Can you guess how hard someone else trains?
Take-aways
The accuracy of your prediction of someone else’s proximity to failure is primarily down to making that prediction as late into the set as possible, familiarizing yourself with the participant’s training and is facilitated through heavier loads.
Coaching experience plays little role in the accuracy of your guesses. Exercises with a smaller technical component may be slightly easier to gauge.
Generally, if you make your predictions under the right circumstances (see (1)), you’ll likely be within a couple of reps of how close to failure someone else truly is. This is still less accurate than someone is at guessing their own proximity to failure, at least in lab conditions.
On this blog, we’ve covered how accurate you are at gauging your own proximity to failure. In short, under 12 reps, you typically overestimate how close to failure you are by just about 1 rep. If you do a set of much more than 12 reps, you tend to become less accurate, over-estimating by up to ~5 reps when doing sets of 20-30 reps.
However, what about coaches? Can other people accurately estimate your proximity to failure? Or, if you’re a coach, can you accurately estimate your clients’ efforts and proximity to failure?
This is what a study by Emanuel and colleagues sought to answer. They showed 250 coaches videos of trainees performing sets of preacher curls and barbell squats to failure with 70 or 80% of their 1 repetition-maximum to technical failure over two sets. Coaches gave estimates of proximity to failure after 1/3 of the set, 2/3 of the set and 90% into the set.
Crosses represent actual number of repetitions left to perform in a given set of preacher curls. The coloured boxplots represent the predictions made by coaches.
Coaches tended to underpredict how many more reps trainees could do when predictions were made earlier in the set. As the set went on, coaches generally became more accurate, but tended to start overpredicting how many more reps trainees could perform (see 90% predictions). Coaches became substantially more accurate from set 1 to set 2 for a given trainee, indicating some familiarization to how a trainee’s sets tend to look is helpful.
Crosses represent actual number of repetitions left to perform in a given set of barbell squats. The coloured boxplots represent the predictions made by coaches.
When trying to analyze the data further, a few interesting findings emerged.
First, on average, when guessing proximity to failure only 1/3 into a trainee’s set, coaches were off by ~4.5 reps (underestimating trainees). However, by the time they made predictions 2/3 or 90% into the set, they started overestimating trainees by ~1.5 and ~2 reps, respectively. Generally, coaches were also about one rep more accurate in their prediction when going from a lighter to heavier load being used (70 to 80% 1RM).
Coaches were also generally more accurate at guessing proximity to failure on the preacher curl vs the squat, potentially owing to a smaller amount of moving parts and a smaller technical component. With that being said, the study also seemed to underestimate their 70%1RM, as the participants in the study were able to perform 22 repetitions with this load, on average. Therefore, with coaches being told this was 70% of 1RM, that information likely played a role. So, while I still think it is easier to estimate RIR on a simple movement, the difference may not be that large.
Besides the above, the only other significant predictor of accuracy was coaching experience, where each additional year of coaching experience increased accuracy by.. 0.02 reps. Not exactly a meaningful improvement compared to load, familiarity with the trainee or making the prediction later into the set.
