Saves are a stupid statistic...except...

...when you cannot convert them. Most thinking baseball writers abhor the save statistic. It is akin to touting batting average as the best way to tell who the good hitters are. And this post is not going to say those writers are wrong. Taken on a whole, the save statistic is not the best statistic. Probably WPA is a better way to evaluate relief pitchers. But the thing is, games are being lost in the last inning or two at a frantic pace this season. Take the Miami Marlins for example. They have five blown saves in seven opportunities. If you give the Marlins those five wins back, they are 12-6 and among the leaders of the NL East. Instead, they are 7-11 and in last place.

Here is the thing: Scoring has become at a premium in baseball. Runs are down again this season for the third season in a row. Teams are giving up an average of 3.78 runs per game. That is down from 3.91 runs per game a year ago which is down from 4.04 runs per game the year before that and 4.28 the year before that. The games are tighter. More games are coming down to the last couple of innings and what happens in those last couple of innings is determining who wins.

For all the grief he is getting, Bobby Valentine is correct. Many games are won and lost in the last couple of innings. As someone who picks games every day, the blown save is a killer and they are very noticeable. Saves are being converted at a rate of 65 percent this season. The converse of that is that there is a 35 percent fail rate. Last year, saves were converted at a 68 percent rate. In 2010, that rate was 69 percent. So, yes, teams are failing to convert the saves more often than in the past.

Among their many problems so far this season for the Angels, a glaring one is their bullpen, especially in tight games. They have converted only one of five save opportunities. That is killer. Other teams struggling at the back end of games are the Marlins, Giants, Cardinals, Dodgers, Royals and Astros. As good as the starts have been for the Cardinals and Dodgers, some of that momentum is slowed by converting close games at the end.

Some of the surprise teams this season are the Orioles, Nationals and Mariners. All three are closing the deal at the end of the game. The Tampa Bay Rays were thought to have a huge question mark at the back end of their bullpen. But Fernando Rodney has been amazing and they have an 86 percent success rate in closing the deal. That is a huge reason for why they are one of the top two teams in the AL East. The Orioles have ten saves already! The Mariners have only blown two saves thus far.

Another statistic to look at is inherited runners. When a manager goes to the bullpen with runners on base, those relievers have to somehow shut down the opposing team. Most of the time, it works. League-wide, only 28 percent of inherited runners are scoring this season. The Angels and Red Sox, two teams that have struggled out of the gate, lead the majors by allowing 45 percent and 43 percent of those runners to score respectively.

Saves is a statistic that is lightly regarded among analysts of the game and this post is not going to dispute that. What is glaring, however, is that teams that fail to convert them have a lot of wind taken out of their sails. With runs at a premium, when you have a small lead at the end of a game, you have to hold it. When you cannot, the results are devastating. Just ask Ozzie Guillen.

Holds and Saves

The Cardinals - Diamondbacks game on Sunday was a perfect example of how screwed up the Holds, Saves and Blown Saves are calculated. One pitcher, Chad Qualls, probably the games most unreliable closer at this point in time, actually blew a save opportunity, but got a Hold and the guy who took his place, Esmerling Vasquez, came in and saved the game for the Diamondbacks, but got a blown save. Here's how it happened...

Chris Carpenter started for the Cardinals and got a "quality start" (another discussion we need to have) by pitching six innings while giving up three runs. In what is really rare for Carpenter, he had trouble with his command and walked five batters which caused him to have to exit earlier than normal for him. His replacement, Dennys (does Denny have multiple personalities?) Reyes, gave up two more runs in a third of an inning to give the D-Backs a 5-2 lead heading into the top of the ninth.

Edwin Jackson started for the D-backs and held the Cardinals to only two runs, but he too walked on the wild side and threw 115 pitches in six and a third innings. That meant that the porous D-Backs' bullpen needed to get seven outs to win the game. That's a scary thought. Heilman came in and gave up three hits in an inning and a third, but didn't give up any runs (give that man a Hold!) and then closer, Qualls, took over in the ninth.

