Talent vs Steroids — what matters more?
Talent? Steroids? Or both? That is the question that I’m going to answer today. Specifically, I will look at Powerlifting and use real data from openpowerlifting.org competition database to see how much of what matters in this sport.
First, a brief disclaimer. There will be a lot of speculation! Like, a lot! Sure, with numbers and math, but speculation and assumptions nonetheless. I’m also not a statistician, so whatever I’m doing is probably not correct. Please take this as a fun home project, not as a fact or truth. I’m not claiming that, even though I believe the results I show are within the realm of possibility.
So, that out of the way, let’s start.
Data
Before going deeper, I’d like you to understand what Normal Distribution means, since I will base all of my assumptions on this concept. In lay mans terms, it means how often we encounter a certain variable. For example, human height follows such a curve where most people (middle of the graph) are of a certain height and some are taller or shorter, with outliers (gigantism/dwarfism) being uncommon. Using standard deviations (SD) we can determine the probability of having a very tall or short person in a population. For example, SD 0 means that roughly 50% of the population is of a certain height. While SD 1 — roughly 16% (the sum of all percentages to the right) and so on. You can find the exact numbers at www.z-table.com
Strength, also follows such a curve. Here, look at IPFs (and affiliates) 83 kg weight class tested Mens Open athlete results (7680 total).
Note, I removed far left outliers (10 total). Because I believe they were more of an anomaly (injuries, purposefully light lifts).
It looks roughly normally distributed, doesn’t it? Good news for us. That means we can assume that most lifters lift an average total of 525 kg, while extreme outliers can lift a 3–4 SD more (780–865 kg as shown later) where SD is 85kg.
Assumptions
The first big assumption we need to make is whether the results are truly drug-free. The obvious answer is no. But saying that everyone or most of the athletes used steroids in drug tested competitions is overzealous in my humble opinion. So, I will take the liberty to assume that these results ARE from drug free athletes. And later on you will see why that can be true!
The second assumption is even bigger — how much relative strength do steroids add? Here, I borrow some data from Greg Nuckols article Steroids for Strength Sports: The Disappointing Truth. To cut it short, he believes that relative advantage one can get from using steroids strength wise is roughly ~10%. I agree with him. Read up his article for the reasoning, but for my intent and purpose I will only need the percentage difference of tested and untested world total records.
Now, it is known that people respond differently to steroids. Some more, some less. This means that there’s probably a normal distribution in there somewhere. Given this, I can calculate the average (~4.9%) and SD of 3.64%. This means that roughly 50% (SD 0) of the powerlifting population will get roughly 5% relative strength boost, while the outliers 0.003% (SD 4) will get a 20% boost. Take, note, I say relative. Do not confuse this with absolute strength gains which are large even for an average responder.
While I base the two previous assumptions on some numbers, I have nothing for the third one. Basically, I assume that it doesn’t matter whether or not the person is strong, the response to steroids can be anywhere from 0% to 20%. For example, a gifted natural can get a tiny boost, while an average Joe in terms of strength can get a big boost. Surely, this is probably not entirely true. But this assumption allows me to calculate probability of what kind of total one can expect at a given SD for natural potential and what steroids can give as a plus.
Results
As you can see, assumptions were wild and should not be taken seriously, but let’s pretend they are true enough and dive into results:
There’s a lot to unpack here, so bear with me. First, the green block, these are real numbers based on data I have from openpowerlifting.org. As you can see, I start with what 50% of the people can do and move until only 0.003% with the SD 4. 1 in X it’s just an inverse of probability — for readability. For example, out of all powerlifter population 1 in 43 totals 695 kg. The other blocks are with different boosts from steroids separated by half a standard deviation, starting at a 5% boost for half the population.
Now, what numbers here seem unrealistic to achieve naturally? Anything above 600kg? 700kg? 800kg? Good question, but I can only speculate from personal experience. I am 176cm tall, weigh 80kg. I have a wrist size of 17.5cm. FFM of 70.5kg and FFMI of 22–23. Vertical jump of 66cm. My last total was 660kg (450 wilks). Now is my total unrealistic naturally? Well, clearly, not, since I am a lifetime natural. What about 700kg? I believe I can reach it, even though with exception of vertical jump, which is above average, I’m average by all other accounts. In fact, I’d say I’m severely underweight for my height in powerlifting. I should be in 93kg or even 105kg class. But I can’t because I can’t add anymore muscle.
What does all of this mean? Well, you can say that I’m not average, but slightly gifted, maybe because of muscle fiber type distribution, neural coding or muscle insertion points or some other unknown factors. Or maybe because I trained for 22 years (I’m 34, started at 12, not seriously though). That or the average I get from data is too low. But it doesn’t matter! Both cases still allow someone to reach 800kg naturally. Here’s why.
