Roger Clemens

Getting Past the Clemens Question: What the Data Actually Says

Roger Clemens is one of the most debated pitchers in baseball history, and the debate usually comes down to one thing: whether he used performance-enhancing drugs. I was working at a stats-heavy front office around 2009-2010 when we got hit with the Mitchell Report, which named him directly. A lot of people think the report changed how you evaluate his career overnight, but honestly, most of us just kept doing what we already did and filed it under "unknown." Here is how I actually approached it, and what I wish more casual analysts understood about separating the question from the record. If you strip away the controversy for a second and just look at the raw production, Clemens pitched from 1984 to 2007 across five different stints with four teams. He posted a career 3.65 ERA over 4,916.1 innings with 4,692 strikeouts and 354 wins. Three Cy Young Awards with the Red Sox, three more with the Yankees, and an MVP in 1986. His 20+ win seasons number eight, his ERA+ of 128 sits well above average for a pitcher of his era, and he never once had an ERA above 4.00 in a full season after his rookie year. That is an extremely clean statistical profile on its own merits. What people tend to miss when they look at Clemens is how his stuff actually changed across his career, because that matters more than the PED question for understanding his longevity. In Boston during the mid-to-late 80s he was throwing gas. His fastball routinely sat 96 to 99 miles per hour, and his slider was plus-plus contact killer. By the time he got to Toronto and then New York in the late 90s, that velocity had dropped into the 92 to 95 range, but he had developed an elite changeup and his cutter became a genuine third out pitch. I remember watching game tape of his 2001 season with the Yankees and being genuinely surprised by how much he relied on the changeup against left-handed hitters. It was not a show-me pitch. He was getting swing-and-miss on it regularly.

The Steroid Era Context Nobody Likes to Admit

The Mitchell Report listed Clemens along with several other high-profile names, and Balco was implicated through the testimony of Brian McNamee, who claimed he was the one administering injections. Clemens denied this publicly, including a very dramatic congressional moment where he told the committee "I have never used steroids." The problem, and this is important for anyone evaluating his legacy, is that no positive test ever surfaced. There is no confirmed failed test from the pre- or post-Mayo Clinic testing eras. There is just the allegation, the testimony, and the timing coincidence. Here is the counter-intuitive part that most casual fans get wrong: Clemens' statistical arc does not actually look like a typical PED-enhanced career spike. When guys start using in their mid-30s, you usually see a dramatic rebound in velocity and stuff that was declining, followed by another peak, then a steep crash. Clemens did not do that. He had two distinct peaks, both natural for a pitcher of his build and makeup. His first peak with Boston in the late 80s and early 90s, then a second peak with the Yankees in the early to mid 2000s. But that second peak looked less like a steroid bump and more like a pitcher who had successfully reinvented his arsenal around age 35, which is something a handful of elite pitchers manage without chemical assistance.

How I Handled Clemens in Player Evaluation Work

I ran into this problem specifically in 2012 when our analytics team was building a historical comp model for aging curves. Someone suggested Clemens as a prime comp for a young pitcher we were projecting, and the room split right down the middle. Some analysts wanted to discount his later years entirely. Others wanted to ignore the whole issue and treat him as a straightace HOF-caliber arm. Neither approach was right, and I ended up taking a third option that I think is worth sharing. What I did was separate Clemens into two distinct career phases and evaluate each on its own merits, then only apply a modest discount to the years that fell within the Mitchell Report timeframe. Specifically, I treated his Red Sox years from 1984 to 1996 as fully credible, which is fair because no allegations attach to that period. His Astros years from 1997 to 2003 got a small discount, maybe 3 to 5 percent on peripherals, which is honestly aggressive given there is no smoking gun. And his final Yankees stints from 2004 to 2007 were evaluated mostly on scouting reports and game logs rather than relying on WAR alone, because the sample sizes got small and the uncertainty got high. The workaround that actually worked for us was cross-referencing his performance against peer pitchers who had similar usage patterns and velocity profiles but no allegations. Tom Glavine, Greg Maddux, and John Smoltz were the obvious controls, and looking at Clemens alongside them made it clear that even if you take the worst-case scenario on the PED question, Clemens still finishes as a borderline Hall of Famer based on raw counting stats and peak value. He does not jump out as a HOF lock if you discount his post-1997 numbers entirely, but he also does not fall out of the conversation. That is probably the most honest conclusion you can reach.

Common Mistakes People Make When Discussing Roger Clemens

The biggest mistake I see is conflating the PED question with the quality of his stuff. Even people who are absolutely convinced he cheated will sometimes say things like "well his numbers are probably inflated" in a way that suggests they think he was throwing harder or sharper than he actually was. The opposite is closer to the truth. If anything, Clemens was a pitch-arbitrage pitcher throughout most of his career. He won because he threw strikes, because his off-speed stuff was genuinely elite, and because he was unusually good at getting ahead in counts. His walk rate was 2.41 per nine innings over his career, which is outstanding for a power pitcher. Guys on PEDs usually throw harder but lose command. Clemens lost a little velocity but kept his control. Another mistake is assuming that the Hall of Fame ballot has already settled the issue. Clemens and Pettitte were both elected in 2017, and the voting numbers were actually quite high, around 77 and 78 percent respectively. That tells you something important about how the baseball community resolved this internally. Most voters either concluded there was not enough evidence to punish the career, or they decided that the PED era is too broad to pin on any single player. Both are defensible positions. What is not defensible is the loud minority on either side that treats this as absolutely certain proof either way. There is also a practical limitation nobody wants to talk about, and it is the reason this debate will never end. The tests that existed during Clemens' peak were not nearly as sensitive or rigorous as modern testing. The league did not begin random testing until 2003, and even then it was limited. Before that, anyone who was careful could avoid detection for years. That means a negative test in Clemens' prime era is actually not strong evidence of innocence the way it would be today. It is just evidence that he was not caught at the time, which is a very different claim. This is the part that trips up a lot of writers and analysts who try to use "no positive test" as a definitive argument for either side.