Lopez was the first batter and he doubled. Rasmus then hit a slow comebacker to Qualls and the pitcher threw badly to first for an error and made it first and third with no outs. Pujols singled and Lopez scored and the score was then 5-3 with sill nobody out. Matt Holliday just missed a homer and flew out to deep center. Rasmus tagged and went to third. Randy Winn--picked up by the Cards after the Yankees dumped him--singled to drive in Rasmus. Pujols went to third on the hit to again make it first and third with only one out. The score was then 5-4.

Here is where it gets messed up. A. J. Hinch had seen enough of Qualls and called for Esmerling Vazquez to get the Diamondbacks out of that mess. Vazquez then threw a wild pitch which allowed Pujols to score and the game was tied. Remember now that Qualls left the game while the Diamondbacks were still officially ahead in the game. So since the Diamondbacks never actually relinquished the lead while Qualls was in the game and eventually went on to win the game, Qualls blew the lead but was still credited with a Hold. All the runners that scored were guys he put on base. We won't even talk about the unearned run he received because of his own error.* Vazquez had one misguided pitch that went astray and allowed Qualls' last base runner to score. Qualls still got his Hold and Vazquez, who still at that point hadn't actually recorded a single at bat against, is lumped with a blown save. But Vazquez did his job and ended the threat by getting a ground out and a line out to end the inning.

*Posnanski asterisk ripoff: The Fan still thinks it's ridiculous that a pitcher is not given an "earned" run when the error was his. He certainly earned the run by making the error, right? It just doesn't make sense.

At least Vazquez got a win for his job well done as Chris Young hit a walk-off homer in the bottom of the ninth to win the game. But here's the thing. The casual fan and the general manager using the stats at contract time will look at this box score and see Esmerling Vazquez with a blown save and a win. That usually happens when a guy blows the lead but then his team rallies back to win the game. The implication here is that Vazquez didn't do his job and simply got the win because his teammates saved his butt. But he did save the game. It was Qualls that blew the lead, but again, the casual fan and Qualls' agent and contract time will point to this event as successful, because Qualls got the Hold, which means that he held the other team from getting the tying or winning run. But he didn't do that.

Put this in the same category as the reliever that gets a Save when he comes into a game with a three run lead, gives up two runs, but records three outs so his team wins. That wasn't a Save. That was a Whew! The Fan recommends that we modify the Save rules so that such events can't be considered Saves and create a new category called Whews!

And these scenarios are the exact reason why the stats of Saves, Blown Saves and Holds are discounted by sabermetric people everywhere. The Fan thinks those relief appearances that preserve wins should get positive statistics. It's just too bad the rules are so screwy that it takes away from what is a very tough job to do.

Closer Brought in for Games that are Tied

There is this "book" that everyone seems to talk about. This "book" has nuggets such as bunting for a sacrifice when at home but not on the road (or maybe it's the opposite). One of the newest items in the "book" is that you bring in a closer if it's the ninth inning and the score is tied and you are the home team. The same "book" says that you don't bring the closer in a ninth inning tie score situation if you are on the road. Like all items in the "book," these baseball "rules" are based on logic passed down from one manager to the next. But are they always logical and based on accumulated data? Probably not. Let's take a look at the closer pitching the ninth in a tie game entry.


The Fan has watched Mariano Rivera pitch for a long time. One thing this casual observer has always noticed is that he doesn't seem to be the same pitcher in non-save situations. Doing some research, the data proves that casual observation. In save situations, batters facing Rivera have an 88 OPS+ against him. Advantage Mo. But in non-save situations, he becomes rather ordinary and batters have a 106 OPS+. Advantage batter. To take it further. in save situations, Rivera has a 4.58 to 1 strikeout to walk ratio. In non-save situations, that figure is 3.44 to 1. Of course, this observation was exacerbated by Rivera's loss today in a tie game.

