more on Cadence

Thanks to all those who contributed with their comments. It might be worthy to elaborate more with a new post rather than reply on the comments section.
It is really a interesting topic for me, because I have never seen any really good publication on the matter.
Basically running output is a result of this mechanical principle "Stroke length x revolutions" (stride length times cadence), so it would be very interesting to make studies on the optimal stroke and revolution based on the lenght of the legs of the runner, his speed, etc. Eventually, it should be possible to get charts like those torque/rpm/displacement available for cars...

to Mike:
yes, totally agree that the sensation of reducing the contact time on the ground is like having "stiff legs". Now I feel (especially in those faster workout), that my legs are like stiff levers rotating under the fulcrum of the hips.

to Alien Runner:
In my personal opinion, it is important to evaluate the starting point: what is your current cadence at different speed?
If your current level is still far from the recommended (88-92 depending on the speed), then it might be worthy to work more on improving the cadence. If your stride is already quick enough, than maybe it is time to pass on to improve the stride length.
Obviously we can expect that improving the overall fitness would bring benefit on the 2 directions (stride length and cadence), but working specifically on the personal weakness can give the best return...
If I take myself (as example):

Boston Marathon 2007 - time 2h42'09" => average cadence 85 / average stride 153cm
Fukuoka Marathon (December 2007)- time 2h38'26" => average cadence 86 / average stride lenght 155cm

So the improvement was due both to a quicker and a longer stride. Now, in my quest for running 2h35', I can calculate what I would need.
- Running at the current cadence (86), the stride should lengthen to 158cm
- Keeping the current stride lenght (155), I should spin at 88rpm (difficult...)
=> I guess that the most likely target is to reach 87rpm, increasing the stride to 157cm

Also amusing is calculating what could be my maximum speed if I could reach the theoretical optimal cadence (90rpm) at the current stride length ==> 2h31'
And dipping under 2h30' would only be possible if my stride can reach the 160cm...

to Simone:
in the PPT 5, there is a function "Reports" where you can monitor the trend of several data. You can choose which data to display and the date range, etc, etc. Unfortunately the PPT SW has some limitations in how the data are displayed and you can not draw average or trends.
So I usually prefer to extraxt the data to Excel and manipulate the data.
How to do: once the report is displayed, you can use the function Edit / Copy to clipboard to extract the data. Than I paste them to an Excel spreadsheet. I usually purge the data from all the workouts where I was not wearing the S3 sensor, to avoid the the presence of "0" would affect the real trend. In this way, I can see the trend of data over time and create moving averages to see the real trend.
I find useful: weekly mileage, Running index, cadence, Exertion

About your marathon, I have doubts that your stride lenght could have dropped so low at 130cm. Unless your pace dropped dramatically, the stride can not have dropped that much.

As a reference, I post the data of my Fukuoka Marathon (which was almost 99% perfect as execution). As you can see, the stride fluctuated only by few cm

Distance		Time		Split		Average Cadence			Stride (cm)
5k		0.18.34		0.18.34		87			153
10k		0.37.15		0.18.41		86			153
15k		0.56.00		0.18.45		85			154
20k		1.14.44		0.18.44		85			155
25k		1.33.38		0.18.54		85			154
30k		1.52.34		0.18.56		85			154
35k		2.11.26		0.18.52		86			155
40k		2.30.08		0.18.42		86			156
42.2k		2.38.24		0.08.16		86			154