I come from an Engineering background and I have a strong interest in science. It has always struck me as odd how so much of cycling is so very subjective.

As an example, though this is by no means the only thing, the best method of lubricating a chain. Now I appreciate that there is a difference between conditions that it has to operate in but from an engineering viewpoint there will be an optimum type of lubricant and methodology for applying it. Yet most of the advice you can read about it is laced with conjecture and opinion.

The entire industry appears to be built on arguments that should be resolved with simple repeatable testing.

Does this type of research simply not exist? or do we just enjoy arguing too much!

Edited to respond / clarify

As a species we have now sent a number of spacecraft out to the edge of the solar system. One of breakthroughs of New Horizon was getting it to leave earth orbit so quickly. I imagine the number of variables that were contemplated, researched and tested where considerably more complex than the handful that we can consider for a cycle chain.

The concept of testing isn't to account for every variable, it is to account for individual variables so eventually they can be merged together to get a complete understanding of the underlying system.

So imagine for a minute a box containing a cassette that can provide some constant friction. Then a powered front ring that turns at x rpm. Every y amount of time the gears are automatically changed.

At that point you can measure how much the chain is stretching, and whatever other criteria you're interested in.

You could run the same test with different lubricants and get an idea of what lubricant changes which result.

Once you have that data, you could have a mechanism that drops / throws / shoots z grams of dust / dirty / water / banana at the chain and run the test again.

At the end of a few chains you're going to be able to categorically say that for riding in mud, a is almost certainly better, for riding in high temperature, b is better.

It slightly staggers me that we can look at a photograph taken from billions of miles away and say that this is too complicated to figure out.

Further edited

It seems that somebody has already come up with that scenario, A guy named Jason Smith at Friction Facts spent $50k building a test lab to test chains. Velo Magazine had a report done on assorted lubes which articulates my point that facts are much better than anecdote. That gives a report that anyone can read and make an informed choice about what would best suit their situation.

My point in asking this question is that it surprises me that such data isn't used more in conversations about technology and that there isn't more high quality quantitative data about this thing that costs us so much bloody money!

  • 1
    There is testing but mostly it's not widely known. For your particular example, see here and here.
    – R. Chung
    Commented Jul 16, 2015 at 15:41
  • 1
    These test ignore two key factors: lubricant performance in wet conditions (like when it's raining?) and also collection of contamination in the lubricant. Pure reduction of friction is not the end all for a lubricant. If the lube collects debris and accelerates wear of the drive train, it's performance in a sterile environment is irrelevant. Commented Jul 16, 2015 at 16:03
  • 2
    Since when is engineering not similarly subjective? "Optimum" will depend on a variety of conditions that vary from person to person and situation to situation. Commented Jul 16, 2015 at 18:44
  • 1
    Testing of anything that humans interact with requires humans as well: car accident injury mitigation, effectiveness of medicines, or ergonomics of desk chairs, @whatsisname. There's an entire field of knowledge that deals with the uncertainty of having a limited number of error-prone measurements, and where it's not possible to use actual humans, proxies can be developed.
    – jscs
    Commented Jul 16, 2015 at 19:49
  • 6
    Testing requires motivation, and primarily that motivation is money. The money comes from one of two places: Competitive cycling and bicycle manufacturers. Both groups do testing, neither group has any motivation to disclose the results of said testing. Commented Jul 16, 2015 at 22:48

9 Answers 9


As far as outside testing goes, Friction Facts is an independent company that does exactly what you ask for, testing components against each other to find out what is best, including chains in a variety of conditions (new, re-lubed, wet, dirty, etc).

The top manufacturers presumably have unpublished data, considering that bikes continue to get faster and lighter. This research probably does contain blind spots, and I haven't seen a lot of manufacturer's recommendations about detailed advice on how to maintain a bike.

In addition to this, a quick google search for "research bicycle efficiency" turned up a significant amount of published research about cyclists, but not about bicycles.

I think the reason there are so many arguments on the web about the best way to do anything is that people just like arguing, although you could certainly argue otherwise.

