T Nation

Advice/1-RM Prediction Equations


I am a student who is conducting a simple small scale test to predict 1-RM from multiple rep predictions. I have observed that the strongest people in most gyms (presumable with a high ratio of fast twitch muscle fibres) aren?t represented by 1-RM prediction tests in that they are not usually able to perform as many repetitions with a % of their 1-RM as would be predicted from theses equations. For example I can incline bench 400lbs 1-RM and on this basis should be able to do 5 reps at 90% and 8-10reps at 80% when in reality all I am able to do is 3 reps at 90% and 7 reps at 80% and I have witnessed this phenomenon on a number of occasions. Therefore if these prediction equations were used for me they would under predict my 1-RM. My intention is to work backwards by measuring 1-RM in a strength trained athlete and then 48 hours later perform as many reps as possible at 90% of the 1-RM and then another 48 hours later perform as many reps as possible at 80% and then compare this with the current literature and recommend a prediction equation which is specific to this individual.
Any thoughts on theories on this phenomena such as fibre type, Habitualisation, etc would be really appreciated. Additionally if anyone can think of ways to improve the testing protocols and suggest relevant references I would be grateful.


I'm sorry, but in my experience these equations are worthless.

However, for me increasing any number usually will cause the other numbers ot increase. If my 1 rm increases, my 10 rep max and so on increase.

The problem is taking these %s and applying them to someone else.


Such equations seem inherrently flawed given what I've read about Type I and Type II muscle fiber and how it affects lifting capacity to have more of one versus another.


Thats tricky. The closest I've come to a formula for that was by recording the numbers the using a Ti-89 and plotting rep max numbers to create an equation that closely matches. That was only correct though as y approaches one. Once the curve changes direction, the graph reads your max increasing with reps.(parabola)

The problem is that the rise of weight is inversly proportional to the drop in reps. This would require a rational function with a diagonal asymptote, and that is a prety darn hard equation to model. You need polynomials in the numerator and denominator.
Even then it will only be correct within a small range.

Maybe a mathematics major could help.
Good luck.


What does the warm-up protocol look like?



That's one hell of an incline bench!


Look at "The Black Book of Training Secrets" by Christian Thibaudeau

Brzycki, Epley, Lander and Weis all wrote their own formulas for calculating 1rm. If I remember correctly University of Southern Mississippi also wrote bench press formulas for ranges of reps. NSCA also has tables to calculate 1rm. There are others as well just do some searches in the literature.

There are programs to fit any number of equations. I know of 12 types of equations that can be reformulated to linear regression. Polynomial equations to 12th degree are common. Rational polynomials are of the same order of difficulty as regular polynomials to solve. Logistic equations are harder to solve. I am listing just some of the tools you should have available at any good school. I have seen software for many more types of equations.


I have found similar results with myself and the "almighty 1RM table", My best suggestion would be to try using the 80% test that one poster aluded to; its in CT BBTS. Basically you take 80% of a heavy compound lift and it helps you figure out you fiber make-up. From there you can better desgine your training to fit you muscle fiber type.
Have fun


I think you have a pretty good plan already. My advice would be to put a number of people through an 80% max rep test to help determine their dominant fiber type. i.e. if they do less than 8, put them in the fast twitch group, greater than 10 put them in the slow twitch group, or just weed out everyone but the fast twitches, and concentrate only on them. You could then do your test as planned, which is what I would suggest. By staying above 80% with your rep tests, you should see a linear correlation, where reps increase as %max decreases. You can just plug all your numbers in excel to get a best fit line, and it will give you a linear equation.

As I said, you could do this just for the fast twitch types, or you could come up with an equation for fast twitch type, slow twitch, and everybody in between.

If you wanted to estimate max based on a higher rep test, things would start to get difficult as it definitely would not be linear.

Be sure to keep in mind that you will get different results for different lifts.

Hope this all makes sense.