The View From Wisconsin

It's not too often that Bill Simmons (aka The Sports Guy for ESPN:TM) brings up a question that really intrigues me. And, in his latest column for The Magazine (June 18th ), he tackles a very intriguing question: "What was the purest baseball era, from a statistical perspective?" He gave some reasoned arguments as to why the era from 1988 to 1992 was the relatively "purest" era in major league history.

Of course, being the stathead that I am, I figured that Mr. Simmons might just be on to something here. And, if it was indeed true, it would show up in statistical analysis over the history of MLB. Lucky for me, I had the Lahman Baseball Database at my disposal – and, consequently, the league totals for every season from 1871 to 2008.

I decided that I didn't want to go too elaborate, since I couldn't easily measure everything that he mentioned in the articles (innings by starters and pitchers used per game, for example). So I decided to limit my season-by-season analysis to two basic areas: runs scored and batters reaching base. Those two were measurable both on the offensive and defensive side, which was an added benefit. I figured two sets of totals – one for the AL and NL combined from 1988 to 1992, and one for the major leagues from 1871 to present. The stats I used were:

• Runs per Plate Appearance (.11009 for 1988-92, .11894 for 1871-2008)
• Men on Base per Plate Appearance [(H+BB+HB)/PA] (.31832 for 1988-92, .32295 for 1871-2008)
• Earned Runs per Inning (.42041 for 1988-92, .42423 for 1871-2008)
• Hits and Walks per Inning [(H+BB+HB)/IP] (1.35220 for 1988-92, 1.38604 for 1871-2008)

You'll notice that each of the four were "rate" stats. That is to even out totals affected by the number of games played over a season, which changed drastically over the course of major league history. I took the individual totals for each league/season and compared them to the total for the "normal" era. Next, I averaged all four categories to get a league "normalization" total. For example: 2008 NL vs. 1988-92 ML totals are 1.06657, 1.02867, 1.13331, and 1.05682, for a 1.07134 Normalization Average.

To make it easy to understand, I took this total, subtracted by one, multiplied the result by 100, and removed the negative sign (for you math geeks, that would be called "taking the absolute value of the result") to indicate difference from the zero-point, calling this the NORM total. In our example, the 2008 NL had a NORM percentage of 7.13%. That means that the 2008 NL was significantly far off from the target totals from the 88-92 era.

I broke this down two ways, once for each individual league and once for the majors as a whole: first, I compared individual seasons to the "norm" totals; next, I averaged every five year period of "norm" totals. I did it this way because it was both easier and because the 1988-92 era wouldn't automatically jump out from the statistics. Last, I did all four of these comparisons for both the 88-92 "norms" and the overall "norms".

First of all, let's look at the five "most normal" single seasons in MLB – in each league and the majors as a whole:

• Based on the 1988-92 standard: 1952 Major Leagues (0.17% from the normal)
• Based on 1871-2008 Major League Totals: 1962 Major Leagues (0.24%)

That last one should be a surprise, as it was the first year of the 20-team majors, and featured the return of National League baseball to New York City. The "longball" era wasn't quite over, while teams hadn't gone crazy with steals like the White Sox had in 1959. Teams were starting to become more integrated, and pitchers hadn't yet taken dominance over hitters like they would six years later. Astroturf wasn't even invented yet, and contracts hadn't become mail-it-in level – at least, not yet, and not knowingly.

Looking at individual league/seasons, things get more interesting – and we see the first appearance of seasons from Mr. Simmons' "ideal era" make the top five in each of the norm standards:

"Most Normal" Seasons in any Major League
Based on the 1988-92 standard:
1. 1990 National League (0.061%)
2. 1911 National League (0.068%)
3. 1979 National League (0.091%)
4. 1958 American League (0.098%)
5. 1952 American League (0.128%)

Based on 1871-2008 Major League Totals:
1. 1962 American League (0.035%)
2. 1945 National League (0.058%)
3. 1958 National League (0.151%)
4. 1992 American League (0.170%)
5. 1959 National League (0.219%)

The 1962 American League's comparison ratios, by the way, were .968 for RPA, 1.039 for ER/IP, .996 for BRPA, and .9978 for WHIP. Those are the closest any single season has come to matching the overall ML rate stats in our four comparison categories.

Now, as to five-year periods in baseball: you would expect that the 1988-92 era would still stick out statistically from the rest in comparisons, even though I wasn't measuring it the exact same way.

You would be wrong.

"Most Normal" Five-Year Periods in Major League Baseball
Based on the 1988-92 standard:
1. 1956-60 (2.15% from norm)
2. 1957-61 (2.24%)
3. 1973-77 (2.30%)
4. 1974-78 (2.34%)
5. 1958-62 (2.71%)
Based on 1871-2008 Major League Totals:
1. 1958-62 (1.23%)
2. 1957-61 (1.59%)
3. 1955-59 (2.00%)
4. 1956-60 (2.08%)
5. 1954-58 (2.33%)

I'm pretty sure you can see a pattern here. MLB was as close to the norm – both overall and in comparison to the "Golden Age" of Simmons – in the era right around 1958-61. In fact, I'd even be as bold as to stretch it to a seven-year period between 1956 and 1962. That era saw the greatest amount of change ever seen in the history of the game: complete integration, new ballparks, new cities, new teams. The clincher is the five-year spans in each of the two leagues:

"Most Normal" Five-Year Periods in the AL or NL
Based on the 1988-92 standard:

• 1956-60 National League (1.67%)
• 1957-61 American League (1.96%)

Based on 1871-2008 Major League Totals:

• 1988-92 American League (0.73%)
• 1958-62 National League (1.10%)

And it's this little piece of information that tells me volumes about why our Sports Guy believed that the era from 1988-92 was the "Golden Age" of baseball: Mr. Simmons has a small amount of bias in his sports view. Namely, he sees only one side of the game – the American League side.

During that five year period, he probably was right on that the AL was as balanced as you could get. However, the NL during that time frame was far from balanced. The league, if you recall,
consisted of seven teams that played on Astroturf – one, of course, who was the namesake for the stuff. 10 teams played in cavernous multi-purpose ball fields. Another park, Candlestick, was a nightmare of a place for the entire era. Teams were rewarded for slapping the ball around and manufacturing runs instead of getting runners on base and then driving them home with base hits.

That's not the case for the Majors in the late 1950's and early 1960's. If I had to choose a five-year period that was truly the "purest" in comparison to what baseball should be, I'd say 1957-61. I'd be more willing to extend it out to a seven-year period from 1956 to 1962, where the game went from being 16 teams with three based in New York City, to 20 teams with teams literally from coast to coast. Add to that the amount of integration that occurred during that time, and I'd say you've probably got an era that is as "pure" as you're going to get.

Labels: Golden Age Baseball Bill Simmons