Worley and Liriano ... A Fangraphs case study for why change is needed in Minnesota

[glunn]

This is one of the best threads I have seen on TD.

 

Thanks to jokin for starting this and to all who have participated!

[TheLeviathan]

it looks like the Giants implement the most effective shift with 494 shifts and 25 net runs saved. Fewer shifts than the Twins and significantly more net runs prevented. What makes the Giants so effective at shifts, and How can the Twins adapt philosophies to do both the Pirates go rate strategy and Giants shift strategy well? How to balance with effective hitting? So many questions...

 

Excellent questions.  I hope the Twins' staff in charge of such matters are asking and investigating these very questions.

 

I'd also like to commend everyone, very informative discussion here.  I wonder though if there isn't still synergy from top to bottom of the administrative/managerial staff.  The problem seems to be that the talent being brought up no longer fits the synergized philosophy we've seen the last ten years.

[Otto von Ballpark]

Just to add to the discussion about the infield shift, Inside Edge's Steve Moyer showed their data in the Wall Street Journal for all teams this year. The Twins shifted about 300 times fewer than the Pirates but the Twins' shifting saved them 15 hits while the Pirates' cost them 2.

Thanks for the data and link, Parker.

 

However, I am slightly confused by the data and some of the conclusions here; specifically, what happened in the non-shift plays seems very important.  If the Twins shifted ~300 fewer times than the Pirates, what if that cost them more than 17 hits?  Then their shift strategy is still a net loser compared to Pittsburgh.

 

Also, the timing of the shifts are ignored -- if the Pirates are more aggressive shifters, it suggests what hits they do prevent could be worth more than those of the Twins.

 

The data in the article is cool, but it is similar to comparing raw basestealing percentages -- a team with a lower percentage of success may easily be getting more value if they are taking more attempts and timing them better.  I think I'd have to see more data on this subject before I conclude (as some have here) that the Twins know more about effective shifting that they've let on.

[Otto von Ballpark]

it looks like the Giants implement the most effective shift with 494 shifts and 25 net runs saved. Fewer shifts than the Twins and significantly more net runs prevented. What makes the Giants so effective at shifts, and How can the Twins adapt philosophies to do both the Pirates go rate strategy and Giants shift strategy well? How to balance with effective hitting? So many questions...

First of all, it is hits saved, not runs saved.

 

Secondly, like my response to Parker, I want to caution about jumping to conclusions like this.  The Giants could have saved those hits by only shifting for the most obvious hitters and situations.  Makes their standing in this particular report look good, but it really says nothing about whether a more aggressive shift strategy would have been more effective for them overall.

 

We have the number of hits saved on shift plays; to really draw any conclusions, we would also need to know the number of hits teams allowed on non-shift plays, and how many of those could have been prevented by shifts.  Perhaps we'd also need adjustments for pitching staff / park / league.  And if the number of shifts are relatively low to the total number of plays, leverage/WPA effects would also be worth considering.

[Otto von Ballpark]

This analysis is curious. Rick Anderson is a huge 2 seam guy, to a fault in the past. He was routinely shoving it down the staff throats. I think that this 2 seam hard on the pirates have is successful because of their personnel. It's been great in the regular season and I'll be interested to watch how this "weak contact" works in the postseason. If anyone else remembers, this is almost exactly like what the twins had in the 2000's..... and we all know how well that staff did against good lineups.

On balance, I think the Twins pitchers performed as expected in the playoffs.  Everybody but Reed was good in 2002.  Silva scuffled in 2004, but Blackburn survived in 2009.  Santana was generally strong throughout.  Pavano and Liriano did OK with their opportunities too.  Some of the other failures were Duensing (twice!), Bonser, last career start Radke... not exactly guys you would expect to succeed anyway.

 

I always felt that our offense was the biggest letdown in those playoffs.  A few guys had decent series, but rarely at the same time, and there were some big underperformers at times.  There were also a few dregs in our starting lineups every series, sometimes due to injury, sometimes by design, but probably way too many for a consistent division winner.

