Drawing conclusions from training data

Training Q&A [Moderated]
, ,
1

Hey BBM,

given the large variance in response between individuals I’d like to figure out what sort of training works best for me for my goals specifically in some sort of controlled scientific-ish process. For example, I might want to investigate what rep ranges work better for me or which exercise variants work better, what intensity and volume produce better results, etc.

Questions:

1: For strength, at least, I can use singles @8 and/or estimated one rep maxes to monitor progress. My question: how best to go about changing variables over time? To make any sense of the data I feel like you’d have to run the same programm week to week for 4-8 weeks say and then only change one or two variables for the next block. Then there’s also the issue that I won’t be the same person after 4-8 weeks. How do I know that the change of variables produced the different result and not just the fact that I’m also adapting to the stimulus over time?

2: What about measuring hypertrophy? Given that my arms aren’t going to grow by centimetres month to month, thus making circumference measurements less useful, how would I go about measuring progress? Is the solution to take regular photos in the same place and lighting? I often hear bodybuilders talking about how this exercise or that workout really worked well for them so I feel like it must be possible to draw some conclusions about what works better, but I’m at a loss as to how.

In short, how can I use training data to improve my training?

2

Paul,

You’ve identified problems with standard, reductionist programming approaches, e.g. “just change one thing at a time” and “post hoc ergo propter hoc.”

We discuss these things quite a bit in our programming podcasts, but the short answer is- it is very difficult to feel confident that what you did programming-wise is directly tied to an immediate outcome and is free from confounders. If the difference in outcomes is large and sustained, you can feel more confident that what you did worked, sure.

In any case, a few ideas here for each question:

  1. In general, modifying the formulation, e.g. the exercise selection and average intensity, prior to the dose, e.g. the volume, is a reasonable place to start. If that doesn’t seem to work based on metrics you’re tracking, volume would be the next tweak.

  2. I find anthropometric data very useful, but it needs to be measured over a longer period of time. I find them more useful than pictures, which are subjective.

-Jordan

1 Like
3

Hey Jordan,

thanks for the suggestion about altering formulation before dose and it’s good to know that I’m not crazy here. It really is very difficult to draw clear conclusions in the absence of large sustained effects. I listened to your programming podcasts, btw and am a big fan of your guys’ stuff in general. I find myself recommending it a lot.

Some questions on the second point: you say anthropometric data is useful. How do you use it? I’m imagining you might change some programming variables if you were regressing. But how much emphasis would you place on the rate of progression? Are you calculating standard deviations to get a feel for how significant changes are? Finally, given that you would likely have to collect data over months to be confident about estimating your rate of progress, what variables would you change? Would you still just start by switching out some variants/playing with intensity, etc. or would you make larger changes so that you could be more confident about future conclusions?

4

Yea I think you’re asking for a longer explanation, one with caveats, examples, etc., than I’m in a position to give right now.

That said, I wouldn’t expect that individuals respond uniquely well (or poorly) to a particular variation with respect to hypertrophy given similar levels of volume tolerance, ROM, mechanical loading (not weight, necessarily), and ability to train near failure. This also ignores the contributions of energy balance and dietary pattern.

To be clear, I don’t think intensity or variations are that important with respect to hypertrophy if the above are in check. Additionally, hypertrophy outcomes are typically delayed a few weeks as tolerance to the training develops.

So, I use anthropometric data just to make sure growth is occurring. If it’s not, I typically play with the variation first again- just to double check it’s not a formulation issue. Then, weeks later, I’ll play with the dosing.

There isn’t a standard deviation that could be applied to a population either, so no, I don’t use that. In any case, I wouldn’t expect a regression of muscle cross sectional area without substantial weight loss.

5

Yea I think you’re asking for a longer explanation, one with caveats, examples, etc., than I’m in a position to give right now.

No worries. Thanks for taking the time. I already got more than I expected :slight_smile:

That said, I wouldn’t expect that individuals respond uniquely well (or poorly) to a particular variation with respect to hypertrophy given similar levels of volume tolerance, ROM, mechanical loading (not weight, necessarily), and ability to train near failure. This also ignores the contributions of energy balance and dietary pattern.

To be clear, I don’t think intensity or variations are that important with respect to hypertrophy if the above are in check. Additionally, hypertrophy outcomes are typically delayed a few weeks as tolerance to the training develops.

So, I use anthropometric data just to make sure growth is occurring. If it’s not, I typically play with the variation first again- just to double check it’s not a formulation issue. Then, weeks later, I’ll play with the dosing.

This is the kind of advice I was looking for. Since I’m still a beginner making fast progress it’ll probably be a while until I need to get into more detail anyway. So, thanks for this!

There isn’t a standard deviation that could be applied to a population either, so no, I don’t use that. In any case, I wouldn’t expect a regression of muscle cross sectional area without substantial weight loss.

Sorry, should have clarified. I was thinking of taking all comparable hypertrophy blocks and then calculating standard deviations in muscle growth (as measured by circumferences and other anthropometric measurements). That way, for example, you might be able to deduce that even though you weren’t regressing, you were progressing 1.5 standard deviations below average, thus making you think that some programming variables need to be changed.

Basically, I’m wondering how much you mess about with statistics and how much I can expect to get out of messing about with my own training data. Another example: you could calculate correlations between various programming variables and your own progress. My fitbit does a fairly decent job of detecting when I’m asleep and I weigh myself regularly, so I can correlate those variables with training outcomes as well :). Regardless of your answer I’ll probably keep collecting and playing with data just because it’s fun. If it turns out to also be useful, even better!

6

Yea, you can’t really do this productively, given that hypertrophy results are not linear (or even consistent) over time.

I keep track of some variables I find useful, but I think we should be generally cautious in how we apply it.