Exorcising the ghosts of 2007…

One of my biggest memories of 2007 is the feedback I got regarding the Jays not using the sacrifice bunt with men on (or a runner in scoring position) when there was nobody out in certain situations (to be dealt with shortly). The thing is, a lot of it came due to folks thinking that I was advocating the Jays become the 1985 Cardinals or the “Hitless Wonder” White Sox of the early 20th century.

I can see why fans who tried to picture Frank Thomas, Troy Glaus and Gregg Zaun playing “HBP + steal + bunt over to third + ground ball to the hole = run!” might be inclined to think I hadn’t read the instructions to a K-Tel home lobotomy kit. Hey, it was only $2.99 and included a set of Ginsu knives (endorsed by O.J. Simpson, no less), a “Goofy Greats” LP and a Fishin’ Magician/Bass-O-Matic kit (great for shore lunches).

Probably the main reason for the flood of critiques was that folks who read my first two MSN Canada columns dealing with it had not been been reading me all year on THT and hadn’t seen the development of my thought process (such as it is) on last year’s edition of the Blue Birds. Some of them thought I did it in hindsight—bottom line, there was a lot of misunderstanding about where I was coming from. I did an ’07 review of the Jays on my blog as a reference point to how the thought process evolved.

If you’re so inclined, you can read Misty water coloured memories of the way we blew… Initially back in May I wrote here:

On to more current matters: I know J.P. Ricciardi is loath to give up outs for runs or to sacrifice. I hope John Gibbons realizes he has to be pragmatic. Right now, due to injuries, the Jays may have as many as three major out-producers at the bottom of lineup. If Troy Glaus’ legs are being rested it’s entirely possible that your 7-8-9 hitters could be some combination of Jason Phillips (78 OPS+), Sal Fasano (27 OPS+), Royce Clayton (78 OPS+), Jason Smith (44 OPS+) and John McDonald (67 OPS+).

Thus far this season (as of this writing), this quintet has drawn just 19 BB in 322 AB and struck out almost as often as they get a hit (77 K/ 78 hits).

Occasionally there will be runners at first and second with none out and one of these five players coming to the plate. Suffice it to say, the odds of a popup, strikeout or double play ball are much more likely than that of a hit or walk.

What Gibbons should do in this scenario is try to bunt the runners over. Every day, Gibbons should be giving these guys bunting practice. When the odds of an out (and possibly two) are so strong, a double play more probability than possibility, the Jays would be better served in making sure that the almost inevitable out (the aforementioned five players have an aggregate OBP of .292) at least moves runner along. That way a ground ball or ball hit to the outfield at least gets you a run.

My articles on MSN were an outgrowth of what I began to observe by the season’s second month and wrote about in this space.

I’m the first to admit that analysis is not a strong point and that the numbers are only one data point (among many) I consider when breaking things down—I’ll look at both objective and subjective information before drawing conclusions. Looking back, I’m quite unrepentant about advocating this particular approach.

I don’t think it takes strong analytical skills to discern when something just isn’t working—all that is needed is an ability to read scoreboards and standings. Probably what made me shake my head (and not feel so bad about my own mediocre mad analysis skills) were the e-mails I received that said, in effect, the Jays should continue with their philosophy until it does work—because it will … eventually.

Of course, you can’t afford to wait until eventuality comes—there is a finite period of time and it’s called “the regular season.” The thing is, those whose devotion to numbers is nothing short of remarkable never batted an eye at the numbers of the Jays’ backups both in 2007 and in their careers.

The funny thing I learned is that I’m not alone in my paradoxical quirkiness. Most of whom I would consider hardcore sabermetricians are quite liberal and open-minded in their worldview. They strongly dislike dogmatism. However, when it comes to baseball, any apostasy, however slight, from the sabermetric orthodoxy requires a cyber-jihad that involves explosive messages in my inbox.

More than once, I was tempted to put my own spin on a common putdown used against those into sabermetrics and write back “get your head out of following your team’s season and watch a bloody Blue Jays game once in a while.”