 

Of course, by the end, it felt like there was some other factor consistently holding the Twins back as well, not necessarily from winning a series but just from a occasionally winning a game and competing on a series level.

[drjim]

First of all, it is hits saved, not runs saved.

 

Secondly, like my response to Parker, I want to caution about jumping to conclusions like this. The Giants could have saved those hits by only shifting for the most obvious hitters and situations. Makes their standing in this particular report look good, but it really says nothing about whether a more aggressive shift strategy would have been more effective for them overall.

 

We have the number of hits saved on shift plays; to really draw any conclusions, we would also need to know the number of hits teams allowed on non-shift plays, and how many of those could have been prevented by shifts. Perhaps we'd also need adjustments for pitching staff / park / league. And if the number of shifts are relatively low to the total number of plays, leverage/WPA effects would also be worth considering.

Doesn't "hits saved" imply analysis of results of a shift vs. non-shift situation.

 

It is quite possible that teams shift too much and negate the advantage. One season of data certainly won't provide the answer either way.

[Richie the Rally Goat]

First of all, it is hits saved, not runs saved.

 

Secondly, like my response to Parker, I want to caution about jumping to conclusions like this.  The Giants could have saved those hits by only shifting for the most obvious hitters and situations.  Makes their standing in this particular report look good, but it really says nothing about whether a more aggressive shift strategy would have been more effective for them overall.

 

We have the number of hits saved on shift plays; to really draw any conclusions, we would also need to know the number of hits teams allowed on non-shift plays, and how many of those could have been prevented by shifts.  Perhaps we'd also need adjustments for pitching staff / park / league.  And if the number of shifts are relatively low to the total number of plays, leverage/WPA effects would also be worth considering.

my mistake, it is hits, not runs.

 

Don't confuse effective and efficient. In this case the Astros are very efficient, taking advantage of economies of scale. For shifting, effectiveness is what to strive for. A team doesn't want to shift too much because it leaves them vulnerable for a player to "beat the shift" the more a team shifts, the more opportunities to beat the shift, thereby reducing it's effectiveness. Ideally a player would provide subtle fielding position variations to put them in the right fielding position without the coaches/managers calling it. Think more soccer than baseball. Inherently, more is less.

 

Now, I agree 100% not to jump to conclusions with this data. It causes more questions than answers, but progress starts with a question, not an answer.

[Willihammer]

The whole article is a good read on the subject.

The link in that article is another good read. I'm wondering how you would reconcile it with the article.

http://online.wsj.com/news/articles/SB10001424052702304679404579459233064277204

 

According to the link(referring to the 2013 season):

 

BIS measures this using its "defensive runs saved" calculation. This measures how many runs were prevented from scoring (or led to scoring) by individual fielders and team defense. The average MLB team has a runs-saved figure of zero. The Pirates defense "saved" 77 runs in all, or 77 runs better than an average defense, third-most in Major League Baseball.

 

Are the Pirates' shifts cutting out extra base hits while allowing a couple extra singles, thus netting a positive gain in runs saved?

 

Its interesting that the Pirates have had for two seasons running, the largest DRS-UZR spread in baseball. This suggests to me that the Pirates are using shifts to compensate for the poor defensive skillset of their infielders (with great success).

 

 

UZR handles "shifts" as an on-off switch; the player is either shifted or he isn't. If UZR deems that a shift is on, the play doesn't count for UZR's tabulations. If they determine that a shift is not on, the player will be scored as though he were standing in the middle of his position's defensive zone, regardless of where he's actually standing.

 

 

Whereas DRS takes no account of whether a shift is on or not:

 

the numbers determines (using film study and computer comparisons) how many more or fewer successful plays a defensive player will make than league average. For instance, if a shortstop makes a play that only 24% of shortstops make, he will get .76 of a point (1 full point minus .24). If a shortstop BLOWS a play that 82% of shortstops make, then you subtract .82 of a point. And at the end, you add it all up and get a plus/minus.” (Joe Posnanski, Sports Illustrated)

 

 

In 2013 the Pirates' DRS-UZR spread was the greatest in baseball at +63.6. The second most was the Brewers at +58 (Twins were +29.6). In 2014 the Pirates' DRS-UZR is again the most in baseball at +72.2 (second is the Astros at +45.3. Twins are -40.3).

