The Z Files: The Cole Park Facts

HUGE upgrade for Gerrit Cole. Without adjusting innings (not sure it'll be a big difference)PIT: 3.79 ERA, 1.25 WHIP, 199 K, 12 WHOU: 3.53 ERA, 1.21 WHIP, 212 K, 15 W

— Todd Zola (@ToddZola) January 10, 2018

Ignoring my original projection for Cole, which many consider optimistic after his final two seasons in the Steel City, the point is the move to Houston, on paper, should improve Cole's stats across the board. I say, "on paper" because park effects don't always manifest linearly. Some players are influenced more, some less. The park factors average everyone on that team, that year, keeping in mind we use a three-year average of park factors to help flesh out some of the luck and bias from a single season. For a deeper look at park factors, check out this archived piece, The Necessary Evil of Park Factors. I'll be refreshing it in an upcoming Z Files.

Let's look at each component individually.

ERA

A common misconception about park factors is that home runs equate to runs. There may not be better diametric examples of why this is false than PNC Park and Minute Maid Field.

The three-year park factor for homers at PNC Park is 89, making it the sixth toughest park on homers in the league. Over that same span, it played 96 for runs.

When doing projections, only half of the factor is applied, since half the games are played at home. This assumes the collective away venues play neutral, which they don't. Making this adjustment, the factors for projection purposes are 95 for HR and 98 for runs.

To get a better conversion, I use composite park factors, taking the actual away parks into account. Doing so, the HR factor for Pittsburgh players is 96 while the run index is 101. These are close to, but not the same as, the version most everyone else incorporates. Regardless, the big-picture takeaway is that collectively, Pirates pitchers give up fewer homers but more runs than they would working exclusively in an all-around neutral park. Fewer homers did not result in fewer runs.

Looking at Minute Maid Park is tricky, since Tal's Hill was removed before the 2017 season. For those unfamiliar, this was the goofy incline in center field. Not only did outfielders need to run uphill, they needed to deal with a huge flagpole. The renovation brought the fence in from 436 feet to 409 feet at its deepest point. The industry standard for park factors, the Bill James Handbook, chose to use just 2017 for its three-year average, consistent with what's been done in previous seasons with parks undergoing a physical change. They also warn to be cautious using the one-year number. There's a reason why the three-year average is used – one season could be noisy.

Using some park overlay data, I adjusted Minute Maid Park's HR index for 2015 and 2016, which is nice, but we don't know exactly how that would affect runs. Using my best estimate, the composite HR index is 103, with a 95 run factor. This trend is opposite that of PNC Park. The bottom line is, on paper, Cole should enjoy a park upgrade in terms of runs.

"On paper" and "should", way to take a stand, Lord Vagueness. Keeping in mind not all players realize the exact effect of the park factor, Cole's primary issue with the Pirates was homers. Minute Maid Park inflates homers. There's no guarantee the added homers don't influence Cole's ERA. The other consideration is new team, new coaching. I'll tag in my friend and colleague Eno Sarris and invite you to read his recent blog on Fangraphs. The short version is, Houston may be able to reduce Cole's gopheritis by having him throw more curves.

WHIP

The composite hit factor for PNC Park is 101, with a walk factor of 100. The corresponding indices for Minute Maid Park are both 97. Cole's new home should lower both his hits and walks, and hence his WHIP. Another factor is intentional walks, as the NL rate is a tick higher than the AL, chiefly from allowing the eighth hitter a free pass to first base.

STRIKEOUTS

The composite K factor for PNC Park is 97, compared to 104 for Minute Maid Park. In terms of relative factors, this is a big difference. Despite moving to the American League (more on this in a bit), Cole's strikeout rate should benefit from the park switch.

