I am perfectly happy with my crank length, why should I think I could riding somethng much better?
Bicycle riders come in sizes ranging from under 5 ft tall (60 in, 152 cm) to 6 ft 6 in tall (78 in, 198 cm), a variation of over 30% yet bicycle cranks range in length from 165 mm to 175 mm, a range of 6%. Bicycle frames are readily available in 18 sizes but bicycle cranks on original equipment generally come in 3 sizes, 170, 172.5, and 175. As far as bicycle cranks go the industry may as well be taking a one size fits all approach. It serves the bicycle industry just fine because it is much less expensive to have just a couple of sizes but it doesn't make sense this would serve the athlete well. If you find that you:
1. are having trouble getting into an excellent aerodynamic position (low up front) for racing…
2. find your knee being thrown out when the pedal is at the top of the stroke…
3. have a lot of knee discomfort when riding…
4. have trouble getting your cadence up as high as you need…
5. have trouble getting your cadence up when sprinting…
then you are probably riding a wrong crank size for you and almost certainly your crank length now is longer than it should be. Be sure to read more to better understand why crank length is the ignored aspect of bike fit, that if you pay attention to it and optimize it for you then it will pay big, big dividends.
When it comes to racing, does crank length really matter?
The first thing you are probably asking is why, after 150 years of racing bicycles during which crank length has evolved to what we all know of as "best", am I asking this question? Haven't the current crank lengths people use stood the test of time? Well, first, we should look at how current crank length evolved. 
Bicycle racing has been around for about 150 years and it started on the "ordinary" bicycle, the bicycle with the big front wheel. The bigger the wheel the faster the rider could go and the longer the cranks the more leverage he had and the faster he could accelerate. Crank length evolved to what it is now by the experience of these racers on these machines in the kind of racing done at the time. Cranks of about 170 mm in length proved to be about optimum, bigger for tall riders, shorter for small riders. Later, after the safety bicycle was developed (bicycles with two equally sized wheels one driven by a chain) cranks could be of any length but these were still single speed bicycles and behaved very much like the ordinary bicycle so riders stayed with what they were used to, because people, by and large, do not like change (plus, it seems tradition really counts in cycling). And, manufacturers made bikes that gave people what they wanted, so every new rider became comfortable with this crank length and for 100 years this crank length is what everyone became used to and liked. Since the best cyclists used the same crank length, everyone thought it would be best for them also. Then, around 1950 the derailleur was developed and riders now could change gearing on the fly to optimize gearing for conditions. Gearing is nothing but changing the leverage between the pedal and the ground, now no longer is crank length important for "leverage." But, not many were thinking of aerodynamics and since power doesn't vary much with crank length, it wasn't obvious that crank length needed to be revisited, so racers stayed with crank lengths that were familiar and comfortable. And, of course, manufacturers saw no reason to change either as they prefer to sell what everyone already wants rather than try to convince people there is something better. We like what we are used to, nobody likes unnecessary change.
But, back to the original question, Does crank length really matter. Well, if I didn't think so, I wouldn't be writing this. The only real question is how much does it matter and in which situations does it become most important. Even Lance Armstrong (after he started racing triathlon and before he got banned by the world) stated he had moved to shorter cranks for many of the reasons stated in this essay. Lance, if he was about anything it was about taking any advantage he could if it would help him win. It appears to me that going to shorter cranks (sometimes much shorter cranks) is simply finding free speed, speed that is sitting within you waiting to be discovered. How did I arrive at this determination?
I got to thinking about this a couple of years ago when I began to think that it would be much easier to get a cyclist into a good aerodynamic position if the cranks were shorter but I didn't know what would happen to the power as one went shorter. I asked the question on the internet to see what "science" had been done in this area and was directed to a study done by Jim Martin at the University of Utah (Determinants of maximal cycling power: crank length, pedaling rate and pedal speed).