But I digress. (The first one of 2008!)

Anyway, the various out-machines employed by the Jays ended the season playing worse than I had noted in May. By season’s end, the Jays gave more than 1,000 at-bats to players with nothing in their past to indicate they could contribute. The Jays’ “backup crew” batted an aggregate .229/.280/.295 with six home runs.

How an Ace Performance Impacts Reliever Workloads
Bullpenning has its advantages, but it's great when an elite starter eats up a bunch of innings, too.

(I’m not including Adam Lind’s contributions since he’s a piece of the future and if you’re going to play currently unproductive talent, it’s best to play those who will benefit from the experience and become productive.)

When a team is struggling to score runs and there are men on/nobody out with two or three below-replacement-level batters due up, it’s best to expect them to perform according to career norms and plan strategy accordingly. A team should be willing to manufacture runs when the numbers strongly suggest the talent on hand is unable to produce runs on anything resembling a consistent basis.

I’m still struggling with why it’s accepted wisdom that a leadoff hitter should have on-base ability and a cleanup batter should be a masher and there is an expectation that they will reach base and hit for extra bases. Their skill set determines their function in the lineup…


… when you’ve got outmakers in the batting order, why is it unreasonable to expect that they’ll make outs (and quite possibly two)? If that is the case, isn’t it logical to accept their dominant “skill” (making outs) and try to improve a scoring opportunity rather than waiting for guys batting .229/.280/.295 to hit/walk/hit for extra bases? Why is it illogical to assume players who cannot crack an OPS of .600 are incapable of run production and to make decisions based on that incapability?

Admittedly, I’m not as savvy with stats as many, but I’m pretty sure .229/.280/.295 indicates a level of ability not known for hitting/walking/hitting for extra bases.

Another difference of opinion I have with aspects of sabermetrics is something that strikes me as a glaring inconsistency. Most realize that conventional stats are poor indicators of offensive value—especially runs and RBI. To use a simple scenario: Batter A reaches on an error, Batter B doubles him to third (Batter A is a bit of a sloth I guess), Batter C hits a long fly ball to score Batter A. The man reaching on an error is credited with a run. The man who made an out is credited with an RBI. The batter who had the key at-bat in the sequence gets bupkis. Sabermetrics is about separating the performers from the poseurs—correctly identifying the value of baseball’s “Player Bs.”

The bottom line is, a team needs its batters not to just score runs and drive them in, but to get baserunners around the … well, bases.

Here is what I find a bit inconsistent. Men on first and second, nobody out and John McDonald is due up. Sabermetric philosophy states it’s bad to give up outs. McDonald swings away and pops up on the infield and it’s men on first and second, one out. However, at least the Jays didn’t give away the out.

Now, same scenario and McDonald bunts and is thrown out at first and the Jays have given away an out—that’s bad. However, instead of first and second with one out, it’s second and third with one out. Like the batter with the key hit in the above paragraph, McDonald has given his team two bases and now two men are in scoring position.

So, a sacrifice bunt from an outmaker (.279 career OBP) that gets two bases is giving away an out while a popup on the infield that doesn’t improve a scoring chance is not giving away an out.

Again, I’m not advocating this as an across-the-board philosophy, but rather something that has to be done on occasion (as with a below-replacement-level bottom of the lineup).

At any rate, I received a fair bit of e-mail regarding run expectancy, regression/progression to the mean, etc. all of which demonstrates bunting suppresses run production. Here is where there is a disconnect between true blue sabermetricians and poseurs like myself.

I do agree that over the course of the season sac bunting will depress run production. The thing is, run scoring isn’t measured in one unit (the season) but 162 units (individual games). To state the obvious, they don’t tally runs for/runs against and award the season’s won-loss record. It’s decided by which team scored more runs more frequently than the other team in each individual unit of the schedule.

