Bike Fitting – Short Crank Arms

Accurate fitting for a bicycle is just something that most of us never consider. There seems to be far too many variables involved so we probably just feel that we will adapt and make the most of it. I can’t remember what specifically spurred my interest in this subject – probably just surfing the topic on the Internet out of curiosity and somehow landing up on the subject of  “crank arm length”.

DuraAce Binned

The net result is that my expensive 170 mm DuraAce crank arms are in a cupboard and I’m now on a set of very cheap (compatible) 165 mm crank arms – which are there to stay because this apparently small difference is surprisingly, enormously beneficial. 

When stumbling around on the Internet trying to find sensible information even on crank arm length the only coherent source I could find was Living in France made arranging an appointment out of the question so I purchased the DIY guide and initially just used it to work out crank arm size.

BikeDynamics DIY

The little 45 page book is a treasure and really raises your awareness. There may only be 45 pages of information – but it’s because it cuts out the BS and brings the benefit of a huge amount of experience. There is a clear avoidance of impractical “perfection” – yet the recognition that when you find your optimum even a couple of millimetres can make a huge difference. 
For crank length I used my inside leg measurement and another measurement to the top of the hip bone (trocanter) – inserted this into a given formula and out popped a figure of 167 mm. The advice is to round down to the nearest available size because it is always apparently better to go smaller not bigger. Other ways of estimating or calculating were all suggesting to go lower too so I settled on 165 mm.


When cycling I’d always had a problem of bouncing on the saddle when spinning out at over 70 kph. When asking others about it nobody else had this problem. It turns out as Mike from BikeDynamics explains in his book and on his website that it’s due to the cranks being too long. There was also a constant problem of my thighs hitting my lower ribs when trying to pedal in a more aerodynamic posture – and not being able to tilt my pelvis forward (more efficient) because of this. Just 5 mm reduction of pedal crank arm length and those issues were history – but not without a battle!


Unfortunately initially I didn’t read the rest of the book – it implied setting up the bike on a turbo trainer, filming and measuring and seemed like a real pain to do. Mistake! This is not only essential it turns out to be fascinating when guided by the little book or probably even much better if you can have a professional work with you directly. Consequently, for some unfathomable reason, I decided to lower the saddle a full 2 cm from my “normal” height. I simply figured that there had been so much messing around over the years trying to compensate for the “bouncing” etc that the saddle had gone higher and higher so it was best to get it back down to a reasonable height. Sure enough,  the usual problems had disappeared anyway, but now my quads burned like mad and were on the verge of serious cramping after only 70k. Despite all that sprinting was easier and there was no discernible difference in times over longer distances – yet now I needed to drop one gear lower in general and use a higher cadence. Several different issues were being muddled up now.

Gain Ratio ( )

Sheldon “Gain Ratio” Brown seems to have done a great job of figuring out what happens to your effective gearing when you change the crank lengths. My initial experience of the shorter cranks was that they caused a bit of a nightmare climbing if I stayed in the usual gearing and I was forced to lower by at least one gear – but in compensation the shorter cranks enabled a naturally faster cadence so in the end the speed was the same. On the plus side sprinting appeared to be better but on the negative side the quads were destroyed.

Superficially the problem seemed to be a reduction in the leverage provided by the short cranks. Sheldon Brown however puts that misunderstanding to rest. The critical issue really is the distance travelled by the foot (in a circular path) related to the distance travelled by the bike – the amount “Gained”. Shortening the cranks means that for the foot to travel the same distance it has to go a few centimetres further than one rotation – and so the bike travels further on the ground.

If you look at it the another way – say the bike keeps everything the same except the cranks – then to have the same speed the actual feet are travelling slower (shorter cranks) – which is why you can eventually use a lower gearing and higher cadence comfortably and maintain the same power as before.

My quads were certainly being affected by the greater force required when using the same gearing as previously – but in reality the effect was being amplified unreasonably by lowering the saddle.

The improved sprinting probably had a lot to do with the shorter cranks allowing greater bike speed for the same foot speed as before.

The Inevitable Bike Fitting Session

The question was whether to continue with “adaptation” and wait until the legs sprouted new muscles or to try to see if setting up the bike differently could fix the exploding quads. Grudgingly I set about studying the bike fitting book and preparing equipment. First of all this meant digging out my old turbo trainer that I’d collected in Texas in the 1980s and dragged everywhere with me – then setting it up level. (Thankfully I’d failed to throw it out though it was on the current outbound junk pile). It’s actually quite good – makes far less noise than my stupid Tacx (computerised piece of overpriced annoying crap) Trainer with the worst software on the planet. (Do I really need a motor to spin my wheel on downhill sections?)

