Like so many who have gone before me in the fight against flab, I am engaged in an unending war with my body. I don't mind exercise - I jog, cross-train and swim - but I do love food. Children learn early that when your stomach is full it is a good idea to stop. It is a skill I have yet to develop.
These two sides - the exercise and the eating - are finely balanced. For months at a time, exercise will prevail, but it takes only one missed gym session for the discipline to fall apart.
Enter MiLife, a web-based system that claims to be the world's first "personalised online coaching system". As I am never going to get an actual personal trainer (why pay for someone in a tracksuit to shout at you?), I thought a virtual one could keep an eye on my progress and shame me into action.
The system comes with a wristband that records all the movements you make in a day and, when connected to a computer via bluetooth, uploads this data to a personal profile on the MiLife website. Every week, you track your performance with a plethora of bar charts and line graphs and the MiLife software advises you on how to get the best out of your exercise.
To start, you tell MiLife what your goals are. Perhaps you want to raise your activity levels or lose some weight? The website's virtual trainer will come back with a personalised plan, broken down into daily targets. As you progress, the software automatically adapts the plan during a weekly coaching session to take into account the exercises you seem to be good at and those you're not.
I chose to give myself both exercise and weight targets, but rapidly regretted the latter. Weight control involves recording a daily food diary, an activity as tedious and irritating as filing tax returns. Every day. I tried, I really did. MiLife even allows you to use your mobile phone to text in how many calories you eat but, seriously, how do you know exactly how many are in a salmon mousse? I gave up after just a few weeks of semi-completed diaries and, during my weekly online coaching sessions, the software duly reminded me of my laziness.
I was more successful with the wristband, which I wore obsessively. MiLife breaks down activity into low, medium and high. Shuffling around my flat was low activity, a brisk walk counted as medium-to-high and a jog or even the odd dash for a bus would rack up minutes in the high-activity section. Like anyone given a target, I did everything I could to get the daily totals up: I walked into work more often, went walkabout at lunchtimes, and avoided buses for all short journeys.
All the information about my activity was recorded with no need for my intervention, and it was useful: days when I took the bus home, for example, instead of walking, appeared as conspicuous gaps among the skyscrapers of activity in the days where I had been more diligent. I could monitor my minutes of high activity from jogging or cross-training to ensure that I kept up the levels suggested by the software. All of this was motivational, too - I was surprised how far I would go to get a perfect set of bar charts.
If you choose, MiLife will email or text to get you exercising, and chide you if you miss too many sessions. The virtual trainer is powered by something called the "Idapt engine", a computer model that MiLife says is the result of five years of research collating data from hundreds of people to tease out successful strategies to, for example, lose weight or keep motivated to exercise. During the first few weeks of use, this builds up a profile of the kinds of exercise that seem to work for you. By matching this to the profiles it stores, it can suggest exercises or ways to break consistent bad habits. I was advised, for example, to try an exercise bike and do more gentle jogging, but the longer you use the programme the better the suggestions should be.
In a randomised controlled trial of 77 people over nine weeks, published in the Journal of Medical Internet Research in 2007, those using the MiLife system ended up doing, on average, two hours more physical activity a week than the control group. This is a good result, but bear in mind that these were probably active volunteers, so likely to be motivated to exercise.
There are niggling problems with the system: the website is slow, badly designed and frustrating to use. As a Mac user, I found the software a small nightmare to set up and the system lost two weeks of my weight and activity data. That meant my programme was all but shot to pieces because the software assumed I had been lying down for a fortnight.
I didn't manage to make MiLife record my activities for long enough to complete a 12-week programme but, on the evidence I do have, my feelings are mixed. Just knowing that all your movements are being recorded is surprisingly rewarding and motivational. Small bits of low-level exercise can add up, and visualising all the jogging and cycling with the bar charts every day was (when I was wearing my geek hat) addictive.
The weight-loss part of the MiLife programme was defeated by my lack of willpower. But the exercise plan definitely recorded an increase in my activity in the weeks that I used the system. Whether that was entirely due to MiLife, I'm not so sure - most of the increase came in the low-level exercise - the jogging or other aerobic exercise I would have done anyway.
So a partial success for me, but is it worth the £99 it costs for the basic equipment and a year's subscription to the website? It might not be as expensive as a personal trainer, but if MiLife is hoping people will put their hands in their tracksuit pockets, the technology needs to be more impressive.
• For more information, see milife.com
• This article was amended on Wednesday 14 January 2009. Milife, a computer based personal training system, costs £99 for a year's website subscription and all the basic equipment, not £200, as we said above. This has been corrected.