Projecting the Giants’ 2016 Rotation

The Shark could help put it all together in 2016 | Photo Credit: Ron Leuty

I spent some time earlier this offseason detailing my research into pitcher “Game Score” and its effectiveness in evaluating starting pitchers. You can find those earlier posts here and here. My purpose then was to compare the various free agent starting pitchers on the market, as well as to propose an idea for calculating a “true”  win-loss record for pitchers.

More recently, however, I’ve turned my attention to analyzing the effect Game Score (GSc) had on team performance. Is there a correlation between GSc and team wins, and could I use it to project team (more specifically, the Giants) W-L records going forward? Indeed, I feel like I can. It’s been a lot of work, but I’m excited to share my findings with all of you.

Quick primer if you didn’t read the posts from November. A Game Score is calculated for every starting pitcher in every game, yet rarely gets mentioned aside from comparing no-hitters and other dominant starts. A SABr study I found from a few years back proposed the idea of awarding pitchers wins and losses based on their GSc for each start, and showed the correlation between certain scores and win probability. A GSc of 55 or higher is considered a “Game Score Win,” and that’s important for the information and opinions I’m going to provide you.

So, how does all this relate to our Giants and their spending spree this offseason? I’ve made no secrets about my excitement for next year’s team, and specifically the rotation after the signings of Jeff Samardzija and Johnny Cueto. But there’s a crowd out there who’s a little weary heading into 2016, and wondering if Evans and Sabean really did bolster that staff enough. Fittingly, people around the web have been trying to answer that exact question.

Grant Brisbee wrote an article at MCC last week comparing the projected WAR improvement changing Hudson/Vogelsong/Pagan to Cueto/Samardzija/Span. According to Fangraphs WAR, the results were an 8-game improvement from last season. DrB followed up with a team-wide WAR projection at When the Giants Come to Town, showing the Giants at 84.5 total projected wins in 2016. DrB also compared the team WAR projections for LA and Arizona in his post. While Game Score (calculated solely from counting stats like IP, H, R, K and BB) certainly seems archaic compared to WAR calculations, I still think you’ll consider my findings from GSc very promising.

Back in November, I included the Game Score W-L records (and average GSc) for each member of the Giants rotation, and noted how poor those marks were in comparison to the majority of the arms on the FA market. Madison Bumgarner (and Peavy when healthy) is the obvious exception, and even Chris Heston’s 31 starts were league average. The rest of the staff was far below average. After putting in hours of new research and compiling the GSc totals for all 30 MLB clubs (as well as the last 10 years in Giants history), I can now give a little more perspective to those totals.

The Giants, as a team in 2015, earned 73 GSc wins (a score of 55 or higher). In other words, the starting pitcher’s performance made them very likely to win in only 45% of their games last season. That total was tied for 8th in the National League, and tied for 16th in the Majors. Bumgarner (22 GScW) and Heston (14 GScW) combined for 49% of those wins themselves. Considering the Giants had the 5th ranked run-scoring offense and the 3rd lowest bullpen ERA in the NL, it’s pretty clear that the rotation was the most glaring weakness on the team. Really, the offense and bullpen (and Bumgarner’s 22 wins) are probably the main reasons the team outperformed its GScW by 11, winning 84 team games.

While finishing 16th in the majors isn’t horrible by any means, you have to go back to 2007 t0 find a Giants rotation that posted less GSc Wins than that of the 2015 staff. That ’07 team, featuring young Matt Cain (20 GScW), rookie Tim Lincecum (12 GScW), and Barry Zito’s first season in orange and black, logged a combined 72 GScW. Its actual W-L record? 71-91. That team was in the middle of a rebuilding period… the 2015 team was defending a World Series title.

So, how do Game Score Wins correlate to team wins? Obviously, teams don’t win every game a pitcher throws a gem, and they don’t lose every game the starting pitcher bombs. There are plenty of bullpen games, extra-innings contests, and late-innings lead changes over the course of a 162-game season. But I was still surprised to find the connection between GScW and team wins quite high, especially in the pitching-heavy National League. In fact, the 5 teams who made the postseason out of the NL ranked #1-5 in GScW as well. Those rankings were as follows: #1 St Louis & LA (95 GScW); #3 NYM (91); #4 ChC & Pittsburgh (85). Washington had 85 GScW as well.