To anticipate the question, venues influence walks and strikeouts a few different ways. I can't pinpoint the exact reasons for the above numbers, but foul territory, batter's eye and atmospheric conditions are three common influences. More foul outs mean fewer strikes from foul balls and more PAs ending at contact, hence fewer strikeouts. A better batter's eye results in more contact, thus fewer whiffs. Movement on pitches is affected by humidity, atmospheric pressure, etc., which in turn can improve contact, ergo strikeouts.

WINS

This one's easy: fewer runs allowed, more run support with a better bullpen leads to more wins.

Now it's time to deal with the 400-pound elephant (or is that the gorilla?) in the room. Either way, the matter of how the move from the National to American League impacts pitchers needs to be addressed. Conventional wisdom is pitchers' numbers decline going from the Senior to the Junior Circuit, due to having to face a designated hitter and not a fellow hurler. The underlying assumption is the rest of the player pool in each league are of equal talent.

Over the years, this has been borne out, except for the last two seasons:

I'll admit, this has my wheels spinning, so much so I've already sent two notes to my editor, letting him know filing will be delayed. My initial reaction was, "Do we still need to account for the league switch as well as change in venues?" To be honest, I'm still not 100 percent sure. Here's some of the research undertaken to help me decide. Feedback is encouraged in the comments.

As cited earlier, the assumption is the non-DH talent in both leagues is a wash. That is, run-scoring differences between the leagues emanates primarily from pitchers hitting in the NL, hence fewer runs.

What if the rest of the talent isn't the same? Further, what if it isn't just the talent? What if it's the difference in skills for AL hitters and AL pitchers compared to NL hitters and NL pitchers? By means of the simplest example, the skills of the non-DH hitters in each league could be the same, but the AL hitters face worse pitching. The numbers would suggest the AL hitters are better, but it's really the pitchers are worse.

Thinking about this, my head started to hurt, so I decided to look at some numbers, hoping this would help elucidate what's happening. I only went back three years, since that's what most projections use as a basis. The study looks at overall league data, interleague data (NLP vs. ALH, ALP vs. NLH) and same league data (NLP vs. NLH and ALP vs. ALH).

ERA and WHIP

In each case, the NL ERA is lower when facing exclusively NL hitters. The factor bringing the overall league ratios together is interleague play, but the gap has narrowed. That said, the American League has still dominated interleague play but has won fewer each of the past two campaigns:

So, what does this mean? The American League is winning more games, but are their hitters better, meaning Cole (or any pitcher switching leagues) will suffer? Are AL pitchers better, but the hitters are the same, so Cole won't be affected? Now my head really hurts, and this filing is really late.

Perhaps focusing on skills will lend some clarity:

It only makes sense the skills trend dovetails with run scoring. The gap in strikeout rates between the two leagues is narrowing. The walk difference is about the same, with the National League seeing more bases on balls due to intentional walks. Homers flip-flopped in 2016.

Does anyone have any aspirin, preferably extra strength? Can I get a couple more for my editor?

After all this, I keep coming back to the ERA and WHIP for the two leagues being nearly identical. Does it really matter why? Isn't the operative point that it is? I actually use league translation factors in my projection engine, but all they're doing is canceling each other out. I very well may be missing something, and therefore wrong. Please feel free to show me the error of my ways, but at least over the past two seasons, no league correction factor was necessary.

Of course, there's a "yeah, but...". For projection purposes, this assumes whatever is overcoming the DH factor will persist in 2018. We've already seen Giancarlo Stanton head north from South Beach to The Bronx. It's only one player, but he's the perfect example to highlight that the makeup of the leagues could change. On the other hand, perhaps Ronald Acuna and Victor Robles spearhead a more impressive set of NL freshman, tipping the balance back?

When we don't know what will happen, the proper action is regressing towards what seems like should happen. In this case, all other things being equal, pitchers batting in the NL and not in the AL fuels the difference. This is the mean and all calculations should be regressed accordingly. That said, the effect won't be enough to detract from the main point, which is that Cole enjoys a significant uptick from the trade.

Now, you'll have to excuse me, I must adjust my projection engine. But first, I need a drink.