The Martin Study
The Martin study looked at the effect of crank length on max power in cyclists. This study concluded that while the power was maximum with a crank length of 145 mm there was little lost by most riders when using 170 mm cranks because the difference was only about 1%. Below is the most relevant figure from that study:
It is strange to me that even though the difference in power between the 145 and 170 mm cranks did not reach statistical significance Martin would have concluded that crank length didn't make any difference to the racer. Failing to reach statistical significance may only mean the study was not powerful enough to pick up a real but small difference. There is a clear trend in the data and there is a big difference between racing significance and "statistical significance". It also seemed obvious that if there had been more participants that the 145 mm cranks superiority would have eventually reached statistical significance. At a minimum the authors should have concluded this potential existed and this area needed additional study. But, they did not and "everyone" now believes, based upon this one study and its flawed conclusions, that it has been proven that crank length makes "no difference" in bicycle racing. Nothing is further from the truth. If the differences were real but small why would a racer want to give up any additional power?
Even if their was no power improvement, how does one explain the fact the power didn't drop? Everyone knows that longer cranks offer more leverage so should offer mover power, correct? Yes, until one considers that the total power during the pedal stroke involes the sum of the powers around the entire pedal stroke. There may be more leverage pushing but what happens around the rest of the stroke. It just so happens I have some data that looks at that. A coach in Italy that just couldn't believe that shorter cranks could offer an advantage. He put himself on an Excalibur ergometer, which measures pedal forces around the entire circle, and compared what happened between 170mm and 150 mm crank length, close to what Martin did. Here are his results (the 150 crank length results are on the left, the 170 crank length results on the right).
Note, the total average power for this 90 second test was slightly greater (313 vs 308) on the 150mm cranks despite the fact that the average maximum power was greater (633, 646 vs 610, 620) on the 170 mm cranks. This is hard to explain until one looks at what happens on the upstroke where the negative forces are much less (-73, -145 vs -96, -166) on the 150 mm cranks. It is the ability to gather this kind of data that makes having a 2nd generation power meter worth the cost to the serious athlete. If you would like one of these for your own training you need to read more about the iCranks and all the wonderful things it will be able to do.
Beyond Martin
Once I saw the Martin data it got me to thinking. Why would any serious racer want to give up any power, even if it were small, just so they could ride the cranks that came with their bike? And, even if there were no power advantage, the Martin study only looked at power production and ignored the effect of crank length on bike fit and aerodynamics, a huge factor in determining how fast you can make your bike go. This bike fit issue was really the original problem that got me thinking about this. This got us to doing some of our own experimenting and our early data shows that for ordinary sized men, when in the aero position, that power does not start to drop until crank length is below 130mm crank length. After experimenting with this Pro triathlete Courtney Ogden won Ironman Western Australia racing on 145 mm cranks. Another recent experiment by an Italian rider with access to an Excaliber Ergometer (one that measures crank torque around the entire pedal circle for each crank) showed similar results to the Martin study but helps us to explain why (see above). See the data averages for the two different lengths.
I have experimented with cranks as short as 85 mm but have found a crank length between 130 and 145 is probably optimal for me. I am 6'2" tall with a 34" inseam and as a result of going to very short cranks have been able to drop the front end of my bike more than 5 inches without any comfort issues (the main reason aerodynamics will improve with this change). Based upon our early data and theoretical considerations (see below) I believe the following axioms will be true:
1. Everything else being equal, a rider putting out 150 watts will test " shorter" than a rider putting out 400 watts. (This means beginners will test shorter than pros, Endurance athletes will test shorter than sprinters). See illustration of these basic relationships I believe to be true even though the numbers for any particular person will be somewhat different than shown, why you need to experiment.
2. Everything else being equal, a rider with less flexibility will gain more aerodynamic benefit going short than the circus contortionist.
3. Everything else being equal, a rider trained to pedal in the PowerCranks fashion will be able to go shorter than another rider who gets all their power from "pushing" alone. (more about this below).