  • 1
    It is the friction-facts site data that is exactly the sort of thing I was talking about actually. I'd never heard of those before. I will put a note about these in the question. I think the responses to this question have significantly opened my eyes to the issues that scientific / engineering data has when it meets up with the anecdotal evidence of religion!
    – Michael B
    Commented Jul 17, 2015 at 0:55
  • It's good that friction-facts has done a lot of research but data should be easy to access and widely available. It's funny that the reports should be purchased ! (friction-facts.com/test-results/individual-reports). More importantly, the data should be translated to writing that general people can understand. Without proper education/explanation, data is just data.
    – azer89
    Commented Jul 17, 2015 at 10:21
  • 2
    lol that last paragraph! Commented Jul 17, 2015 at 17:40
  • 3
    @azer89 - I agree, but who will pay for all this? Science is expensive and costs need to be recouped or future experiments cannot be run.
    – Rider_X
    Commented Jul 17, 2015 at 18:35

As a completely different answer, I am sure a lot of these type of tests are done by manufacturers. The information likely remains proprietary and never sees the light of day. The job of marketing is not to disseminate scientific facts, but to convince the public to purchase item X. The business model of cycling publications is to entertain, not run scientific journals. In the end a few dedicated souls try to do their own experiments, but this information is never largely disseminated. In the end the consumers never know the real truths, and relies largely on anecdotal evidence.

  • 1
    Some time ago I stumbled upon a great example: fietsersbond.nl/sites/default/files/test_schwalbe.pdf
    – ojs
    Commented Jul 16, 2015 at 18:59
  • You might be surprised, but when a lot of companies do "journal-article" caliber science, they usually will publish it. The problem is more often than not the tests and experiments aren't of that kind of setup to begin with. Commented Jul 16, 2015 at 19:04
  • @whatsisname Did you have some examples? I am doubtful they would publish it if it compromised their competitive advantage. What benefit would it be for Lube Y to publish an article saying Lube X works better than Y in wet conditions?
    – Rider_X
    Commented Jul 16, 2015 at 19:06
  • Apparently the test is done by Schwalbe itself, and rolling resistance numbers don't match at all with product descriptions on their website.
    – ojs
    Commented Jul 16, 2015 at 19:06
  • @ojs - And they were likely done on a smooth roller so we have no idea the true rolling resistance under more dynamic conditions (e.g., rougher roads where suspension/vibration losses become meaningful). That said it is consistent with my view that marketing and science have two completely different set of facts.
    – Rider_X
    Commented Jul 16, 2015 at 19:08

A lot of the research would be hard to do if not useless. For one thing, most bicycles just don't get used all that often -- plenty of people I know will likely never wear down any of the original parts on a bicycle even if they neglect the maintenance. Moreover, even among regular cyclists, we don't have enough people who would care about such a thing. So we don't have that many samples in the real world. We also have a huge amount of variation in parts and even more, a huge variation in riding conditions. See this article by Sheldon Brown on chains for example (particularly the first paragraph).

For example, if you wanted to see how long a tire lasts, most of the time its not that the tire wears out but you cut the tire due to debris. It also depends a lot on braking style (fixie riders who skid can burn through tires very quickly, for example) and other things so just running the tire on a machine wouldn't really be a good representation of how long a tire lasts.

Brakes have a similar thing -- if you ride in the wet vs dry (and type of wet vs type of dry), you'll get vastly different results. Same things for chains.

Also, cyclists have a wide range of what they deem acceptable. For example, some people will replace their chain when it hits 1% stretch according to manufacturer X's rules, while others will wait till it slips. Theres a wide variation in parts (esp. since you get the hot new thing every year with different rules) so building up averages over time is hard too (and boring for people to cover in press). See for example Heltonbiker's answer in this thread.

In your "optimal" scenario, you have to define what optimality is with respect to (i.e. the cost function) -- waxing a chain vs dripping some oil vs doing nothing can have vastly different costs making the optimal one different for everybody since everyone can have a different cost function. With different cost functions you'll get vastly different results. This is true in engineering as well.

  • I would like to emphasize the differences in riding conditions. It's hard to pick an "optimal" lubricant when people ride bikes in a 80F Hawaiian downpour or just as well in a -30F environment with next to no precipitation or humidity. No product is likely to be "best" under all conditions. Many products will outperform others under extreme conditions, but in the overlap where most people ride, there are advantages and disadvantages to everything that come down to rider choice. Commented Jul 16, 2015 at 16:16
  • 1
    @ChrisinAK You're right, optimal does mean different things to everyone, but with knowledge you can make a much more informed decision. Instead of the current practise of 'yeah we kinda suspect it works like this, but Joe next door thinks that's wrong, but he's an idiot anyway' (to paraphrase a recent conversation)
    – Michael B
    Commented Jul 16, 2015 at 17:54
  • 2
    Most consumers are not capable of making informed decisions; that's why there are marketing and advertising firms. People have learned to distrust studies and "facts" because they are usually skewed to whomever is paying for them. Studies (like the one in the first comment) are done under conditions unlike the real world and can produce irrelevant results. Commented Jul 16, 2015 at 18:04
  • The cost of getting that information of exceeds the value. If someone earns $20/hour, the best chain lube might make a 50% difference in chain life compared to whatever product they choose at random, saving them say $10. If they can't find and apply that knowledge in half an hour it's not worth trying.
    – Nuі
    Commented Jul 17, 2015 at 7:00
  • For people who ride a lot it's often the multiple needs they have - I need a very long-lasting "preservation lube" for my touring bike that gets used every year or two, something that extends chain life even if that means re-lubing every day for my commuter, and something in between for my other bikes. I could read the research and then try to find the products, and apply the right one to the right bike. Or I could just use whatever lube is readily available from my LBS and probably replace transmission parts more often.
    – Nuі
    Commented Jul 17, 2015 at 7:04