I’ve always felt that run distribution was more important than run production (like, no duh, eh?). To use a couple of examples: let’s look at 10 fictitious games. Suppose the Blue Jays played 10 games with the following scores, using their 2007 philosophy and personnel (Jays listed first):


The Jays outscored their opponents 39-21 but were 3-7 in those games. They maximized run scoring, but it got them only three wins. Multiply that by 16 (160 games) and Toronto outscored its opponents 624 to 336. If they had 10 such streaks, the Jays’ won-loss for those games is 48-112.

Sure, it’s an extreme example, but it does demonstrate that you have to focus on run scoring in that particular unit (the game), rather than worrying about what it does to run scoring when it comes time to tally the club’s total at season’s end.

Another quandary I have is regression/progression to the mean. To me, it assumes that players will revert to their norms within a tidy April-October time frame. As I understand this, it means that over time (and not a predetermined period of time) numbers will do this. In both cases, neither data point has much utility in assessing a strategy in a single unit (game) when these things require a large sample size to before the mean is realized.

Don’t misunderstand me. (I keep using this caveat because if you don’t, then folks will think you’re suggesting guys like Frank Thomas should bunt more. You know who you are.) I’m not discussing a rigid but rather a flexible approach to run scoring based on the situation at hand. Just because a strategy may cost runs over the course of the season doesn’t mean the lost runs were crucial to a particular win.

A single run in a 1-1 game is worth a lot more than the final 20 of the 30 runs scored by the Texas Rangers in one game against the Orioles last season. You cannot transfer excess runs from one game to use in another. The only runs that truly matter are the ones scored above the opposition’s total in each individual game. Why does it matter if a strategy suppresses run scoring over the course of a season if it helps win individual games?

The thing is, runs have an expiration date—once the final out is recorded in a game, their utility ceases. Wins and losses expire when the season ends and determine who takes their cuts in the batter’s box or the men’s tee come October.

In 2007, the Jays underplayed their Pythagorean won-lost by four games. The Diamondbacks overplayed theirs by 11. According to runs scored/runs against, the Jays should have been 87-75, the D-Backs 79-83.

Of course, we’re all well aware of the problem using Pythagorean projections—they don’t take into account run distribution. The Jays lost 40 games in which teams scored slightly more than a league average number of runs per game. They lost 23 in which the opposition scored four runs or fewer. These are games in which every run can be huge. These are games in which replacement level hitters destroy you. What I was trying to figure out in 2007 was minimizing the damage that hitters on the level of Hector Luna, Ray Olmedo, Phillips, McDonald, Fasano and Clayton inflict on a team’s fortunes.

Ah, well. That’s what I love about baseball—there’s still so much to learn and understand. I hope that in 2008 the Jays won’t have to give below-replacement-level hitters more than 1,000 at-bats and that first and second nobody out means a rally, and not a double play. Not to worry, there will be unstable weather systems generated in this space in 2008—some things never change.

For those preparing to celebrate the New Year by popping the cork on a bottle of vintage whoop-ass on me for the above, I have good news—THT is gonna double your opportunities!


“Heroes and Zeros” will be a feature in which I’ll be passing along props and snark for those in MLB who have distinguished themselves for better or worse in a given week. I’m encouraging my readers (the editor, mom and I when the urge to proof-read hits me) to submit candidates. If there’s someone you wish to nominate drop me a line. If you wish to be credited for the nomination let me know—if you have a blog and would prefer that it receive the credit pass along the URL and we’ll attribute it to your site (link included).

For the moment I haven’t come up with a catchy name for the feature. Ideally, I’d like to name the awards after people in baseball. For example, the award for somebody who has made a positive contribution could be called “The Good Pujols—Albert” and the negative “The Bad Poo-Holes—Loria and Samson.” If you’ve got a suggestion— let me know . (Actually, Pujols and Poo-holes ain’t bad—mind you I’m writing this early in the morning and caffeine hasn’t reached what passes for my brain as of yet.)

There may be weeks where nobody does anything positive, so it will be a week with nothing but snark—the opposite applies as well. For the moment, it’ll debut next Wednesday and depending on what happens between now and next week, I may just do a “Heroes and Zeros” of 2007.

Print This Post

Comments are closed.