Finally arrived in Kona after a long but reasonably relaxing 2 days of travel. It gave me lots of time to work on my new training Database and try out some ideas I’ve had.
I am a firm believer that consistency over a period of months, years is what really is the key to Ironman success. It out trumps volume to a point. By this a mean consistent week in week out periods of moderate volume are better than doing some monster weeks and then being a couch potato for one. I’ve been spending a lot of time recently moving all my training data into a database I’ve built. This harps back to my old life as a data modeller and programmer and it’s given me a great deal of pleasure. I remember that I did actually enjoy my work!
Having my data in this form and allowed a load of “Data Mining” – I’ve seen some interesting things and I may blog on that at a future date. It has allowed me to experiment with trying to measure and track “consistency”. I wanted to try and capture the essence of what I feel is required and back test it against my performances. A big part of this is to be able to start to use this with my athletes. This is very similar to the idea of monotony (standard deviation of daily load) and strain (average load divided by monotony) which I track. There are problems with this
- I don’t want to look at consistency on a daily basis. How the week is split up doesn’t matter so much as much as overall volume for the week. NB: this is not a fact, it’s just my opinion that I will try and test against my data
- I’m saying consistency is good so I low standard deviation is good. For what every reason I like bigger numbers to mean better things … so I went with 1/standard deviation
- Other problem is that standard deviation would mean doing nothing was good … thats bad. So I multiply the lot by the average load squared. Needed to square it as consistent very low volume produce such a low Stdev that it outweighed the low load.
I tried loads of different methods of trying to measure what I wanted and none quite worked as well as I’d hope… I will still play with it but so far what did I get ? Well a reasonable surprise – it kind of did the job. I’ve set it up so I can tweak the period I look over. So the graph above looks at consistency weekly by looking at how consistent training has been for the previous 12 weeks. It shows my Ironman Marathon times and where they sit compared to consistency.
My first three marathons are more or less my best and they show they follow a solid period of consistent running. After that the times have dropped and really the steady improvement post injury has come off the back of consistent running. After Lanza I had a decent break this year but now I’m seeing consistency figures which I’ve not seen since Lanza 2007 which was probably my best Ironman Performance.
How does this fit with the volume I was doing ?
This shows consistency in red with weekly bike mileage in Blue. You can see during period of good consistency and fast running I very rarely went above 80 miles per week. I certainly fell in to the trap of more is better with my running. Hammering out 100 mile weeks (this work for Ironman NZ 08) but I think long term is just so training that consistency is sacrificed as very low volumes are needed to recover. I’ve bourne this in mind in this build period (more in a later post) and really held myself back. If I’ve hit volume targets by Saturday I haven’t used Sunday to go beyond them I’ve taken it easy. I’ve had a few very pleasurably Sundays taking it easy these last months something I never used to do. My run mileage this last 4 weeks has been 50,52,51,52…. This consistency really shows up if I do the analysis using rolling 3 weeks. Since this period was more consistent than I’ve ever managed it outweighed any other period of running despite the relatively low load. Here’s the graph based on 3 weeks.
Thats consistent. Lets hope I prove it’s worth come a week Saturday.