Of those postseason clubs, only the Mets and Dodgers underperformed their GScW total, winning 92 and 90 games as a team, respectively. Three other NL clubs won less games than their GScW would have indicated – Wash (-2); SD (-7); and ATL (-3). That means 10 teams, or 2/3 of the NL outperformed its GScW totals. The Rockies, not surprisingly considering Coors Field, led the way with a +20 difference (48 GScW, 68 team wins). The Giants, at +11, had the 4th highest differential in the league. When I combined the totals for all 15 clubs, I got exactly 70 more team wins than Game Score Wins for the season, an average difference of only 4.7 wins per team. That’s… a pretty sweet win correlation, if you ask me!

The AL Game Score data didn’t come out so squeaky clean, but that’s understandable for me considering the greater offensive output. The playoff clubs, by GScW, were ranked #4 (Houston); #7 (Toronto); #10 (Yankees); #11 (Kansas City); #14 (Texas). The WS champion Royals, literally the best team in the AL all season, outperformed their GScW by an insane 29 games. Hey Friedman, how’s that Dodger bullpen looking these days anyway? Still, despite the more volatile differences in GScW and team wins, the average differential for all AL clubs was still only 2 games higher on average than the NL (6.7 to 4.7). AL teams won exactly 100 more games than their starting pitching would have indicated.

For further proof of this correlation, here are the totals for the Giants during the past 10 years, with GScW listed first, followed by team wins and (differential).

2015 – 73; 84  (+11) | 2014 – 82; 88 (+6) | 2013 – 76; 76 (even) | 2012 – 84; 94 (+10) | 2011 – 94; 86 (-8) | 2010 – 90; 92 (+2) | 2009 – 88; 88 (even) | 2008 – 75; 72 (-3) | 2007 – 72; 71 (-1) | 2006 – 65; 76 (+11) | Total GScW = 799. Total Team W = 827.

If you didn’t tally the numbers in the parentheses, that’s a combined win differential of +28 over a 10-year period. Wait, let’s count that again. Take out the 2013 season (heck, leave it in if you prefer), and that’s a cumulative +27 dating back to 2012. If not for that 2011 club, which lost Buster Posey and wasted literally THE BEST Giants starting rotation of the past decade (22 GScW apiece for Timmy, Cain and Bum), the positive win totals go back even farther. The way their offense and bullpen have been over the past few seasons, you can almost bank on them outperforming their rotation’s GScW by 5-10 games.

What does all this mean? To me, it means the Giants don’t need a super rotation to be a playoff contender. Padres’ starters logged 81 GScW last season, exactly 50%. They were 8 games better than the Giants in that area, yet finished 10 games below Los Gigantes in the actual win column. If Bochy’s staff could hit that 81 GScW mark this season (an 8 game improvement), the bullpen and lineup would likely raise them into the 87-92 win range – very much in the thick of the postseason race.

So, just how much should we expect Cueto and Samardzija to help the cause in 2016? Well, I totaled the 2015 Game Score performances for the 6 guys who should see the majority of the Giants’ starts this season (Bum, Cueto, Shark, Peavy, Cain, Heston). The group made 157 combined starts, earning wins in 52% of them according to GSc… that’s good for 82 GScW, or the target mark I just suggested. Considering Cueto was an ace in Cincinnati, then well below average in KC, and Samardzija had easily the worst season of his career in 2015, it wouldn’t be a stretch to say this is the floor for a healthy 2016 Giants team.

You may question whether Cueto can stay healthy in the long-term, or whether Shark will regain his ace form, but do you honestly think both will repeat their 2015 second-halves next season in San Francisco? That’s the absolute most pessimistic view of the situation, and that’s not how I choose to look at it. I didn’t start a blog to be a downer.

If 87-92 wins is the floor for this team, then you probably want to know what the ceiling is… right? You only have to go back to 2014 to do that. Ready for the exciting part? Combine the 2014 numbers for Bum, Cueto, Shark, Peavy, and Cain (+1 start from Heston), and you get 95 GScW over 148 starts… or a .642 win%. Holy smokes! You put that group together in 2014, and you’ve got the best Giants staff in decades (and maybe ever). That staff, with this offense, is likely a an easy 100-game winning juggernaut.