[Willihammer]

I should also mention that Worley has a .266 batting average on ground balls in play which is well above the league's average of .245, so it is not like the shifting has been extremely beneficial to him. Where he's really excelled is in line drive defense -- his .600 batting average on liners in play is much better than the league's average (.648). 

Line drives would more often go for extra bases than grounders, no?

 

How does the team overall compare to the league average in converting line drives to outs? Grounders? Fly balls?

[jokin]

The link in that article is another good read. I'm wondering how you would reconcile it with the article.

http://online.wsj.com/news/articles/SB10001424052702304679404579459233064277204

 

According to the link(referring to the 2013 season):

 

BIS measures this using its "defensive runs saved" calculation. This measures how many runs were prevented from scoring (or led to scoring) by individual fielders and team defense. The average MLB team has a runs-saved figure of zero. The Pirates defense "saved" 77 runs in all, or 77 runs better than an average defense, third-most in Major League Baseball.

 

Are the Pirates' shifts cutting out extra base hits while allowing a couple extra singles, thus netting a positive gain in runs saved?

 

Its interesting that the Pirates have had for two seasons running, the largest DRS-UZR spread in baseball.

 

 

This suggests to me that the Pirates are using shifts to compensate for the poor defensive skillset of their infielders (with great success).

 

 

UZR handles "shifts" as an on-off switch; the player is either shifted or he isn't. If UZR deems that a shift is on, the play doesn't count for UZR's tabulations. If they determine that a shift is not on, the player will be scored as though he were standing in the middle of his position's defensive zone, regardless of where he's actually standing.

 

 

Whereas DRS takes no account of whether a shift is on or not:

 

the numbers determines (using film study and computer comparisons) how many more or fewer successful plays a defensive player will make than league average. For instance, if a shortstop makes a play that only 24% of shortstops make, he will get .76 of a point (1 full point minus .24). If a shortstop BLOWS a play that 82% of shortstops make, then you subtract .82 of a point. And at the end, you add it all up and get a plus/minus.” (Joe Posnanski, Sports Illustrated)

 

 

In 2013 the Pirates' DRS-UZR spread was the greatest in baseball at +63.6. The second most was the Brewers at +58 (Twins were +29.6). In 2014 the Pirates' DRS-UZR is again the most in baseball at +72.2 (second is the Astros at +45.3. Twins are -40.3).

 

Thanks, Hammer.  This really puts the science of the shift into better perspective. It's becoming apparent that the Pirates have put a great deal of thought and man-hours into their overall GB-heavy-inducement strategy, relative to run prevention vis-a-vis payroll dollars committed to pitchers and infielders, as well as to game-planning of opponent hit f/x.  And it seems replicable with the Twins existing personnel, it seems that the Pirates IF, defensively, are only better than the Twins at SS.

[The Wise One]

First of all, it is hits saved, not runs saved.

 

Secondly, like my response to Parker, I want to caution about jumping to conclusions like this.  The Giants could have saved those hits by only shifting for the most obvious hitters and situations.  Makes their standing in this particular report look good, but it really says nothing about whether a more aggressive shift strategy would have been more effective for them overall.

 

We have the number of hits saved on shift plays; to really draw any conclusions, we would also need to know the number of hits teams allowed on non-shift plays, and how many of those could have been prevented by shifts.  Perhaps we'd also need adjustments for pitching staff / park / league.  And if the number of shifts are relatively low to the total number of plays, leverage/WPA effects would also be worth considering.

The shift is employed when the batter's spray chart shows a tendancy and usually no base runner. So to get the "you coulda shifted there" number you would have to have an understanding how predominante of a hit pattern  you are willing to shift for. It would be a safe guess given the Twins have given up  about 150 more baserunners that the Pirates have there is less chance for shifting by the Twins. To really break it down, you would have to look up when during an inning a hit/walk  was given up More baserunners on means less shifting..What other factors wouls limit shifting. If your pitcher does not have good control and leaves a lot of pitchs where a hitter could go the other way (or over the fence) versus pulling the ball, it might be a good idea not to shift.

