Saturday, 1 August 2015

On Oval Chainrings

On a forum to which I suscribe, one poster got all bright-eyed and bushy-tailed about oval chainrings, and one brand in particular. The poster assured us that they were the greatest thing since sliced bread and had clear, and clearly proven, benefits, and linked to a report by some suitably qualified engineers to prove it. the poster also assured us that they must be good because Team Sky (who have just clocked up ANOTHER Tour de France win!) use them, and as anyone knows, Team Sky are big on marginal gains (apparently, they even take their own beds with them so the riders sleep better!)

So I read the report.
I am not sure how to say this.
The report is based on a mathematical model, not actual cycling.
"So what?" you might ask.
Well, the model doesn't seem to cover the way I cycle.

1) I move my bottom on the saddle (apart from comfort issues, I move right to the back of the saddle when climbing - a common technique, and although I don't do it, many folks move to the front of the saddle for bits where they want to go faster)
2) I move my feet on the pedals (I have platform pedals, without toe-clips or stuff like that, so I can put my foot pretty much where I want!)
3) I flex my ankles (like many folks I lift my toes a bit at the top of a pedal stroke, and I "follow through" somewhat at the bottom - indeed, on steep climbs I "claw" the pedal back to help bring the other pedal to a better position for my other foot).
But these three things aren't in the mathematical model,
I'm guessing that another three independant variables would make the math really horrible, so they just simplified it!

Then I did quick a bit of research and found out that only a few Sky rider currently use oval chainrings, and the rest are round, and "Wiggo" has apparently stopped using them, and gone back to round chainrings (and is still winning!).
So while it is true to say that SOME pro riders use oval chainrings, some certainly choose not to.

Anyway, here is what I wrote on the forum in reply to the overly-enthusiatic poster and his odd-shaped chainrings:

A quick look at the analysis you linked to seems to suggest that it is a "model" exercise, rather than testing using actual cyclists.
It appears to omit at least 3 factors relevant to MY cycling style:
1) On any significant climbs (can be shorter and steeper, or can be longer and less steep, or can be both - i.e. long and steep!) I move my butt backwards, so I am right at the rear of the seat, This is a pretty common climbing technique. This changes the angles of the hip, knee, and ankle joints. No account of these changes appear in the model.
2) I use platform pedals, so I can (and do!) move my foot about on the pedal. On significant climbs I move my foot so I am pedalling just behind the ball of my foot, rather than the ball. This allows me to extend my leg just a touch further, and also, of course, slightly changes the angles at the hip, knee and foot, again, but not so much as moving my butt to the very rear of the saddle, It also allows me to use the stiffer part of the sole of my footwear - adidas sneakers.
3) There is no allowance for the fact that pedals can turn independently of the crank position - "pedalling in circles" does not seem to be represented! While I cannot lift the pedal (platform pedals, remember), I can certainly employ the toe-down and claw-back technique to extend the length of my pedal stroke. Similarly, I can tilt my foot back and apply power to the pedal at a point earlier than the model would suggest.

So while the model seems like an interesting attempt at an analysis, one must enquire whether the authors actually ride bikes on a regular basis! The three oversights I have mentioned above were no doubt omitted to prevent the math becoming terribly tedious - there are more independent variable in the action of a real cyclist than they model.
This undermines fatally the validity of their report - their report is valid if you cycle in a particularly stiff-legged manner, but many cyclists (including ALL the pros) just don't cycle like that!
Time for something like a GCN test, methinks - at least they use real cyclists who know how to actually cycle efficiently in the first place (they are all ex-pros!).
Maybe I read the report wrongly.
If so, can someone point out where the model allows for the three apparent omissions I mentioned above?

I might also mention that there is some debate (and evidence!) to question whether the authors of the report are right to so readily accept the earlier work they mention that suggests that the optimum cadence is in fact 90! Not sure how much that changes the figures, though (unlike the other three omissions!)

 How about what Team Sky actually say about oval chainrings?
"Tim Kerrison is the performance director at Team Sky, the sports science chief who oversees the training of the team, with particular attention paid to the ‘GC group’. Kerrison has been instrumental in the successes of Wiggins and Froome, both of whom have used Osymetric rings on their time trial and road bikes. Kerrison isn’t a mechanic — bike or bio — but he obviously pays close attention to his riders’ power outputs and equipment. What is his opinion of the rings?
“I’d say that, performance-wise, there is very little in it either way,” said Kerrison. “A few riders have a preference for the Osymetric rings, but many of our riders have tried them. Only a few continue to use them.
“That said, both Wiggins and Froome used them in the 2012 Tour, so they are unlikely to be significantly detrimental to performance.”
Wiggo used them in the 2012 Tdf, but went off them and apparently did his recent World Championship Time Trial (which he won!) on circular rings. (according to the same article)


If the benefits were so clear cut, and there was all that "free" extra power just waiting to be had, why would Wiggo not want it?