Before we go too crazy though, it’s worth remembering that Cueto was one of the absolute elite pitchers in MLB in 2014, while Samardzija was having a career year himself. It’s pretty unlikely either repeat those numbers next season (65 avg GSc for Cueto; 59 for Shark). But it’s also worth remembering, however, that Bumgarner is still getting better. And as bad as Cain pitched in 2015, it’s hard to see him pitching that poor again if he’s healthy enough to make 20-25 starts.

I tried to take everything into account here (age, # of starts, 3-year avg’s) to realistically project the Giants 2016 rotation. Note that I used Fangraphs projections for games started for each pitcher. Here’s what I came up with:

Madison Bumgarner: 33 starts | GScW = 23

Johnny Cueto: 32 starts |  GScW = 20

Jeff Samardzija: 32 starts | GScW = 18

Jake Peavy: 26 starts | GScW = 13

Matt Cain: 23 starts | GScW = 10

Chris Heston: 8 starts | GScW = 3

All Other Starters (Clayton Blackburn?): 8 starts | GScW = 2

That’s a grand total of 89 Game Score Wins. Using that information, I feel very comfortable projecting this team at 93-97 wins, and an absolute threat to win the NL West. If the Giants win the World Series in 2016, it won’t be because of the calendar (ok, it might be…). It will be because they have the powerful, consistent rotation to match their young, exciting lineup and steady bullpen… and now the numbers back it up. Of course, a masterful manager and an unmatched team postseason experience won’t hurt the cause either.

I hope all of this helped to solidify your enthusiasm for the new season. I know it did mine. If you’re feeling giddy about the club right now, it’s because you should be! You heard it here: 2016 is lining up to be another very successful year for Giants baseball. As always, thanks for reading…. and go Giants!




11 thoughts on “Projecting the Giants’ 2016 Rotation”

    1. Hey thanks Moses! I’ve been working on my hitter projections again. Will post them before S.T. I’m not sure I’ll do pitcher projections… this is basically them. Pitcher ERA is incredibly hard to predict/forecast.

  1. Nice look at Game Score, enjoyed reading.

    I disagree with your statement regarding the Giants outperforming their Game Score wins by 11 games (73 vs. 84). You forgot that there were a bunch of no decisions limiting that Games Score win figure.

    If you add up all the starters, they were 73-54 over the season with 35 no decisions. If you treat each no-decision as 50-50 or .500, then we get 17.5 more wins making the win total 90 or 91, which means that they underperformed. And this number is actually close to what Pythagorean calculates that the Giants should win given their Runs Scored and Runs Allowed, which is 89 wins.

    While the starting pitching was not stellar last season, it was enough to win with, given some better luck. Just via Pythagorean, they should have been tied with the Dodgers for the NL West division title at 89 wins, they benefited and we fell short. We went 34-17 in Pence starts, showing again our pitching was good enough, given how good our offense was, when Pence was in there with the other regulars. Plus, we were not that far from the division lead until late August, when all the injuries costed our lineup tremendously: Pence, Aoki, Belt, Panik, Pagan, Crawford. Only Duffy and Posey were healthy enough to play most of the season.

    Using the projections from Steamer, Baseball Forecaster, and Bill James, I also get a projected win range from 93 to 103 wins (Bill James, for some reason, is always the outlier for offense, which ironically is what he feels most comfortable projecting). And just based on what they did last season and adjusting upward for the additions of Cueto, Samardzija, and Span, I can see the Giants winning somewhere in the mid-90’s.

    Of course, there will be injuries, but that’s not really something one can foresee. Nobody could have foreseen Pence being out for most of the season.

    Lastly, I would note that the 2011 team also had Vogelstrong with 16 GSwins, which would have been pretty sweet to have on the 2015 team.

