I just watched your “What 17 Years of Data Tells Us About Strength Progression” video, and I’m confused about the conclusions you’re drawing from the Latella study data. If I’m understanding correctly, they were studying competitive powerlifters, and the average gains across a 10 year period were only around 20%. In the novice/intermediate/advanced framework, wouldn’t most of those lifters be “advanced” already? If that dataset included novice lifters, I would expect to see more like a 20% increase in a month or two rather than in a decade. I think this is the equivalent of trying to draw conclusions about overall basketball skill progression by looking at NBA players, when most of them are already close to their skill ceiling before entering the dataset.
I agree with your position that novice/intermediate/advanced don’t represent biologically distinct categories, but I’m not sure how the Latella study supports that at all.
Howdy!
That video is a segment of our long-form podcast and article.
You’re right that IPF meet 1 isn’t day 1 of training, and the Latella dataset can’t see what happens before a lifter’s first meet. Where I’d push back is on how much weight you’re putting on Latella in the overall argument.
The article describes five physiological lines for why the same adaptive machinery runs at the same rate year one and year ten: neural, hypertrophy, connective tissue, bone, and fatigue. Those do the load-bearing work. Latella corroborates this.
On the dataset gap you’re pointing at, Steele and colleagues 2022 ran the same growth model on 14,690 fit20 members over up to ~6.8 years. Gains were 30 to 50% in year one and 50 to 60% total after 6 years, roughly 5x the powerlifting cohort in the same paper, consistent with a much less trained starting population than IPF entrants. Same linear-log shape. Their changepoint analysis did find a slope change around 26-31 weeks with wide individual variability, but that’s a change in the rate of gain. The fit20 protocol is completely unvaried from session 1 through session 350, and it continues to produce substantial adaptation across the whole time period. A single unvaried program working across ~7 years cuts against the framework’s claim that tiers require categorically different programs.
Your prediction that true novices would show ~20% early is roughly right, and Steele’s fit20 numbers match that (fast early gains, then slowdown). Where I’d separate is what that shape means. The same reactive adaptive process operating against more remaining headroom produces exactly those fast early gains without requiring a biological switch. Fast novice gains are consistent with both the framework’s biological interpretation and a ceiling-on-a-continuous-process interpretation, so they don’t distinguish between them. The physiology sections do.
We agree is that the categories aren’t biologically distinct. Where I think we differ is whether Latella and Steele together support that. My take is that combined with the physiology, they do.
Appreciate the response; I think the long-form article clears things up quite a bit. The Steele study in particular is quite interesting.
1 Like
Yea. The whole thing is very interesting to me obviously. As you may have predicted, some SS diehards did not take this very well and have engaged in some very strange mental gymnastics. It would be interesting to have a dispassionate conversation with a NIA supporter.
I’ll admit that I’m not sure I fully understand what the NIA position is. It’s been a while since I’ve read any of the SS-adjacent works, but I always internalized those terms as something very similar to what you write about how “NIA labels track how often a lifter’s gains are large enough to be obvious.” Maybe they articulate it a bit differently, but I don’t see much more than a semantic distinction.
From my perspective, the real disagreement is between proactive and reactive programming. That was the big insight that helped me realize why I had struggled so much with the SS-style of progression.
I think if it was just about different adaptation detection intervals, I probably wouldn’t say much about it. They don’t really discuss it like that, and their position adds/conflates it (the adaptation window) with how often weight should be added to the bar, and the need for fundamentally different programming to accommodate that. I think it would be better for them (and others with similar views) to concede that the rate of adaptation is highly variable, making fixed progression schedules problematic.
1 Like