4. Those with significant knee issues should do better with short cranks than long cranks - even though short 
cranks may require the rider to push harder, the knee is bent less so overall stress should be lowered on the knee.
5. The major racing benefits to come from going to shorter cranks are to be found in the aerodynamic improvements that are possible and improved bike comfort. While power might be optimized also, the power variation seen with crank length is small (10-20 watts), as shown by Martin and our own data.
But, the aero benefits seem to be huge such that this change should provide huge dividends to the serious cyclist. The picture shows a super imposed photo showing how pro triathlete Courtney Ogden was able to improve his aerodynamic position when moving from 172.5mm cranks to 115 mm cranks. Not only did his power improve (slightly) as he went shorter but look at the improved aerodynamics. He would be faster even if his power dropped a bit with this change. But, what is best for Courtney may not be best for you. Another rider willing to push this envelope reported this:
"I went for a ride yesterday for 75 min @ 120mm lengths then did a 15 min run. It felt really good - I averaged 90 rpm and did a 5 min time trial in the middle @ 300W @ 29.7 mph avg. Compare to a TT in (a few weeks earlier)w/170mm cranks @ 280W @ 27.7 mph avg., first time on the tri bike since July it was 7% faster @ 7% more watts so aerodynamics seem to be improved. (ed: a 7% increase in speed would normally require about a 20% increase in power) I'll continue to play around with cadence and power but it felt pretty good yesterday. It feels a lot more aero with an extra 10cm drop to the bars and more comfortable than the old setting too. I need to measure the hip angle to check position. Sitting forward on the seat felt more powerful so I will test moving the seat forward. I'm at about 23cm drop from the seat to the pads. I think I'll stick with 120mm cranks for now because I don't think my neck could take any more drop and I'm concerned about bike handling if I go to 100mm crank and move the seat even more forward."
What is optimum for each person can only be determined by trial and error experimentation.
But, what about climbing?
Everyone knows that to climb well you need the leverage that a longer crank offers. So, what happens when you try to climb with short cranks. First, it is simply a myth that you need long cranks for the leverage. Well, the leverage that is important is not the leverage between the pedal and the bottom bracket axle but the leverage between the pedal and the rear wheel. Since there are several links in the leverage chain it is possible to adjust the leverage using gearing to make up for "loss" of crank arm leverage. And, there is one more lever in the chain that people generally ignore, the knee. The more you bend the knee the more "leverage" you lose (can you press more weight with a full squat or half squat?) Climbing is about power so you should have the crank length that optimizes power, not a crank that optimizes one aspect of the overall leverage equation.
To illustrate this here is a real world example. Drew Peterson is an experienced ultra-cyclist. He had done the Everest Challenge for many years. He believed he needed long cranks to perform optimally in a race that involved 28,000 ft of climbing in 208 miles. His times were typically just over 12 hours for this effort on 180+ cranks. In 2010 he decided to do the race on PowerCranks, using the longest length available to him, 182.5mm and he cut about 30 minutes off his usual time, doing 11'47". This effort convinced him that racing on PowerCranks was the way to go for him. I then convinced him to experiment with shorter cranks. He gave it a try and kept pushing the limits and decided to race the 2011 edition on 110mm PowerCranks (he also used arch cleats because he felt they were more energy efficient, stressing the calves less, and allowed him to fit on the bike better). In 2011 he improved considerably using 110 mm PowerCranks he improved his placing from 26th to 9th, beating most of the pro/1/2 riders in the race. The course was slightly shortened this year but we think his time improved about 30 minutes compared to most others who did both races. So, yes you can climb just fine on very short cranks if you are geared properly and train yourself to do so.
More benefits - Moutainbikers
So, using shorter cranks will usually give the rider: 1. More (or, the same) power. 2. More ground clearance. 3. Better aerodynamics. What is the mountainbiker not to like?