The article also says (and I think this is VERY relevant!): "Professional cyclists tend to have very efficient pedal strokes, with minimal dead-spots. Perhaps these rings may be more beneficial to the amateur or beginner, helping to iron out a more pronounced dead-spot."

So oval ring have clear advantages for folks who don't want to learn to pedal efficiently in the first place, apparently.
Probably good for folks with injuries and impared joint movement as they seem to give clear advantages if you pedal "stiff-legged"!

The summary of the article:
"The jury is out regarding the mechanical basis for non-round rings. However, from a psychological point of view, the case in their favour is more compelling. A placebo has the potential to offer a significant performance gain. That said, the disadvantages of lower-quality shifting and the potential chain drop are significant, especially in a race situation"

Update, 29th October 2015:
I came across Sheldon's article on oval chainrings again.
He makes a series of VERY valid points (as always) about why the oval chainring is not at all a new invention, indeed it has been around for about 100 years.
Sheldon also comments on why the "Biospace" oval chainrings are, in his opinion, better than the type that some sky riders use.
Sheldon says that the problem with the type where the gearing is higher half-way down the pedal stroke is that it is more likely to lead to knee injuries, both because of the inclination to pedal that bit harder on the section where the gearing is highest, AND because knees shoot past the section where the gearing is the lowest (and thus the legs will be moving faster at the point where the "upper" leg is the most bent).
Sheldon says that the Biospace rings even out the power delivery, and can be of use to riders on loose surfaces (such as MTB riders) as a consequence.
Part of the mechanism for that happening is that the legs move fastest during the "power" part of the stroke, and this momentum of the legs carries them through the "hard" part at the top and bottom.
This is in complete contradiction to the supporters of the Osymetric chainrings claim.

The Osymetric supporters point to their use by some of the Sky riders (apparently the rest tried them but didn't like them!) and some ill-founded research which uses an unrealistic pedalling model (I mentioned this above - I have read it, and it would just about pass for an undergraduate project, but nothing higher! - well-intention, but oversimplified to the point of obscuring the results)
The Biospace fans, like Sheldon, claim their design is based on "dynamic" pedal loads, as opposed to the motly static loads analysed by the "Osymetric" testers.

They can't both be right.

Interestingly, Sheldon suggests that folks fitting Biospace to recumbents should rotate their position forwards by one bolt-hole. as the most strident advocates I have found of Osymetric rings ride recumbents, if they are not making a similar rotation, then perhaps they are feeling the same benefits that they would get if they fitted Biospace rings the way that Sheldon recommends. Whether the "Osy" fans would feel the same benefits on "uprights" remains to be seen

As for me, this just throws more confusion into the mix, and my money is staying "round" :-)

Update 2, 29th October 2015:
I came across a piece which analyses the forces involved in pedalling, (in the journal of the Internation Federation of Sports Medicine, no less) and it is nothing like the model that the supporters of the oval chainrings suggest.
The model certainly is NOT just the classic half-sine wave follwed by a flat bit!
Read the report and see for yourself.
I would copy the picture, but it is from a copyrighted publication.
the two pictures on page 4 of the report tell all you need to know.

The relevance to me is that I tend to use a cadence of about 60 to 70 rpm on longer rides, switching to a higher cadence of about 80 rpm on hills.
This report would suggest that
  • I have already accidentally stumbled upon the very best way to pedal longer distances, given my modest power output, giving me both the apparently most efficient cadence from an energy expenditure point of view, combined with lowering muscle fatigue with my higher "hill" cadence. Of course, the fact that I have a VERY wide range of gears on my bike allows me to do this (my smallest gear is about 22 inches, and the biggest is about 118 inches, nore than a 5:1 spread!)
  • at slower cadences (such as the 60 to 70 rpm I mentioned above) there is a much bigger "tail" to the power generation with an efficient pedal stroke than there is at a cadence of 100+
  • the benefits of ANY sort of oval chainring are likely to be less for me because of my "ultra-endurance" cadence.
Again, my conclusion is that oval chainrings currently have nothing to offer me, and my money is best spent elsewhere!

No comments:

Post a Comment