Is Lemond's saddle height method outdated?

Question

Do you think Lemond's saddle height measurement method (which is taking the 0.883 of inseam measurement socked feet and use that to set the distance between the BB center and the top of the saddle) is outdated considering the changes on pedals, cleats and varying crank arm lengths? If so, how can we adapt this method to today's standarts?

Answer 1

(The Lemond method is to measure inseam by 0.883 in socked feet and use that to set the distance between the BB center and the top of the saddle).

The Lemond method gets into the ballpark of correct as successfully as lots of other methods. If you're comfortable with the measuring and the math, it makes it easy to get to a reasonable starting height quickly, by yourself, and without a lot of knowledge about bikes. That's always been true and hasn't changed.

It has never had any validity whatsoever as an absolute predictor of optimal final height. There are a lot of reasons for this, including:

1. It doesn't account for crank length
2. Pedals and shoes have different amounts of offset
3. Humans have different amounts of joint mobility
4. Saddles flex different amounts

As with many rule-of-thumb discussions in cycling, if you pretend everyone is a stereotypical road racer, the world gets simpler. Here the Lemond method probably hits more often. Even then, if you only take away the assumption that everyone is on road shoes/pedals (which are relatively homogenous in offset), things still fall apart because of the higher variance in offset between other kinds of pedal systems alone.

If the question were has it gotten more or less useful over time, the answer is less, because today there is more variance in crank length and pedal offset.

Answer 2

Consider the formula of 220 - age for maximum heart rate.

This formula is typically presented as a formula for your maximum heart rate. It categorically is not. It is an estimate of the average maximum heart rate for a given age. You may have encountered alternate formulas, like 207 - 0.7 * age, which are claimed to be better - but even so, they're just refined estimates of the average max HR for a given age. Many biological characteristics have a lot of variation, and max HR is one of them. The formulae are produced by a statistical technique called regression¹.

What on earth does this have to do with saddle height? Well, any numerical rule like this would have to be a rule of thumb. It should never have been presented as a reductive rule. I don't know exactly how the rule was derived. Perhaps Lemond observed that it was close to how most of his colleagues, who were professional road cyclists under age 40 and of European descent, set their saddle heights. We don't know how consistently he measured saddle height. We don't know the quality of records that Lemond kept. So, even if that rule was empirically derived, there are issues of measurement error and generalizability at minimum.

I would use rules like this as a starting point for your saddle height. After that, you need to see how it works in practice, and you should consider iteratively setting your saddle height based on observation. In this answer, I overviewed some empirical tests that bike fitters use for saddle height and saddle fore-aft position. Those tests can potentially be done by laypeople, although some of them need you to enlist a friend (the ones with smart trainers being especially useful). If the explanation doesn't make sense to you, then you probably want to seek a professional bike fitter.

In sum, regardless of how dated Lemond's formula is, it shouldn't have been used to prescribe an exact saddle height. If it was presented as an exact formula, that was an error.

How was that formula derived?

Reportedly, there is at least some explanation in Lemond's Complete Book of Bicycling, which I don't have access to. This blog post says that the research behind the formula was actually done by a Dr. Ginet and by former French pro racer Cyrille Guimard.

It was the product of research from Dr. Ginet, and supervised by Cyrille Guimard, who determined the ideal leg extension for maximum efficiency and power output while cycling, and made famous by the legendary cyclist, and 3 time Tour de France rider, Greg Lemond. Basically the ideal optimal saddle height is approximately 96% of full leg extension. The chart numbers come from a calculation of the inseam multiplied by .883 and then measured along the the angle of the seat tube from the center of the bottom bracket to the top of the saddle and take into consideration average shoe and cleat thicknesses.

If the account is accurate (which means that the book described the process accurately and that the blogger summarized it accurately, keeping in mind that neither of them are scientists), then it sounds like Ginet did an empirical study. It was later used by Guimard and Lemond. However, that study would have been done in the late 1970s or 80s. The human body is complex. Our knowledge of kinematics was definitely less advanced then, and there is likely still a lot we don't know about the body. For example, there is still no clear, definitive, empirically-based rule for setting crank length.

Therefore, the rule might be simultaneously dated and still informative (but not absolute).

---

Footnote 1: The coefficients in that formula and probably most alternate formulae are rounded off so that they can be remembered more easily. This article tried to track down the original source for 220 - age. It's probably a 1971 article by Fox and Haskell, which showed a scatterplot and had this quote:

“....no single line will adequately represent the data on the apparent decline of maximal heart rate with age. The formula maximum heart rate=220–age in years defines a line not far from many of the data points..”

This study measured the max HR for a sample of people and plotted it versus age. They may have fit a linear regression model and then rounded off the coefficients for easier recall. When the parent article's authors approximated the datapoints from visual estimation and fit a linear model, they got predicted HRmax = 215.4 - (0.9147 * Age). For those familiar with regression, the intercept is 215.4 and the slope for age is -0.9147. By comparison, the 220 - age formula sets the intercept at 220 and the slope at -1.

Linear regression may not always be explicitly described as producing a conditional mean (in this case, that means the average max HR given age) in learning materials, but it is that. For example, go to the Wikipedia description and search for "conditional mean". Thus, the formula should always have been interpreted as an estimate for the average max HR for a given age. You can calculate a prediction error for any linear model fit to any dataset, and the paper says the prediction error was over 10 bpm in that study. However, once the formula got into practical use by lay users, misinterpretation was inevitable.