Curves, colours, shapes & killer robots.

A recent comment from a rider got me thinking. I was reviewing the data from one of her sweetspot sessions and I wrote ‘all in parameters, good shaped HR curve’ in the Post Activity Review box helpfully provided by Training Peaks . ‘Great!’ she said, ‘What does that mean…?’

Fair comment, I’d slipped in to using unhelpful jargon and worse still. I’d slipped in to writing notes in training records as 'feedback' that were more use to me than her. Bad coach, bbbbaaadddd coach. Lesson learned, but it did remind me to check my use of language and force me to explain what the jargon meant to my bemused rider. It also prompted me to reflect on just how much effective use of data has become vital in my coaching practice.  

If you’re in the cycling matrix – by which I mean using GPS enabled data recording including power and heart rate measurement -  and you regularly upload it to an App (Strava, Garmin Connect, Training Peaks) you’ll be familiar with the colourful blocks, graphs, maps and pictures produced. Whether you bother looking at them or not is another matter, but the the App will collate your data and try to turn it in to digestible performance information. Some of it is useful, some of it less so.
​
For me as a coach, the most powerful performance metrics are power, heart rate (HR) and time. For our purposes HR can be defined as an input measure (an indicator of how hard you’re trying), power as an output (the product of your efforts) and time is that thing Einstein talked about although we can define it here as three 8 minute efforts with 4 minutes rest in between. Phew.  

In the diagram above, the boxes describe the session, the blue line is power and the red line is HR. Every rider is different but here you can see power remains fairly constant for the efforts and HR arcs upwards gently, but progressively more so, for each block. This indicates increasing effort is required to maintain consistent power. A good session completed within the prescribed parameters.  

Sweetspot efforts tend to be in multiples of 8, 10, 12 or 20 mins at 88 – 93% of FTP with varying rest in between dependent upon what effect you’re trying to create. They mimic volume riding without all that pesky getting dressed up and going outside for hours on end. Very generally, longer sweetspots tend to work best at the end of a block of training maybe just before an FTP test, shorter less stressful ones are better soon after an FTP increase to see if the power/HR correlation makes sense.

Over a period, the delta between the HR and power curve alters as fitness changes. This is true of all sessions but sweetspots are useful because of the relatively long length or the effort giving time for HR to respond and stabilise. Increases in fitness are often indicated by the HR curve flattening with power remaining constant, or power increases over zone (or both if you’ve left it too long to boost the FTP settings). Steeper HR increases or drops below the power zone indicate the session is maybe too hard if that’s not what was intended. This can indicate an adjustment to the sequence or intensity/duration of training sessions is needed to ensure fitness matches the planned goals.
 
Over a period of a few weeks, a data trend for a rider develops meaning better targeting and sequencing of sessions during any week to reflect the stress of a previous week and allow for appropriate recovery or aim to peak for a forthcoming event. The evaluation of data at this level allows fine adjustments to sessions to ensure progress is being made without constant FTP testing which is psychologically and physiologically stressful and can be a rather blunt and overused tool. 

So what does this mean for you as a rider? Your data is powerful stuff,  it is ammunition to make you progressively stronger and faster so... 
​

• If you do not record your data, perhaps start recording it
• If you do record your data, use it
• If you use the data, use it objectively and implement it with knowledge to improve the quality of your training
• If you struggle with objective interpretation or are unsure you have the knowledge, get a coach to help or...
• Think about if you really need to go to the bother of recording the data.

Used effectively, training data will help protect you from injury and over training and give you empirical evidence your fitness is improving or, if it isn’t, what needs to be changed to bring it on track.  Fundamentally, accurate interpretation of your data means productive use of your time to achieve maximum training benefit and bigger improvements in fitness than you could hope to get without it.

With the rise of Big Data and Artificial Intelligence, could you let the matrix do this for you – avoiding inconvenience of finding, and paying, a coach?. Algorithms are learning fast but they don't cut it, yet, although as AI becomes more sophisticated and we upload more human biometric data, it is likely it will become better at interpreting human behaviour and responses over time.  I guess how scary this is depends on whether you think of Brave New World as a good idea or not. However, the machines aren't there yet and, as much as it irked me at the time, I was gratified to see my Garmin 520 once telling me I needed 42 hours rest in the middle of a 10 mile TT. Even if the advice was good, I’d seriously question the machines message management technique.

---

​Rich Smith likes machines and is a fan of the first two Terminator films. He thinks the others were, frankly, rubbish. He has coached the GB Transplant Cycling team for 10 years, is a British Cycling qualified Level 3 coach and a mature psychology student. He spent 30 years responding badly to people in authority in senior roles for Barclays, HSBC, British Waterways and National Grid Property and will probably help the robots out when the time comes.