 I think the Twins have a preference towards not shifting when the pull hitting tendancies are less prominent. The numbers might say it is a good idea.

[Jham]

Pitchers also pitch differently when the shift is on, which the stats don't show.  This may effect the outcomes as well.  Where a pitcher may typically choose to attack David Ortiz with breaking balls away, the same pitcher may elect to run fastballs inside with a shift behind him.  A tactic which encourages a pitcher to pipe fastballs into a power-hitter pull-wheelhouse is a gamble to say the least.  For instance, I was at a Cleveland series this year where we were shifting on Carlos Santana.  He dropped down a bunt single in game 1, a seeming win over the shift.  However, the Twins also thought they won by turning Cleveland's clean up hitter into a bunter.  In game 2, we continued the shift and Santana turned on one and crushed it over flag pole.  There's no shift for that!

It would be nice to see a stat comparing singles against the shift vs. extra base hits and HR in order to determine its overall effectiveness.

[Kirby_waved_at_me]

Yes, that's a good point - the shift is a self-defeating strategy if it forces (or even just encourages) the pitcher to throw pitches to a hitter's wheelhouse.

In a similar vein, if a shift can change a hitter's approach at the plate, it's worth exploring further. Some batters don't really change their approach when the shift is employed, and it seems to really mess with the head of other guys. 

[jokin]

Yes, that's a good point - the shift is a self-defeating strategy if it forces (or even just encourages) the pitcher to throw pitches to a hitter's wheelhouse.

In a similar vein, if a shift can change a hitter's approach at the plate, it's worth exploring further. Some batters don't really change their approach when the shift is employed, and it seems to really mess with the head of other guys. 

 

I'm not following your logic, here.  If a team shifts, it's with the expectation that the batted ball has a high probability of ending up in the teeth of the shift.  I would think the pitcher's game plan is to emphasize at delivering the location and type of pitch that most likely results in that outcome for that particular batter based on spray chart/hit f/x.

[jay]

I'm disappointed I missed this conversation until now, it's right up my alley.

 

It appears the Pirates have been able to synergize many of their approaches.  What stands out as strange to me is the fact that PNC Park has been one of the hardest ballparks to hit HRs in since opening in 2001.  That would lead me to believe that a FB-heavy staff could be especially successful there, but they've gone the opposite direction.  Perhaps they are choosing to focus run prevention elsewhere knowing the park will naturally limit HRs already?

 

The Twins don't need to directly mimick the Pirates.  However, they do need to have an approach that is coherent and aligned across the organization to whatever corner of the market they think provides the best value.  That approach needs spread to their hitters, pitchers and defense both on the current roster and in player development.  I don't think we've seen that.

 

Park factors (analyzed to handedness, batted ball profile and section of the field) would seem to me to be the biggest constant that a team could build around.

[Kirby_waved_at_me]

I'm not following your logic, here.  If a team shifts, it's with the expectation that the batted ball has a high probability of ending up in the teeth of the shift.  I would think the pitcher's game plan is to emphasize at delivering the location and type of pitch that most likely results in that outcome for that particular batter based on spray chart/hit f/x.

 

Yes, the shift is using the spray chart to increase the likelihood that the hitter will hit into the shift. If it's a power hitter who feasts on pitches that are on the inner half, and the pitcher is throwing to the inner half to try to induce the hitter to ground the ball into the shift, but instead is giving up 400 foot bombs to RF, maybe it's better to not shift and have the strategy for pitching that guy be based on scouting hot and cold zones.

 

More importantly, if a pitcher is taking a different approach with the shift behind him but he doesn't have the best control on the hill, I would argue that the pitcher is more likely to make a mistake pitching to the shift than he would be pitching just to the batter's weaknesses.

 

If you're using the shift to emphasize your team's strengths and minimize their weaknesses, it's a great move.