More benefits - Good for Bad Knees
And, did I say that short cranks should be good for bad knees? For any given muscle force, the stress on the knee goes up as the knee bends more. The stress increases substantially as the knee bends beyond 60º. Therefore, if you have bad knees one way of minimizing the stress on them is to shorten your cranks. Here is a link to a web site that lets you explore exactly what happens to the different joint motions as you change crank length and saddle height. Another part of this site has a very interesting discussion of bicycle power production. Despite it being primarily associated with recumbent pedaling, the general principles apply to all pedaling.
WHY THIS WORKS
How can this short crank length benefit be explained scientifically? There are two major scientific papers here. First is Determinants of maximal cycling power: crank length, pedaling rate and pedal speed by Martin and Spirduso of the University of Utah. This paper concluded that while the power was maximum with a crank length of 145 mm there was little lost by most riders when using 170 mm cranks because the difference was only about 1% and the difference was not statistically significant. Further, the paper concluded that the most important variable that affected power was pedal speed, not crank length.
The most relevant figure from that paper as regards this discussion is again reproduced below.
The trend is clear (in both the average power for each crank length and the power range for each crank) but the differences didn't reach statistical significance (a mathematical analysis by which the research can say with certainty that the results have less than a 1 in 20 chance of being due to chance). As a result of this failure to reach statistical significance most observers have concluded that crank length is of no consequence. But, let's look at this "failure" more carefully. There are only two possible reasons this data did not show statistical significant differences. 1. There are no differences or 2. The sample was too small to show real differences to the degree required by scientific papers. The reader must look at the data and decide for himself, which is most likely? To my reading, there is a clear trend and, to my reading, the only reason statisical significance was not reached was because the sample size was too small. From this data alone it is reasonable to infer that if there had been enough subjects in this study the 145 mm crank length would have been "proven" to be more powerful than the 170 mm crank. Even if there were not a "power advantage" to shorter cranks (at least to 145 mm) this study "proves" that 145 and 170 mm crank arms are equally powerful.
But, there are a couple of more issues with the Martin paper. First, it was done in cyclists riding in the more upright "road racing" position and not in the more aerodynamic time-trial position. We theorize that this effect will be even more dramatic when in the time-trial position. Our initial testing supports this theory as a pro-triathlete saw continued small power increases for the same effort as he shortened the cranks until he got shorter than 115 mm.
Second, as cranks become very short it takes a bit of time to adapt to this new length to learn how to ride them with good power. It requires learning the proper cadence and gearing to optimize the crank length. Martin tested without any training time at each length.
We have yet to have an athlete who has tried going shorter report a drop in power with the change, even with very short crank lengths, down to 130 mm. (One issue that prevents most from going much shorter than 130-145 when testing is how high the athlete can raise their seat with their current seat tube.)
The second paper by McDaniel, et. al. entitled Determinants of metabolic cost during submaximal cycling looked at the effects of crank length and pedaling rate (cadence) on the metabolic cost of cycling. In this study, crank length had almost no effect on metabolic cost.
There are some other reasons to think that crank length might affect power. That has to do with leverage. The average cyclist thinks shorter crank length affects leverage to the wheel, but this is not so if one adjusts the gearing to keep the total leverage the same. But, where crank length does affect leverage is in the leg joints. Leg muscles are most powerful and efficient in a relatively small range of motion. (Stresses on the knee start to increase greatly if the knee bends more than 60º. Standard size cranks typically bend the knee more than 90º at TDC.) If a crank is so long that it put the muscles and joints (knee and hip) in a less favorable position then power and efficiency can be lost. Everyone "worries" about what crank length does to the "leverage" to the crank axle (which is easily compensated for) but don't think about what it does to the leverage of the knee and hip joints and muscles. An "easy" test to look at what range of motion is optimum for you is to put yourself on a Stairmaster (a climbing machine that lets you choose how high each step is) and see what step height works best for you when trying to optimize your climbing rate for various periods. I predict you will find you take bigger steps if you are trying to optimize your climb rate for 30 seconds than if you are trying to optimize your climb rate for an hour. And, I predict your 1 hour step height will be on the order of about 6 inches, subtantially less than the 14 inches the typical bicycle cranks force you to take.