Let’s take world champion as another example. Brett Gibbs. 830kg total in 83kg. 164cm tall. Have you seen him (First pic)? Built like a tank. Legs like tree trunks. I don’t know his other measurements, but I can speculate he has an FFMI of 26–27 (13–15% BF) and probably an FFM of 70 kg. Which is the same muscle mass I have, but for height that is 12cm shorter! Wrist size is probably also bigger than mine (check out the pic again!). He is by far superior to me in terms of natural talent. So, if I can squeeze out 650–700kg naturally, what’s stopping him from squeezing out 800kg+? Ok, maybe he naturally was in 74kg and used steroids to get into 83kg. But if you look at his frame size (and competitive history, where he mostly stayed in 83kg), it looks like he can get 70 kg of FFM easily (the bigger the frame, the more muscle mass it can hold). Anyway, we don’t know. But we should assume innocent until proven guilty, it’s not freaking middle ages!
But my point stands. If we are to believe the normal distribution, there are people who are SD 4. And if we follow the data we have (which is not unreasonable), we haven’t seen true talent yet. Maybe it is possible to total 900kg naturally? We well know it that powerlifting is not a popular sport even with the recent 5 year surge. Now, if it had real money in it, the numbers we have now would look like peanuts and a lifter like me would not break out of the 50% range. At least, that’s what I speculate.
Let’s get back to numbers in the picture above. An avid reader would notice that the probability of achieving 830kg is higher in the +steroids block. And that is true. Here’s an example on what is the most likely combination of natural talent + steroids for that total:
So, it turns out, it is more likely that 830kg can be achieved with slightly less natural talent and an average boost from steroids. Does that mean that Gibbs is on steroids? Of course not! It just means that given the distributions of totals and effect of steroids we get a higher probability of having someone achieve that number with the help of substances than without. Regardless, a possibility to have a powerlifter 1 in ~6000 to get the highest natural total is not that far off a stretch in my opinion, considering how many powerlifters are there.
The reason I believe so is that there is no way in hell half the population of healthy males in their prime years that could not total 500kg with enough training. Given this and given the standard deviation of 80–100kg, it is not unreasonable to find a freak of nature or two? Have you not seen men barely in their 20s breaking world records? Or teenagers lifting more than most men can imagine? If anything, my numbers under-predict, as I’m certain that majority of powerlifting population doesn’t train seriously enough (or long enough) to achieve their max potential.
Moving on, I want to look at John Haacks totals both Tested and Untested. In 2016, he got a total of 813 kg, while in 2019–907.5kg. Now if we assume (not claim!) he was drug free back in 2016 and is supposedly on steroids in 2019, here’s what we get:
When we combine the tables, we can assume that Haack has a 3.4SD natural talent, and a slightly better-than-average response to steroids. In reality, probably, just average, because Untested also has 24h weigh-ins, deadlift bars, and different rules (heels up in bench for example) that can help boost the total. The other outcome is that he is 2.9SD natural talent but an extreme responder to steroids (3.5 SD). The likelihood, though, is tiny, so if I had to bet, I’d go with a natural talent. For those who took note that Untested probability is much lower, let’s not forget my third assumption which basically stated that natural talent for strength and response to steroids is completely unrelated. That is probably not true and the probabilities should be much lower for Untested.
Last but not least, in number crunching, a total of 655 kg (#1 DRUG FREE ~1.5 SD 6.67% 1 / 15) or less has the highest probability of being from talent alone. Meaning that if you had to bet whether this total is achievable naturally, bet on YES. Also, if you are an average 83kg powerlifter with a 525 kg total and have an insane response to steroids, your relative strength will increase only to 627kg! That’s less than what has the highest probability of being achieved naturally at 655 kg! Crazy, right?
Conclusion
So, there you have it, folks. Wild assumptions, random numbers, unqualified data analysis and I’m telling you — steroids don’t matter that much. Talent does. Now combine great talent with even average steroid responses and you get top level monsters. Does that mean that all of them are pinning? Absolutely not. Call me naïve, but from the numbers I see, it is possible to be naturally talented and reach the top of the mountain! Not everyone is on steroids. Probabilities exist and are indicative of the outliers in our population who can move those insane numbers.
With that, I’m finishing up. If you’d like to discuss the article, I’m all ears. I had fun writing it, slicing the data, trying different approaches to see what I can tease out from it. So, if you have questions? Blast them! I will be happy to answer.