[jokin]

Yes, the shift is using the spray chart to increase the likelihood that the hitter will hit into the shift. If it's a power hitter who feasts on pitches that are on the inner half, and the pitcher is throwing to the inner half to try to induce the hitter to ground the ball into the shift, but instead is giving up 400 foot bombs to RF, maybe it's better to not shift and have the strategy for pitching that guy be based on scouting hot and cold zones.

 

More importantly, if a pitcher is taking a different approach with the shift behind him but he doesn't have the best control on the hill, I would argue that the pitcher is more likely to make a mistake pitching to the shift than he would be pitching just to the batter's weaknesses.

 

If you're using the shift to emphasize your team's strengths and minimize their weaknesses, it's a great move.

 

I think it works well on an individual basis, we have one of the best examples on the Twins of teams scouting both hot zones and overwhelming tendencies, regardless of a pitcher's control.  The shift on Joe Mauer is now near-ubiquitous around the league.  Night-after-night, DicknBert marvel when Mauer is up to bat.... at how close the LF is to the LF foul line.... and how far away the RF is to the RF foul line.  I'd like to see the study for verification, but the spray charts I've seen indicate that Mauer;s production suffered, largely due to  the shift this year.  

[Mike Sixel]

I think the key message is that the FO and field staff are aligned in ONE CLEAR set of strategies, and are implement it. That, to me, is a huge deficit here. If there is a strategy, it isn't clear, and it doesn't appear to be implemented consistently from top to bottom (see Ryan's statements on Santana as a SS, and him playing CF). that said, I'm not in Pittsburgh, maybe they appear disconnected there also.

[Kirby_waved_at_me]

I think it works well on an individual basis, we have one of the best examples on the Twins of teams scouting both hot zones and overwhelming tendencies, regardless of a pitcher's control.  The shift on Joe Mauer is now near-ubiquitous around the league.  Night-after-night, DicknBert marvel when Mauer is up to bat.... at how close the LF is to the LF foul line.... and how far away the RF is to the RF foul line.  I'd like to see the study for verification, but the spray charts I've seen indicate that Mauer;s production suffered, largely due to  the shift this year.  

 

With Mauer in particular I can cite several examples of him absolutely destroying the baseball for an easy out  - but over the course of the whole season, it may be my own confirmation bias mixed with my hope that Mauer would get better as soon as those hits started to fall in. Is it the shift, or is he not hitting the ball hard as often as he did in the past? The Shift has no doubt played a role in his diminished stats.

 

Since this is the first year the Twins have played the defensive shell game so heavily, I'm hoping that they take away some lessons and refine their approach next year. The numbers show that they were able to take away some hits, and hopefully that will translate to even more efficient shifting and even some improved run prevention in 2015.

[Badsmerf]

I'm bumping this because of the helplessness the pirates are displaying again in the playoffs. This pitch to contact theory does not work in this context. I would have to do more research on it, but when every pitch is magnified this ideology fails. One game isn't enough to put together knee jerk reactions, so more data would be necessary.

[Mike Sixel]

I think the point isn't "pitch to contact is a good idea", I think the points are:

 

1. Strikeout pitchers are really expensive. If you are a lower revenue team, you need to find another way, maybe.

 

2. If you go with another strategy, you need to change your fielders to fit that strategy.

 

That's what the Pirates have done, at least that's what they claim to have done. Isn't that a good idea?

[Kirby_waved_at_me]

Volquez was doing ok for most of his outing, and got knocked around towards the end. Giving up the Grand Slam pretty much sealed the deal for the Giants and sealed the fate of the Pirates. If Volquez had induced a ground ball against Crawford, it's likely a much different story.

[jokin]

I think the point isn't "pitch to contact is a good idea", I think the points are:

 

1. Strikeout pitchers are really expensive. If you are a lower revenue team, you need to find another way, maybe.

 

2. If you go with another strategy, you need to change your fielders to fit that strategy.

 

That's what the Pirates have done, at least that's what they claim to have done. Isn't that a good idea?

