﻿ Objects in mirror are closer than they appear | The Hardball Times

# Objects in mirror are closer than they appear

At approximately the midway point of the regular Major League Baseball season, I’d like to offer some sensible reminders to fantasy owners. Owners, at times, have a tendency to overestimate how wide the divide is in some categories. At the same time, owners also may underestimate how difficult it might be to catch up, or ascribe magical powers to the calendar. I’m going to devote today’s column to some level-headed—but non-revolutionary—advice about assessing leads and deficits.

First off all, at the mathematical middle of the season (or thereabout), all deficits are theoretically conquerable in a team-to-team sense. If the first place team in homers outpaced you by 30 longballs over half the season, it’s equally possible to outpace them by 30 over the second half of the season. If you have to catch up .35 ERA points, you have as much time to erase the deficit as it took for to amass it. So, time isn’t exactly your opponent’s extra man—not yet. One of the main differences in these examples is that, in order to overcome stark offensive deficits, a profound change of personnel is more likely necessary; and there’s more room to scratch, claw, and poke via strategy on the pitching side.

Let’s look at some general axioms.

Judge yourself by true pace, not points

Don’t prematurely count yourself out of offensive rate stat categories

Two years ago, Eriq Gardiner posited, “Don’t give up on batting average just yet”, and this is more easily understood if average is thought of as a counting stat rather than a rate stat. I think this is valuable perspective. To climb the summit of a .10 pt. midseason batting average deficit, it may really come down to amassing 30 more hits than your opponent from a dozen players over the course of 80 games. This is without even considering the possibility of your opponent helping you by dropping in average himself (though, of course he could gain as well).

Sometimes owners don’t fully grasp how rate stats can be so tenuous. Let’s take a realistic hypothetical to hammer this point home. Assume your team is hitting .265 and you’re chasing a team hitting .275. To reach that mark, you’d have to hit .285 the rest of the way. Let’s take a look at how small a difference that really is when you distill it from the team setting to the individual. Suppose that at the approximate midway point of the season, both of your teams are averaging 280 ABs per your 13 active roster spots.

80/280 = .286
77/280 = .275
74/280 = .264

We’re talking six hits per player over nearly 300 ABs. Now, let’s take a look at the effect of two teams simply having mismatched outliers on their respective rosters. Right now Adrian Gonzalez has a BABIP of .394 (+.75 vs. career), while Adam Dunn has one of .234 (-.58 vs. career). Now, let’s imagine that Team A and B both have a .275 cumulative average from their 12 non-1B roster spots, but team A banked on A-Gonz, while Team B went with the Big Donkey. How much would that single difference affect the batting average of these two rosters as a whole?

Team A: 924/3360 + 128/362 = 1052/.3722 (.2862)
Team B: 924/3360 + 43/269 = 967/3629 (.2665)

From this experiment, you can see that a .10 point batting average difference could be largely attributed to BABIPs that belie career norms, and therefore that such a gap can potentially be nearly erased simply by the expected regression of the rates. If Adam Dunn hit .86 points above his career AVG the rest of the way, to meet his career .243 BA, and Adrian Gonzalez his .63 points lower than his career average the rest of the way to even out with his, that alone would erase the near entirety of the current difference in team AVG.

Clearly, reality is not so simple and neat, and there are plenty of variables beyond the simple random variations in distribution of outcome that underlie these two players’ numbers, but the essential point remains. When you crunch the numbers, objects in mirror are closer than they appear.

Midseason is an arbitrary and largely insignificant milestone

The All Star break provides some downtime, and therefore serves as an inflection point for fantasy team owners. Being close to the mathematical middle of the a full season’s games helps put players’ stats into perspective, but the fact that we are now at midseason doesn’t actually mean anything for your players, their upcoming performance, or the future of your team.

The payoff of being polite.

A player is not destined to bounce back in the second half because he had a poor first half, and a player is not “on pace” to hit 30 homers because he has 15 at the halfway point. The statistical truths, mathematical underpinnings, and regressions to the mean that underlie the probability of outcomes going forward are unaware of the date on the calendar. Yes, players performing at outlying rates of production should be expected to regress to their career norms, but the fact that we have crossed the threshold from sunrise to sunset does not inherently mean such will occur or signal the beginning of such effects.

While it may have sounded like I’m distancing myself from the paragraphs above, I’m just simply emphasizing that what is most likely to occur is that which underlying peripherals support. A player shouldn’t be presumed to hit 15 homers in the second half because he did so in the first. Rather, his homer projection is still best formed by weighing his historical performance (of which the first half is a part, but just a part) and core indicators like flyball rate, HR/FB, and the like.

Much like the case of false dominance in pitching categories, don’t rely on the superficial stat totals to determine the core “pace” at which your team is progressing. At the same time, don’t look at a fledgling team and count on regression alone to ensure its dramatic resurgence. Don’t watch, act. Rationally.

Print This Post
Guest
B
Nice article Derek, thanks.  I’m in one of those situations, where I’m below the IP pace with solid W/K per IP ratios.  I am constantly paying attention to reliever situations, but not just anyone that could hurt my solid ERA/WHIP.  And with a few managers narrowly ahead of me in SV and HLD, this could help increase my chances of a few additional ranking points.  If I remain with the status quo, it might not happen. Another example is for starters who are young or recovering from injury (e.g. Zimmermann), and the risk of them being shut down early.  His… Read more »
Guest
George
Allow me to throw a bit of water on this column. First, we are NOT at the “midpoint” of the season.  Most teams have played at least 90 games and have around 45 percent of the schedule remaining, some less.  That means, conversely, that the hill you are trying to climb is 55 percent of the schedule.  Rather than having “80 games” to make up gaps, the reality is that you have more like “70 games.”  That’s a big difference. Second, while some players will have better “second halves” than others, the reality is that production in the end of… Read more »
Guest
Derek Ambrosino
George, I do understand that we aren’t at the mathematical middle of the season, but I think my underlying points hold. Yes, catching up is not easy, but it isn’t impossible, as some people believe. But, more importantly, is your contention about second half production anecdotal, or do you have some sort of cite to back that up? I’m not trying to call you out; I’m genuinely curious. And, even granting that, doesn’t that just beg the question of the composition of your team? If that is true, doesn’t it just represent a strategic opportunity to help yourself increase your… Read more »
Guest
Dave Shovein

Very nice article Derek, I was having a similar argument with a co-owner of mine the other day who thought we were so far behind in batting average that we should just punt the category rather than trying to fix it

Guest
Jeffrey Gross

Great article derek.

Guest
bennythedog

re: the middle relievers.. do you have a short list of 10 or so in mind that one who is trending over the innings limit could target?

Guest
George
Derek, you got me thinking about that.  So I went and fired up Excel with 2010 data divided by month.  I first calculated league totals for counting stats (PA, R, RBI, HR, Hits, SB), then figured the per player share of each stat, and finally took the average and standard deviation.  If the share is lower later in the year that means the league’s production of these stats is more dispersed. What I found was kinda interesting.  There is definitely dispersed production in September/October, which supports my hypothesis to an extent (and makes sense because that is when rosters expand). … Read more »