Proposal for scientific comparison of bicycle hill-climbing performance

Jim Parker, MD, Cruzbike, Inc.
Many Cruzbike owners have made the observation that they climb faster on a Cruzbike vs. other rear wheel drive (RWD) recumbent bicycles. Clearly, there are complex biomechanics involving the upper and lower body when hill climbing on a regular road bike  (watch this video of Lance Armstrong climbing for an example). The Cruzbike pivoting-boom front wheel drive (PB-FWD) bicycle has some obvious similarities to the road bike when hill climbing: both make use of rhythmic pulling on the handlebars to angle/lean a compact tetrahedral drivetrain, and both take a slightly sinusoidal wheel path during maximum hill-climbing effort.  Based on these observations and testimonials from Cruzbike cyclists, we have reason to believe that the Cruzbike PB-FWD design climbs faster than a standard RWD recumbent. Now let’s look at what would be required to prove this hypothesis scientifically.
First, the research would need to be conducted by an independent researcher. Some people have called for Cruzbike, Inc., to conduct the research. However, if we did, and the hypothesis was confirmed, many people would consider the results biased, and disregard them. Secondly, a good research protocol needs to control for many variables, and have enough data points to have statistical power.  In short, a scientific, randomized, controlled study would require considerable planning and proper execution to have merit. However, the effort may be worthwhile.  Recumbent bicycles are acknowledged to offer many advantages over the standard safety bike, but they have a reputation for being slow when climbing hills. If it could be proven that the PB-FWD recumbent design climbed significantly faster than other recumbent designs, it could greatly widen the acceptance and popularity of recumbent bicycles.
Here is a suggested proposal for such a research project that attempts to control for all the important variables, including physical fitness, age, body weight, bicycle weight, and adaptation to the unique pedaling characteristics of each format. Other protocols may be equally or more efficient at testing the hypothesis.
Proposed Method:
Recruit a number of active adult cyclists from the same geographic area with no recumbent riding experience.  They number would be determined by the investigator to provide the needed statistical power to prove or disprove the hypothesis. The cyclists could be volunteers or paid a small fee for participating.  Basic demographic information of the riders would be collected (age, weight, height, gender, etc.) At the initial meeting of the recruited cyclists, they would each perform a time trial on their regular DF bike climbing up a reference hill with approximately a 5 to 10% grade and 1000 meter length.  Heart rate would be monitored. Next, randomly assign each cyclist to one of the test groups (i.e. PB-FWD or RWD recumbent).
Each cyclist would receive a 30 minute introduction to the bike, and would agree to train on the bike 6 to 10 hours per week for 12 weeks, keeping a log of their training. A research affiliate would check in by telephone or email weekly to make sure cyclists were putting in the required training.
At the end of the first six weeks, the cyclists would re-convene with their assigned recumbent bicycle and repeat the time trial, climbing the same hill, again monitoring heart rate.  The bikes would have approximately the same gear ratios and similar tire tread and air pressure. The weight of the bicycles would be equalized by adding weight to the lighter bicycles. Training would continue for another six weeks and the time trials would be repeated at 12 weeks, concluding the study. Standard statistical analysis would determine if there was a significant difference in climbing speed between the two test groups.
Key points here are 1) the randomization of the participants so that both groups would have the same probability of having participants of equal age, fitness, level, body weight, etc.; 2) by having participants with no recumbent experience, there will be no advantage from a longer adaptation period; 3) by equalizing bicycle weight, tires, and gearing, the only variable left is the actual bicycle drive system, and 4) by having one long hill-climb as the test course, speeds will be low and advantages from aerodynamic differences will be minimized. High-tech monitoring of power output and/or oxygen consumption is not needed to test the hypothesis and has been intentionally omitted to reduce the cost and complexity of the study.
Sponsorship of this research by more than one recumbent manufacturer is important to avoid the appearance of bias. If an independent University-based researcher would like to lead such a research project, Cruzbike, Inc., will cooperate and provide up to $10,000 in funds or bicycles if at least one other recumbent bicycle manufacturer does the same.
Any interested investigator or recumbent manufacturer should contact me at

Kenneth Jessett's picture

Interesting posit, however I

Interesting posit, however I doubt very much if the results would persuade more cyclists to take up recumbents or those that do ride recumbents to buy another machine albeit it a FWD. So, you might spend time and money proving a point which in reality would have little bearing on the recumbent market. Just my twopence worth. Smile

banerjek's picture

The proposed study strikes me

The proposed study strikes me as a bit complicated to execute, and I think power should be part of the equation if possible.

Seems like it would be easier to just get riders to switch a PT wheel between two bikes that they ride everywhere for a few months and log all the data on Strava (with a different account for each bike) or some other place. You'll get thousands of miles per rider fast and if there is any benefit, it will show up.

If you're looking for volunteers for a hill climbing study, count me in. My daily commute has over a thousand feet of climbing in just a few miles and I live near one of your dealers. I have other 'bents which I take in the hills as well as a couple DF's, so I could generate some decent hill climbing data pretty fast.

I recently took one of the Cruzbikes for a short spin.  As I was unexperienced with its handling characteristics, my sprints weren't quite working the way I wanted, but I had a blast once I got it up to speed. The bike is outrageously comfortable -- best I've ever tried (and I've been on a lot of bikes). I would look forward to an opportunity to become more acquainted with the Cruzebike.