I think the other problem with trying to both do the research, and then present it, is that there are so many ways to define the "problem" and so many valid criteria for evaluating "best" or "optimal."

To use your chain example, and just off the top of my head, are we talking about:

  • Maximizing the life of the chain, or

  • Minimizing the time spent maintaining the chain, or

  • Improving shifting, or

  • Keeping the chain clean (so that I can ride it work in my white pants…).

I'm sure the list could go on and I was intentionally avoiding issues of different conditions. Plus, I'd bet that there is an element of loving the debate as well.

  • 2
    ...but all of these things are in themselves measurable. You don't ask an engineer "What's the optimum car transmission lubricant?", either, because the immediate response is "...for what circumstance?"
    – jscs
    Commented Jul 16, 2015 at 19:52
  • Agreed, but it leaves lots of room for discussion (or all out brawls) because there are so many measurable things and so many different ways to measure them – your quick method for maintaining the chain is my idea of purgatory, so I don't care that it's a bit shorter and makes the chain last a little longer than the method I prefer (or maybe it costs a bit more in materials or spousal tolerance or …). Fortunately we seem to relish the debate.
    – dlu
    Commented Jul 16, 2015 at 22:20

It's exactly the same thing in the automotive world. There are all kinds of studies showing you should do this and you should do that, but the vast majority of people just don't care. For example, I have no idea what kind of oil is in my car. It could be 10w30, it could be 10w40. It could be real or it could be synthetic. All I know is that there is oil in there. I trust that the guy doing my oil changes knows what he's doing, and as far as I can tell, he's been right so far.

I ride a lot, and on all my bikes, I've used all kinds of lubricants. What I've discovered in the last 20 years is that I can't tell any difference. Those times when I've pulled off a 5-hour century, it wasn't because I had the right oil on my chain. When I set a new PR up my local trail, it is not likely due to switching from dry lube to wet lube. If there really was one lube that was significantly better than the others, then I'm sure we would all know about it. The fact that every manufacturer seems to make that same claim is a good hint that it probably isn't true.

I'm sure I'm representative of most people who not only don't care too much about these things, but also are certain that they don't matter. The truth is, I don't lube my bikes enough, so any lube on the chain is going to be better than no lube.


It all comes down to tolerances and cost/benefit.

On a highly engineered vehicle or machine, parts are subjected to tremendous forces, very near to the maximum point of resistance a piece can withstand. Tolerances are very tight in many aspects. Forces, temperatures, etc.

For those machines, cost of replacement of parts, or repair due to damage from neglected maintenance (or catastrophic failure, for that matters) is high enough to justify the needed research to determine optimum procedures and supplies (lubricants, cleaners, catalysts, specific tools).

But in bicycles, which are a fairly simple mechanism, many of the components are far from material and design limits, specially for non competitive riders, there are very loose tolerances. So, the whole industry doesn't need to be that specific about procedures, the cost/benefit relationship simply does not justify the research.

On the other hand, the simplicity of the bicycle, and its ubiquitous-ness makes it both possible and necessary that there are so many alternatives for maintenance. I mean, there is the top competitive rider in a fully sponsored team, and the humble commuter who lives in a sub-developed region, far away from any bike shop or even a hardware store or gas station.

From that follows that there exist the specific, "optimum" lubricant for an 4 hour race stage, but there is also a rider who has been lubricating the same chain with, grease from some animal for ten years (Exaggerating just for the sake of the argument). And they both have used the best that they could and they both may swear that it works. From that point on, all "objectivity" decays, because you'll find riders all over a spectrum not even limited by this two examples.

But this is not a phenomenon exclusive to cycling. Consider for example how many debates exist in automotive industry. Compare how tight standards are in motoring respect to cycling (in general). Now compare aviation to motoring. Space transport to aviation...

