All Colorado Starts: Ignorable?

I’m not sure of the exact Sleeper and the Bust podcast but Paul and Eno were discussing Jeff Samardzija and how his ERA would look better without a Colorado start. Right now Samardzija’s ERA is 4.31 but if the seven run, five inning start in Colorado is removed his ERA drops to 3.81. I’m not a fan of removing starts from a pitcher’s stats. The good and bad will happen. The deal is that Samardzija owners probably sat him that week like I did in Tout Wars. If owners aren’t going to use pitchers in their Colorado starts maybe owners should start ignoring the stats generated there.

The increase scoring environment in Colorado is about impossible to overcome. In an article at Rotowire ($) last year, I stated the following:

What we do know is that the road pitcher should expect between a 1.00 and 2.00 increase in ERA. I might put the ERA drop near 1.50, but I could understand if an owner wants to use a smaller or larger number. Additionally, I will put the K/9 drop near 0.9 and the WHIP increase at .25.

Those increases are just too much to absorb over an entire season. Sure an owner can get away with a start every now and then but eventually, the odds will even out. Once the hammer drops, an owner will spend a couple weeks trying to recover their overall rate stats.

Most owners don’t take this chance and sit any starter in Colorado. Some won’t even own any Colorado starters. I ran a Twitter poll top find out how often an owner will use one of their starters in Colorado.

What's the percentage chance in a mixed league would you start a pitcher on your roster at Colorado?

— Jeff Zimmerman (@jeffwzimmerman) June 16, 2017

When I wrote this article, 75% of the responders were at 40% or less with some people writing in for 0%. If owners regularly don’t use their Colorado starts, why should the pitchers be evaluated with those starts?

For a simple analysis, here are the pitchers hurt (and helped) the most by their Colorado starts in 2016 (min 10 starts).

I figured more of a correction would have existed but after the 16 biggest drops, the difference is only a quarter of a run. Significant, but not overly concerning.

Stephen Strasburg is the one name which does sticks out. Without the start, his ERA would have been 3.08 which was in line with his FIP (2.92), xFIP (3.20), and SIERA (3.18). This season, his ERA and ERA estimators are in the low threes. The starter justed his talent level too much.

Pitchers in 2016 had months to correct the damage from Colorado but this season some pitchers have not had the time to get their ERAs corrected. Here are the pitchers who have seen their ERA balloon because of Colorado (min 5 starts)

Several names stick out like Julio Urias and Mike Clevinger who see their ERAs drop in the threes. There is almost no chance they were started so why should these games be included in their talent profile? I’m not sure I would change too many but I would consider giving a small value boost to those on the extremes. Your thoughts?