I understand what lactate threshold is, and the "normal" procedure for [guesstimating][1] it (short of a lab trial with actual measurements) seems to be to run a private 20 minute race after warming up and taking the average of the HR at the start and end of that time, and the average pace.

I have been using a Garmin Fenix 6 watch for years (running and other sports), and am often running diverse training sessions - sprints, intervals, fartleks etc. in addition to long slow runs.

Now to the question: after some runs of sufficient length and difficulty, the watch offers a new estimate of my LTHR. For example, today I ran a VO2max training session (45 minutes of easy pace, with two 4 x (30sec max effort - 30sec low pace) intervals broken up by a 5 minute recovery period). After this session, the watch offered a new LTHR estimation, which was only slightly different from my previous one, so I have no immediate cause to doubt it.

Reading [Garmin's info about their LTHR procedure][2] it only mentions some relationship to breathing, HRV etc. - I guess the actual algorithm used is  a well-kept secret. 

Is it safe to assume that a sport watch has enough measurements to guess a relatively correct LTHR based on "whatever" training session? My run was nowhere near a smooth, steady pace, as the usual DIY method implies, but was very spiky, owing to the all out sprint intervals. Are there scientifically proven methods that do not require a steady race pace, or is it more likely that the watch "thought" it was a regular run for whatever reason, and took a wild guess?

  [1]: https://www.outsideonline.com/health/running/training-advice/running-101/do-it-yourself-lactate-threshold-testing/
  [2]: https://www.garmin.com/en-US/garmin-technology/running-science/physiological-measurements/lactate-threshold/