If you take Rivera by himself, then the "book" looks like it's messed up. Bringing Rivera in a tie game doesn't seem as effective. But before making that major pronouncement, we better look at more closers than just Rivera. The following lists the batters' OPS+ for several closers in save situations (the first number) and non-save situations (the second number).

Trever Hoffman: 92 - 115
Billy Wagner: 104 - 94
Francisco Rodriguez: 99 - 101
Joe Nathan: 72 - 88
Brad Lidge: 94 - 107
Bobby Jenks: 88 - 115
Troy Percival: 94 - 110
Jonathan Papelbon: 93 - 97
Lee Smith: 98 - 102

The Fan already knows some of this data is corrupt. For example, Rivera, Jenks and Papelbon each had a year of starting before they became relievers. Lidge, Percival and Wagner have always pitched in relief. But considering that most of these pitchers' statistics come from relieving, then that flushes out most of the offending data (at least in this small mind's opinion).

As easily seen from the numbers in the list, at times it pays to listen to that "book" and at times it does not. Papelbon, Nathan and Billy Wagner are good bets in non-save situations. These three come in under 100 in non-save situations meaning they still have the advantage over the hitters. Lee Smith seems like a wash either way as does K-Rod. But all the rest are iffy at best when brought into non-save situations like a tie game in the ninth inning, no matter if their teams are home or on the road. Hoffman, Jenks, Lidge and Percival all give the hitters an advantage in non-save situations.

So what should a manager do then? The manager should study the data and should save pitchers like Rivera, Jenks, Lidge, Hoffman and Percival for save situations. They may prepare differently or have different mental processes for save situations than non-save situations. It may be harder for them to be pumped up. The batters might be in different states of minds facing these closers when save situation occurs and when it doesn't.

The manager should also know that Papelbon, Nathan and Wagner are good bets in such situations. Since all these figures are readily available, then who pitches in tie games can be based on splits for such occasions and not on some ephemeral "book" passed down for generations.

More Blown Saves

The season-long trend of blown saves continued Thursday night as there were five in twelve games finished as of this writing (The Angels/Mariners and Giants/Padres games are still going as these words are typed). That means saves were blown in 42% of the games in the books for the night. Four of those teams lost games they were winning, meaning that 33% of the games featured a team that was winning and lost the game.

For the record, the blown saves occurred for:

• Kansas City - They lost because of the bullpen in a Greinke start and that lucky son-of-a-gun, Pavano, won again.
• Tampa Bay - But they still won because Oakland's bullpen was bad too.
• Pittsburgh - A blown save cost them the game.
• Florida - Another loss that should have been a win.
• Houston - Oswalt was a wizard stranding runners all night. His manager fixed that by bringing in a reliever to make sure those runs scored late in the game. Remember that Houston is faring less than 50% in save opportunities.

Some notes about the above:

In Tampa Bay, this is going to be a problem unless they can find an answer. The Fan has watched Wheeler pitch on several occasions and the guy has nothing. His fastball is straight as an arrow (just like Farnsworth's) and any success he's had has been a fluke. Meanwhile, the Percival experience has to end. The guy has the heart of a lion, but the arm of T-Rex. He gave up three more runs tonight.

Trey Hillman either has the worst luck in picking his relievers or he is just really bad at it. Greinke was sailing along and the Royals were winning, 3-2. Greinke had give up eight hits in his six innings, but he still struck out eight. Hillman brings in Horatio Ramirez, who entered the game with his ERA already sitting at 5.74. He had, to the point of this appearance, already given up 26 base runners in 15+ innings. Oh! But he's a lefty and it was a matchup thing. Well, it didn't work out. Two runs scored and the game ended up an 8-3 rout.

In the Pittsburgh game, Snell had given the Pirates a decent game and left after six innings with the score in the Pirates' favor, 4-3. The Pirates bring in Tom (ScoresAMany) Gorzalanny, who promptly gave the game away. The pitcher, who had a decent 2007, was bloody awful in 2008 and had only pitched 1.2 innings since he was called up from the minors. You're going to give a guy like that the ball in a big situation? Man.

