Amusingly, the Giant rep concludes the stiffness section by saying, "Beyond a certain point, the stiffness doesn't matter. The typical ride will never flex the bike enough to matter."

So now, the question is that actually an engineering judgement or marketing speak?

I understand that. Which is why I also did a seated sprint effort and saw the same bb flex. I can flex the BB left to right about a inch at will

Not quite right. Accuracy is a measure of average distance from the the bullseye. Precision is a measure of the data scatter (the number of significant figures). If you are trying to do an efficiency calculation, accuracy is of prime importance. With 2% uncertainty and, say, losses of 4% you are trying to measure, you can get anywhere in the range from 2% to 6% for your loss measurement (50% uncertainty). You would require a lot of calibration to make this better, to the point where, if you wrote a paper on the subject, the calibration method would take up more than half of your paper and 3/4 of your experimental time.

_________________

As far as the other comments; there is data, there just isn't numbers, so the picture is a bit cloudy. I stand by the crowd sourced answer. Every frame material which makes the bike lighter (the move to aluminum, titanium, and carbon fiber), did not "catch on" until the stiffness increased to match the prior state of the art. Aluminum frames didn't catch on until tubes became oversized and the joints welded. Carbon didn't catch on until the carbon composite technology led to the required stiffness.

Engineers frequently have to make do with limited or untrustworthy data. This is why we spend 4 years in school learning all that math and basic analysis. Just because something can't be measured doesn't mean you can't make a first order approximation of the solution. You would be surprised just what little information engineers deal with. Testing is expensive. It typically isn't done unless the information is critical.

As a marker, for a 1000W sprint at 90rpm, a deflection of 1.7mm (noted above) is something like a joule per pedal stroke, which is 2 joules per rotation, or about 3 watts. First order approximation, of course. Deflections, even of the very small amount, matters in a bike drivetrain because our legs are very torque driven. And this is a relatively stiff frame of modern design; I've seen movement in the bottom brackets of certain bikes that well exceeded 1.7mm.

In an effort to keep the thread going past 7 pages, I will pose this question. On descents, why would stiffer necessarily be better? At some point, too stiff might cause the bike to lose contact with the road surface on bumpy terrain. Would you want to be descending in a car with no suspension relying on your tires to soak up the bumps in the road? You will lose contact with the surface as you bounce over the bumps.

Both. In engineering terms, you have a bunch of things sucking power out of the load path (shoe, pedal interface, bearings, chain, derailleur, wheel, tire, etc.). Bottom bracket stiffness might dominate until you beat on it enough to where something else dominates, at which point, the focus is off of bottom bracket stiffness until you push on that "something else".

In marketing terms, it might be a bit of a move to blunt another manufacturer who is making a big deal about their frame stiffness.

Ding ding ding. It's not. If you are touring (or in a long road race) stiffer is not necessarily better. As you mentioned, control is an issue over bumpy terrain, and comfort becomes an issue as well, as comfort directly affects your power output at a certain point of discomfort.

Not sure about others but as an old phart of 63, I am enjoying this thread. I will NEVER-EVER have to worry about flex here, there or where-ever but thanks for the civility. All I do know is that my Propel is WAY MORE than I need but I really enjoy it and in the long run, that's all that matters for this geezer.

As long as the bike tracks correctly, you actually want something that can take small bumps with a small bit of deflection

This is actually what the Domane/Roubiax frames address. I suspect that in general you want a very high lateral stiffness, while allowing some vertical compliance to improve ride quality. Modern frame designs can achieve both by shaping stays etc.

So Brian, isn't Big Corp Bike also interested in the question of diminishing returns? It seems to me they wouldn't want to spend all that R&D money to find the next level of stiffness solution (stiffer, lighter, just as or more comfortable) if the "crowd" that we are all relying on for sourcing guidance isn't going to see the value. Short of producing 100 prototypes and giving them to pro teams, how can they validate that the enhanced properties increments will be of value this time as they have been in the past. Are you saying Big Bike Corp really doesn't know the relationship between stiffness and power transmission. I am very skeptical of that.