So, there are plenty of reasons to think that crank length is important to optimizing power but let's assume, for the sake of argument, that crank length has zero effect on power - is there another reason to worry about crank length? Why not just ride what came with your bike if power isn't affected? Of course, this ignores the affect of crank length on aerodynamic position and the affect of aerodynamic position (and crank length) on power and speed. Most experts agree that the lower one brings the front in trying to get into a good aerodynamic position the more the power is going to drop because of the difficulty in getting the thigh close to the chest. We would expect that this "crank length does not affect power" finding would not be so clear if position on the bicycle were added to the mix. Every rider knows they can only go so low before they begin to lose power and/or stop being comfortable. So, an even more important consideration is the effect of crank length on aerodynamics. Simply shortening the crank and moving the seat up the same amount (and doing nothing else to the bike) does three things in this regard.
1. It moves the butt up in relationship to the shoulders so it flattens the back, generally regarded to be a better aerodynamic shape.
2. A shorter crank and reduces the frontal area, important to good aerodynamics, and,
3. It opens the distance between the knee and the chest at TDC, reducing "cramping and improving comfort.
4. #3 also leads to the possibility of lowering the handlebars even more, reducing the frontal area further. This is where the big aerodynamic improvements can be found.
Several anecdotal reports to illustrate the effects of this. Pro-triathlete Courtney Ogden has reported that he has been able to drop his front end about 10 cm between a crank length of 170 and 115-130. He believes this is saving him about 15 minutes on the Ironman bike split. And, an article in Triathlete Magazine in 2008 reported that John Cobb reported reductions in drag, determined in a wind tunnel, of 30% from simply shortening the crank length and lowering the front appropriately.
At PowerCranks we believe experimenting with crank length will become very important to most serious riders in the future. Therefore, we have recently modified our line to allow the user to experiment with crank length down to 90 mm without changing Q factor. Our early work suggests that optimum crank length for the average male for time-trial events will be somewhere in the 100-130mm range. Every PowerCranker will now be able to easily experiment with this to see what works best for them should they choose to do so. The key here is for each serious athlete to not be afraid to experiment with this variable to see what is really best for you. Here is what happened to one who did just that: "I read the write up about shorter cranks and it intrigued me. I didn't believe that shorter cranks would be faster, but nevertheless I couldn't stop myself from trying them out. So, I went out and bought a set of 145mm juniors bmx racing cranks. and did some training rides all within a month of the 200mile Seattle to Portland classic. I found in training that speed, endurance, and climbing strength all went way up on the shorter cranks. And in the Seattle to Portland, I finished 4th, 6 min behind a recumbent rider and two guys who were working together the whole way even though my last 100mi were almost all solo miles. All on cranks and a saddle position adopted less than a month earlier (imagine how fast I will be when my legs fully adapt to this length). Thank you Dr. Day for raising this issue. it has impacted my athletic performance beyond my wildest dreams."
YOU TOO SHOULD TRY THIS CHANGE. YOU CANNOT KNOW THE POTENTIAL FOR YOU UNTIL YOU TRY. YOU MIGHT BE SURPRISED! IT IS WHY MOST MODELS OF POWERCRANKS ALLOW FOR THE USER TO ALSO EXPERIMENT WITH CRANK LENGTH. YOU CAN BE KILLING TWO BIRDS WITH ONE STONE, LEARNING HOW TO PEDAL CORRECTLY WHILE YOU ALSO DETERMINE YOUR OPTIMUM CRANK LENGTH.
The above argument is reiterated and differently in this article written for USA Cycling coaches or in this paper or as illustrated in the following video.