    1. OGC, thanks for reading. At some point along the line I decided to take the “no decisions” out of the equation for individual pitcher evaluations. I definitely could have (and maybe should have) inserted them into the total W-L record here, but I felt the numbers matched up pretty well without those ND’s. You are right though, that the Giants did underperform their Pythagorean W/L some. So many things happen over the course of 162, it’s hard to pin it all on one thing. But they did have players like Tomlinson step into the lineup and help keep things moving offensively… I didn’t feel like the rotation had the same reinforcements this year.

  2. Heya Cove. Long time… I think you know I love delving into stats as much as many others like to. I certainly started paying attention to GSc’s a couple seasons ago. In fact, I used GSc in an early post last season. Forgive me, I forgot if I can embed links here?

    Sure, no stat seems perfect. GSc doesn’t account for batting performance for a guy like Bum or Greinke. Great offensive teams can obviously sway more comebacks away from no-decisions (or blown saves). Stout bullpens offer a field manager better opportunity to replace a fading ace in a game. I think we all get there are nuances. But Game Score ain’t too bad as some sort of “overall performance index” for starting pitchers. Not bad at all.

    As per usual, you offer great insight through strong analysis. Thanks for all the legwork you put into this. Nice job. Personally, I might feel 93-97 wins is “a bit” too optimistic. But hey, 89-92 wins seems within reason IMO. And hey, that aligns even CLOSER to your 89 GScW projection. 87-90 wins as a floor seems very fair, too. Rolling a lot of dice, but mostly I feel they are weighted towards “more upside.” Looks good from he-re. Happy New Year. Cheers.

    1. Carmot, great to hear from you. Thanks for checking in. I took a look at your post from April with the Game Score references. Somehow, I missed that last spring. I spent a little time (ok, a lot of time) catching up on your blog this evening, and left you a comment on your beautifully-written late-season recap vs. Cincy. Your writing is just a joy to read in general.

      I like that you included stat lines as a reference for different GSc values. That’s something I have missed on, and in the back of my mind knew I should have included at some point in all of my talk about the statistic… but my posts just get so long, I’m usually wiped out by the end of them. Things get left out. I know that I’ve been fascinated by GSc this offseason, and hope my work will help others take an interest in it as well.

      Thanks again for reading.

  3. Something to think about:

    2015 Giants Quality Starts

    22 Bumgarner (typical of him, btw)
    14 Heston (started strong, faded hard)
    10 Peavy (out of 19 starts)
    10 Hudson (out of 22 starts)
    09 Vogelsong (out of 22 starts)
    05 Leake (19 in total but only 5 for the Giants)
    05 LIncecum (out of 15 starts)
    03 Cain (out of 11 starts)
    00 Petite (out of 1 start)

    The managed just 78 quality starts out of last year’s rotation. Very close to the 73 GSC you came up with. And despite that rather… non-dominant… lackluster… average-at-best… club pitching performance the Giants still finished with a winning record. We added Cueto who had 19 quality starts and Samardzija who had 15. It was a bad year for both and they both underperformed there more recent history..

    The mere replacement of LIncecum, Hudson and Vogelsong (59 Starts, or about 2-starters worth of starts) with Cueto and Samardzija and you’re losing 24QS to gain 34QS. +10QS in rotation change. Put in a healthy Peavy (which is the problem), who is good for 17, maybe 18, QS when healthy (and I’m low-balling it a bit) and that’s becoming a pretty impressive rotation. (Peavy had 20QS (of 30) in 2014, 14QS (23 starts) in 2013. 23 (of 32) in 2012.)

    I think it’s reasonable to project a reasonable, and possibly low, 90 QS (assuming a 30+ starts for each starter) this year as follows:

    22 Bumgarner
    22 Cueto
    19 Samardzija
    17 Peavy
    10 Cain (Assuming he doesn’t get benched)

    So, coming at from a different data set, I think your GS calculations and your projected Giants win range are both very reasonable.

    Basically, with heath and upgrades, there’s no reason to not think a reasonably healthy Giants team can’t be a 90+ win team. They have a potent offense and they’re really addressed the major issue that kept them down last year — starting pitching.

  4. Since there must be a high correlation between GS, Quality Starts, and/or the DOM%/DIS% that ogc tracks, can you suggest the comparative advantages of these methods?

I'd love your feedback!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s