 

Yep.  Here're the Twins K/9 rank since 2008:

 

08-  26th

09- 23rd

10- 26th

11- 30th

12- 30th

13- 30th

14- 30th

 

The Twins had bad K-rates from 08-10, but their defense was for the most part good enough to make up for the K deficiency to produce three straight winning seasons.  The Twins have the worst K-rates from 11-14 and the defense can't come close to making up the K deficiency. (Interesting to note, with the SP upgrade this year, the K/9 rate of 6.47 in 2014 is actually higher than the average K/9 of 6.38 for the years 2008-10... which obviously shows the K-rates around the league continue to rise while the Twins struggle just to stay above water).

[Kirby_waved_at_me]

I think "above water" is being generous

[jokin]

I think "above water" is being generous

 

"Struggle not to drown" better?  The main point remains, there was no "plan" implemented at the major league level to address the lack of a run prevention strategy,  other small market teams have done so and thrived.  Pitt did it their way, Miami did it differently, but has gone from #1 NL laughingstock just a year ago during the 2013 season, to having one of the most respected pitching staffs in the game.  

 

The Twins "strategy" appears to sign speculative SP FAs along with 2 trades to keep the leaky boat afloat in the short-intermediate term, and in the meantime, draft bushels of K-throwing college relievers and sign 4 HS arms and hope it all pans out to big success by 2017-18.  

[Kirby_waved_at_me]

I guess in 2010 the Twins expected to still have guys like Scott Baker, Kevin Slowey, Liriano and even Nick Blackburn in the rotation in 2014.

 

That strategy has been torpedoed by injuries and regression. Unfortunately the re-inforcements from the minors didn't materialize in 2011-2014. May and Meyer might both contribute in 2015, but it's been a long road back and the Twins aren't particularly close yet.

 

They have been using band aids to try to heal a pretty nasty flesh wound. They've spent like a small market club along the way, the veteran arms haven't provided much more than age, and with the exception of Hughes, the new faces have disappointed.

 

I feel like the rotation will be very good by the time Berrios makes it to the show, but it's been a very long wait for relevance.

[Otto von Ballpark]

I guess in 2010 the Twins expected to still have guys like Scott Baker, Kevin Slowey, Liriano and even Nick Blackburn in the rotation in 2014.

 

That strategy has been torpedoed by injuries and regression. Unfortunately the re-inforcements from the minors didn't materialize in 2011-2014.

If that was their expectation, it was a terrible one.  2010-2014 is 5 seasons.  Radke is about the only Twin to top that in recent memory.  Milton I guessed made it exactly 5 seasons before he was injured/traded.  Even if healthy and successful, contract issues make it fairly unlikely that a guy will last that long.  Santana wasn't even a full-time Twins starter for 5 years.

 

I think more likely, they just over-estimated their depth when they dealt Garza and Santana, and over-estimated their SP prospects/drafts at the time (Gibson, Hendriks... Humber?).

[markos]

I'm bumping this because of the helplessness the pirates are displaying again in the playoffs. This pitch to contact theory does not work in this context. I would have to do more research on it, but when every pitch is magnified this ideology fails. One game isn't enough to put together knee jerk reactions, so more data would be necessary.

If there is a lesson to be learned from last night's game, it is that it sucks to not have your best starter lined up for the wildcard game. Additionally, it also sucks to waste your best starter on the last game of the year in a failed attempt to try to force a tie in the division.

[Mike Sixel]

Think about that....they'd need the third best starter to play in the tie breaker, which they probably lose, then they have to use their 4th best starter in the elimination game.....think about that.

 

Bad, bad strategy in game 162

[Otto von Ballpark]

Think about that....they'd need the third best starter to play in the tie breaker, which they probably lose, then they have to use their 4th best starter in the elimination game.....think about that.

 

Bad, bad strategy in game 162

To be fair, had they been able to force a tiebreaker game, they would have only had to face lesser Cardinals starters Lackey/Miller (albeit in St Louis).  It's hard to blame them for trying to go that route and avoid an elimination game against Bumgarner.

 

I suspect another factor here is that they over-rated Volquez and his recent good run (mostly against inferior competition).