^^^
Your fallacy is in thinking there is such thing as "big bike corp". In reality, there are a bunch of profit driven companies looking to push products on a small niche market. If company A won't make a stiff frame, and people want it, they'll go to company B. The way of the world.

Pros don't test much. They ride what their sponsors give them, period.

All modern frames are stiff enough and pro teams don't make their decision on which bike to ride based on performance or 'stiffness' measurements.

Oh, that's why quite a few riders request custom bikes with more stiffness! Interesting. Didn't know that Cervelo Test Team didn't do any testing for Cervelo. That's good to know

They also request custom paint jobs. Doesn't make them any faster though.

Of course they test things. You spend a gazillion hours on bikes and you'll get a good idea of what you like. You'd have to be willfully blind if you think bike companies don't take advantage of this with their sponsored riders.

Trek, Specialized, and Giant are certainly significiant size businesses with more market than just high end roadies. I am just saying that if more stiffness isn't going to provide additional benefit, do they want to design and build it? By the way that is a trick question. Answer no, and you have to wonder, well how are they going to know when to stop stiffening. Do they have the data they need, and why can't I see it? Answer yes, and the whole rationale for buying the latest, stiffest frame falls apart, because the maker is chumping the public.

Depends on your definition of 'test'. I would be very surprised if any pro or team had done an objective measurement quantifying performance differences between different frames with different stiffness. Most of their testing would be of a subjective nature.

First, like bike weight, there is no limit to how stiff a frame can be (for power transmisson) if you are marketing to the performance crowd. There is no local maximum for stiffness. The perfect bike frame would be one which is perfectly stiff. More stiffness always leads to better power transmission; a small advantage is an advantage, which then means you are just dealing with cost/benefit.

A special note about data and why you can't see it... The general public only has access to a very small fraction of the research going on the world. Research done by a business is commonly referred to as "intellectual property" and is every bit as, and sometime more important, than gaining patents. IP is closely guarded and only released to the public and their competitors if there is something to gain from it. Most research done by universities is regarded by industry as a joke, btw.

Of course. Don't be so quick to discard the subjective opinions of cyclists who spend a better part of their life on a bike. Don't be so quick to discard the crowd sourced answer wrought from collective billions of hours on bikes.

Most information we have about the world is of a subjective nature. Objective observations, much less designed experiments, are a very small portion, not to mention that any time you try to design a test or design the measurements for observations, you are adding your subjective input to the problem.

Cannondale made a big leap (forward?) in stiffness with their oversize aluminium frames. Anyone notice the pros of the era rushing to get similar frames because they were now at a disadvantage?

According to you. Not settled for some of us.

I've done a lot of both, corporate and university research. Corporations can think what they want, but in most cases I have experienced, their work is crap compared to universities. The methodology is highly perverted by the requirements of the business.

Yes, of course you are right about intellectual property and its secrecy. But just as you say that suggests that the information would not be helpful in the marketing effort, else it would be disclosed. Going right back to my original post, I want to know why.

Okay, no local maximum for stiffness, but surely there is a point where the returns diminish. Perhaps a TdF rider isn't interested in that, but it is exactly what the amateur wants, aka bang for the buck or as you say, cost/benefit.

Look, you are representing the status quo and suggesting since it is the status quo it must be optimum. I say BS to that. You're right about the status quo, but not about what is best for us amateur riders. Let's know how this stiffness/power transmission really works, and then we can decide for ourselves what meets our needs.

Chuckle. Some people say the converse.

Will this post get us to Page 7? To get back to the original question - you most likely will never see numbers providing a direct correlation between stiffness and efficiency - it just can't be proven, as shown by 6 pages of comments here, with nobody coming forward with anything but opinons. I happen to have a bet with an acquaintance who works for a large aerospace firm that he can't provide proof of the stiffness/efficiency debate. 6 years and I'm still waiting for his proof.

I, too, would like to see some hard numbers... And is it too much to ask that we have them for both clipless and platforms?