Just for fun

The Surprising Connection Between Sleep Quality and Running Performance

jonathan chng MnsnLMHofpM unsplash

A couple months ago I decided to get a Fitbit Charge 3 because I wanted to accurately track my sleep. For those who’ve read my posts on analysing my own running data, you would know that I already have a Garmin watch so hold on a second.. why would I have two watches?!

We’ll save that question for another blog post. I’ll just say that I find the Fitbit sleep tracking way more accurate in terms of start & end time, duration, and also detection of sleep phases (time asleep, time awake, and all the sleep stages).

Here I try to answer the question: is there a connection between sleep quality and running performance?

The obvious answer that comes to mind is a big fat yes. Bad sleep will result in bad running performance. There are also plenty of studies, articles and websites that say this.

Because my sleep tends to be quite regular, I wanted to see if there would be any particular insights that could be gleaned if I were to try find a relationship between how my sleep was on a particular night, and how my running performance was on the day after that night of sleep.

Going deeper into definitions..

But what does “how my sleep was”,“running performance” and “relationship between” actually mean?

how my sleep wasFrom the Fitbit, I can obtain for every night data, the following metrics:
Time asleep, time awake, number of awakenings, time in light sleep, time in REM sleep, time in deep sleep.
From these I can then calculate my own metrics that are alternate ways that capture sleep quality.
running performanceInstead of a perceived running effort metric, I chose to go with average heart rate (HR) during a run to reflect how hard my body is working.
relationnship betweenGiven that there are only 44 data points (but they’re all my own data points!), I’ll apply a simple linear regression to quickly identify relationships, where a true relationship will be defined by the factor achieving statistical significance in having a non-negative coefficient (in non-statistical talk, that it’s a true relationship).

Data & findings

My complete sleep data was from 1 October 2022 to 5 January 2023, and I filter this down to only the days where I performed a run and recorded it on my Garmin (which I always do) – this brings it down to 44 days where I recorded sleep quality on my Fitbit and I ran as well and recorded the activity on my Garmin – sweet!

You might also be interested in  [Jan 2023] How Sydney Properties Performed in 2022 - Revisiting 2022's Most Affordable & Expensive Suburbs

Next, I performed a simple correlation analysis between my average HR during runs and each of the sleep quality metrics for a first look.

Sleep MetricCorrelation
Hours In Bed18.4%
Hours Asleep17.6%
Hours Awake7.5%
% Asleep4.9%
Awakenings Per Hour28.1%
% REM-6.2%
% Light5.0%
% Deep2.5%
% REM of Asleep-7.6%
% Light of Asleep4.2%
% Deep of Asleep2.0%

Three metrics come out mildly positively correlated with my average HR during runs!

  • Hours In Bed: this would be highly correlated with hours asleep, and I would have thought that the more I slept, the more rested I’d be, and so the lower the HR during a run?
  • Hours Asleep: similar to above..
  • Awakenings Per Hour: the stronngest correlation – although still mildly positive!

These results look contradictory – I performed a simple linear regression (albeit I know it’s only 44 data points) to get an idea of significance.

I saw:

  • Awakenings per hour came through at a 95% statistical significance (i.e. p value of < 0.05)
  • For every extra 1 awakening per hour I have when I sleep, it causes a +6.1 in my average HR during my run
  • No other variables were even remotely as significant as this

This is huge!! My average awakenings tends to be about 4.2 per hour on average from my Fitbit. Does that mean that I can reduce my average HR by 25 by simply having 0 awakenings during my sleep?! I can only dream – this is only hypothetical.


I found a really interesting insight that for every extra 1 awakening per hour I have when I sleep, it causes a +6.1 in my average HR in my run.

You might also be interested in  [Jan 2023] 10 Most Expensive and Affordable Suburbs in Sydney

There are so many open questions from this which can be explored, however I lack the data (for now).

For now, I’ll need to try understand what awakenings mean on a Fitbit (a simple Google search actually isn’t very helpful) so I can try work on this.

Let me know your thoughts!

Leave a comment

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