and one big market team run by Theo Epstein. Just like Oakland adopting OBP and SLG in 2002, the Pirates are adopting defensive shifts this year to put together a winning season. Oakland wasn’t the only team relying on those stats in 2002, and the Pirates aren’t the only team relying on defensive shifts now.
The fact is, any team can do what the Pirates are doing. Other teams are paying a ton for offense, and the Pirates are winning with defense by paying guys like Clint Barmes and Russell Martin. Any team could sign those guys. The Pirates have their pitchers throwing two-seam fastballs, and their defense shifting per advanced scouting reports. The end result is that they’re getting great production from their pitchers, at a much lower cost than what other teams are paying for pitching on the open market. And any team could do the same.
That’s the true spirit of Moneyball. Taking unconventional approaches that work, and using them to beat the odds and win against teams who are taking the traditional and more expensive approaches. The A’s used this approach in 2002 to win 103 games with the second smallest payroll in the league. The Pirates don’t have that small of a payroll, but they have used the Moneyball approach to post their first winning season and make the playoffs for the first time since 1992. They might not get a book or a movie, but they’re getting what counts: wins.