  • Tolerances are not necessarily tight on "highly engineered" machines, whether a particular element's tolerance is tight or loose is an engineering decision like any other. In many situations loose tolerances is the better choice. Commented Jul 16, 2015 at 19:26
  • "lubricating the same chain with grease from some animal for ten years" -- and the animal is getting pretty annoyed by it. Commented Jul 16, 2015 at 21:01

A bicycle without a rider is not a machine, it is just a large paperweight or a piece of artwork. There are very few technical aspects that can be "objective" when the very machine itself has a greatly varying and subjective element inherent to it.

Many of the other answers have touched on this issue, and that's that all bikes are not used exactly the same way.

In one of the OP's comments, is written "If you want variable loads, then create a mechanism on the drive / cassette to create that in a repeatable way.". Well, that's not so simple, because not all bikes see the same loads. A bike ridden by a professional rider is going to see different wear than a bike ridden by a guy in a minnesota winter commuting to work. The moment you build a machine to test your cassetes, you've already decided what "optimum" means, and that won't cover everyone. Trying to test every method with every substance with every riding condition with every maintenance schedule to figure out what's best in all circumstances is completely impractical.

Jahaziel gave the example of someone lubing his chain with animal fat. That very well could truly be optimal for that person's situation. How can optimal involve something that is unavailable or too expensive, for anyone, for anything?

The New Horizons probe was designed to perform a very specific function under very specific conditions in a very specific timeframe. In contrast, part of the beauty of the bicycle, is that it can be used for so many different purposes.

  • 3
    I don't know a lot about rocket science (apart from what I've learnt from Kerbal!) but I imagine the challenges of throwing half a ton of scientific equipment 3 billion miles produces a lot more unknown variables than the difference between the TDF and a commute to work.My point is simply that more data is always a good thing, and a lot of this wouldn't take a lot of work to get more data, but instead we say its way more complicated and variable than rocket science, so we won't bother trying to understand it
    – Michael B
    Commented Jul 16, 2015 at 19:44
  • "A [car driven] by a professional [driver] is going to see different wear than a [car driven] by a guy in a minnesota winter commuting to work. The moment you build a machine to test your [transmission], you've already decided what "optimum" means, and that won't cover everyone. Trying to test every method with every substance with every [driving] condition with every maintenance schedule to figure out what's best in all circumstances is completely impractical." ...and yet much relevant information exists for cars. The question is, what explains the difference?
    – jscs
    Commented Jul 16, 2015 at 20:21
  • This has a lot of truth to it. Maybe using a chain with a pitch of .75 inches vs 1 inches would make things more efficient, but the advantages of using the same chain pitch throughout the entire industry is more important than some small increase in efficiency. Bicycles are really great in the fact that so many of the parts are interchangeable. Compare that to cars where every parts seems to be specific to each model of car. The vast majority of people would rather have interchangeable parts than have parts that are 5% more efficient but are expensive and difficult to replace.
    – Kibbee
    Commented Jul 16, 2015 at 20:28
  • @MichaelB: The laws of physics don't change from day to day. The behavior of people in deciding how to treat their bicycles does. Commented Jul 16, 2015 at 21:28
  • 2
    @JoshCaswell There isn't a difference. Bikes list out weigh specs and have performance stats just like cars. The car industry also has all the same questions. What motor oil is best? The usual answer is whatever weight is recommended by your vehicle manufacturer supplied by whatever brand is sponsoring your favorite NASCAR driver. Commented Jul 16, 2015 at 21:34

The big thing is : you have a very subjective human riding the bicycle.

Even if a study shows that for X leg length you need Y crank length, if the person riding the bike think that Z cranks feel better or feel faster, what can you do? That is mostly why there is so much debate, because everybody is different.

Also people want different things from their bike, some want minimum maintenance, while other want minimum friction. Some people want light wheels, other want stiffer wheels. Some people don't mind lubing the chain after each ride, while other will do it one a season.


To answer your question of why there are so many subjective discussions which are not backed by numbers:

Because there is no accepted standard

I am specifically not saying that there could not be such a standard (of course there could be), or how relevant it is, I am just mentioning that there is none.

I believe the situation can easily be compared to car efficiency discussions.

  1. There are a lot of different circumstances
  2. It depends on the user and usage
  3. There are many different objectives

Now, what happened in the car industry (at least in some countries) is that manufacturers are forced to report their theoretical mileage in a standard way (A certain test track is used, simulating several conditions with reliable drivers).

This greatly reduced the amount of discussion for the particular topic, but of course it is a lot of work and it only solves the issue for one, or a few topics.

It would of course be possible to also set such a standard for each individual aspect of bikes, but there is simply insufficiant incentive to warrant the cost and effort. So, unless there are big changes (bikes become tax deductible if they are expected to last more than 5 years), we will probably not see much development in this area any time soon.

Not the answer you're looking for? Browse other questions tagged or ask your own question.