The bullpen situation this year has been amazing. There are going to be a lot of starting pitchers who start 33 games only to have 20 decisions or less. It's been a bloodbath all over the league. One of the reasons that the Yankees and the Brewers have been on a roll is that their bullpens settled down. The entire league has to figure out how to do that same thing.

P. S. The Giants closed out the night with another blown save.

Blown Saves on the Rise

As mentioned on the daily game picks feature, it has become increasingly difficult to pick games based on the starting pitchers who are matching up. A big part of the equation has become which teams can close out a game once they have a lead. If you are feeling that blown saves are more prevalent than in the past, your feelings are dead on.

Let's look at the past five years:

2005 - American League Save Percentage: 68%, National League: 68%. There were only two teams in the entire major leagues that had a save percentage less than 60%

2006 - American League Save Percentage: 68%, National League: 64%. There were five teams that had a save percentage less than 60%.

2007 - American League Save Percentage: 68%, National League: 67%. Again, there were five teams that were successful less than 60% of the time.

2008 - American League Save Percentage: 67%, National League: 62%. The number of teams that were successful less than 60% of the time jumped to eight.

2009 - American League Save Percentage: 63%, National League: 61%. An amazing 14 teams are currently under 60%. And four of those teams have been successful less than 50% of the time: Washington (37%!!), Houston (41%), Cleveland (47%) and Minnesota (46%).

Some things to consider here: First, only one guy can get a save in a game. So only one guy can be successful. The other guys get "Holds." But any pitcher from the seventh inning on can get a blown save if he loses the lead (under certain score and men-on-base situations). So any team's Save Percentage is going to be lower on average than the closer's statistics because more guys can blow a save than can record one.

Secondly, the National League has seen a decline over the last two years while the American League has been steady up until this year when it has dropped by a wide margin.

The Fan is going to propose a couple of hypotheses on why this trend is occurring. While the Fan would love to take the time and compile numbers to support the hypotheses (and thus make it to Wednesay Wangdoodles), the Fan just doesn't have the will to work that hard. Perhaps someone will take these thoughts and either prove them or disprove them. But here we go.

First, increasingly, the "set up" men in the bullpen are becoming a luxury item. Three of the four teams that are performing at less than 50% this year in Save Percentage are in the bottom half of the team salary statistics. Since these teams have less money to either keep successful pitchers or buy them, they rely more on castoffs, long shots and youngsters to round out the bullpen and hope something works. When it is successful, it's an amazing thing. When it's not, the failure can be spectacular.

Secondly, starting pitchers are on an increasingly shorter leash. To see a pitcher go beyond 100 pitches now is rarer than ever before. With many teams placing more emphasis on OBP than ever before, those starters are getting to their 100 pitch ceiling faster than ever. That places more need on the bullpen meaning that more guys have to work to get the same job done. The Fan would bet that the average pitcher appearances per game is up for many teams. The Fan tried to find this data but was unsuccessful without going to each teams and counting them all up.

Lastly (unless you want to get into the whole PED debate which pitchers no doubt participated in as much as hitters), the current vogue in almost all major league ball parks is the left, right, left parade as managers play the match up game. It seems that every game, a late inning pitching change occurs where a weak throwing lefty (who may or may not be over 40) comes in to face a left handed batter and throws four sweeping breaking balls outside and walks the guy. His day is then done (unless there is another lefty coming up) and he has to hope that the next guy in to pitch doesn't surrender that run.

The Fan would like to back up his theories with cold, hard facts, but will have to settle for someone else doing that legwork. The one fact that is incontrovertible is that Save Percentages are taking a beating and 14 out of 30 teams can't even be successful 60% of the time. It's early yet in the season (a little past the 20% mark), but if the trend holds true, it's going to be a long year for fans hoping that a lead is safe once their team is ahead.