My “uninformed” approach was to try the saddle at its current level and then move it up 1 cm at a time for a total of 4 cm, filming each setting. Subjectively the highest level felt best so I took still shots from the video and imported them into a drawing program where angles could be accurately measured. The knee and hip were on the limit of acceptable extension but surprisingly the ankle was actually the opposite – literally flexed when it should be extended.

Ankle Dorsiflexion Correction

Although I don’t bend my ankles in skiing even a strong stance on the front of the heel requires the ankle flexed to about 12° – so it looks like this had migrated into my cycling. Realising that all the angles – hip, knee, ankle would be improved if the ankle dorsiflexion was removed it was necessary to work with the video to get a more “toes down ” pedal stroke. What felt like being extreme “toes down” to me was in fact just about right or not even enough. Apparently pedalling correctly is very alien to me. This is the first major thing I learned from the bike fitting process – this alone being amazingly valuable.

Handlebars Up – Saddle Forwards

The handlebars were raised one centimetre to the position they were in when supplied by the factory. I realised that I hardly spent any time on the drop bars as they were already too low before even moving the saddle up. The saddle was moved about 1.5 cm forwards (now almost centered – it had been set backwards) to get the knee position correct over the pedal spindle and now all the arm angles fell correctly into place and it felt great. This was the point where I deluded myself into thinking the job was complete. I was just happy that the upper body felt great, the legs were no longer being destroyed and I could return to normal gearing on the climbs despite the increased “Gain Ratio” of the shorter cranks. Progress had certainly been made.

The Crunch

Contacting BikeDynamics to thank them for the advice it was suggested that I might make a comparison against calculated saddle height by using leg measurements. Reluctantly after a day or two I forced myself to make the dreaded leg measurements and then do the calculation. Turns out the formula suggested that my ideal height was 4 cm lower – exactly where I’d started off with the legs being destroyed. Unbelievable! That was clearly not ideal for me – but the discrepancy was far too great as statistically there was evidence that the leg calculation method is relatively accurate. Back to square one – almost. I decided that at least I had two extremes – to start from an upper and lower limit – and the right height would almost certainly be somewhere in between. Back to work!


The first obvious thing to do was to try the two extremes and note the differences. The high setting felt like there was an early loss of pressure during the stroke. The low setting felt like there was never a break in pressure and no relaxation at all of the quads as a result. Both could probably be considered acceptable solutions just though force of habit if someone got used to either of them. Placing the saddle in the middle (2 cm) of the two the positions the sensation was still “loss of pressure”. Lowering by half the distance again to only 1 cm up from the bottom setting it was a return to the constant load on the quads. At 1.5 cm the load was still predominant. It was only at 1.85 cm that the load vanished and and so did the loss of pressure. (3.85 cm up from the seat tube top to the bottom of the “Y” in the “Ritchey” seat post – note for my memory) It turns out that there appears to be an optimum setting – despite all the variables with the ankles etc. The fore/aft position of the saddle was corrected and this didn’t appear to change anything other than the upper body fit – which once again felt right.

(Inside leg 80cm trocanter 88 cm – calculated result 71 cm saddle height from centre of axle – 75 cm being the height worked out just by looking at angles. Final height was 72.85 cm)
01/09/2019 – moved saddle forwards 1.5cm, raised saddle 7mm, cleats pulled back around 1cm on shoes. This has been arrived at in stages – first moving the cleats to where they are now and the saddle forwards only 1cm. Realised I needed the saddle even more forwards to be fine in the drop handlebars and with the cleats moved back. Have more power from being more “over” the legs. 

Going out on the bike for a climb it was clear that sitting back in the saddle allowed that “optimum” feeling where there was just enough muscle relaxation at the end of the stroke but no real loss of pressure. When sprinting a slight slide forward in the saddle allowed the quads to remain engaged all the time and this felt appropriate over such a short distance.

I have no idea if any of this is right – but it kind of makes sense to me. In addition my bum is a lot less sore than when having the saddle high, legs less sore than when having the saddle low and there is plenty of freedom with the ribs/knees for a good aero position (though the stomach need to be reduced!) and spinning at high cadence. There is no bouncing with the short cranks and only apparent advantages with the shorter cranks – no disadvantges.

I’m writing this here more as a diary than anything else – because it’s all very easy to forget. If I  was in the UK though I’d definitely prefer to have the clearly vast experience of BikeDynamics do the job directly. The stunning thing is realising that when close to being accurate then just a millimetre or two can make a huge difference – just the same as with crank length.

One of the most obvious aspects of feedback when the angle is dorsiflexing is that the sole of the cycling shoe pulls down and away from the foot. I’m finding the simplest way to get clear proprioception is to focus on maintaining pressure along the underside of the foot.

Here is a photo of the ankle over-flexing – unconsciously – despite the saddle being way too high…

One comment

Leave a Reply

Your email address will not be published. Required fields are marked *