Experience with daily suggestions for marathon training?

There is an excellent post from Mladen T   right here, but it is about a half marathon.

https://forums.garmin.com/beta-program/fenix-7-series/f/community-discussion/301141/general-daily-suggestion-and-training-readiness

You can see long runs in the 2h range for a half.

I am updating this post in case somebody else wants to use the Daily Suggested Workouts as their marathon training plan. This is only my experience with it so take it as a grain of salt. If you don't want to read the whole thing, my summary would be: the DSW is a good tool to get suggestions of workouts, with a lot of room for improvement from Garmin, but it's just that, a tool and it's more important that you learn how to listen to the signals of your body and not to rely completely on your watch to do so. 

The good:

- The DSW seem to follow the usual phases of a training cycle for a marathon and adapt the length/intensity of the workouts based on which phase you're on.

- My fear about it not suggesting a very long run was unfounded: approximately one month before my race, it suggested a base run of 2:47 (more on this later) and I also had to make quite a few runs above 2:10.

- My last marathon was 8 years ago and following the DSW helped me to get a PR (~26 minutes improvement) so if you are an inexperienced runner as I am, it won't hurt you to follow the DSW.

What could be improved IMHO:

- I don't quite understand why the DSW doesn't give an approximate distance of the workout instead of times. For example, for the easy runs (according to my watch, my easy pace [aka base by Garmin] is 5:08 min/km) the DSW could be 48 mins at 5:08 min/km. Unless you're running in a very short loop, I think that in terms of planning a course, it's easier to have the distance instead of the time, or at least have both! This is even more cumbersome to do when the DSW suggests an interval session.

- I posted about this a while back, but the training load/recovery time from some workouts has too much variability. For example, the DSW often suggested a base run of 53 minutes, so I just decided to follow a particular route for those sessions. However, for some days, the recovery time from those runs was just 8 hours, but other days, with very similar conditions, the recovery time was 28 hours. I get it, some days you're not feeling well and you have to make a greater effort for the same amount of work, but I didn't feel this was the case for most of the days whenever this happened. 

- The DSW for the following day can be updated just after your run in the current day. The DSW sometimes suggested a recovery run, which is ok, but then, as soon as I finished that run, it was recommending another recovery run for the following day as if the recovery day from that day had been too hard on my body but I felt fine. And the first time this happened to me, the following DSW was also recommending a recovery run (so 3 consecutive days of recovery runs!). However, that third run coincided with a long run of 2:16 according to the previous day schedule (see next paragraph), so I just did the long run instead of the recovery run. This made me afraid of having recovery runs hehe.

Depending on your habits/perspective, this could be good/bad

- The DSW gives you the workouts for the full week ahead of you and you can actually choose any of the workouts for that week to do it right away. I do like to run during the very early morning (5 AM), so I plan my routes the day before. At some point, I found that the DSW was quite sensitive to some of the factors (e.g sleep, see next paragraph), so this generated constant/abrupt changes in the DSW for that day with respect to what it was recommending the day before. At some point, this was quite annoying so I just took the schedule provided the day after a long run and roughly use it as a baseline for the rest of the week and choosing a similar workout to the one seen in the baseline whenever the workout for that day was too different. 

- The DSW takes into account multiple factors to suggest the DSW: sleep, recovery time, HRV, etc. I don't follow a perfect sleeping pattern but that's how my body works. That being said, I found that the DSW is quite sensitive to the sleep score you got the night before the workout. If you feel that your watch is accurate in tracking your sleep and whenever you have a bad night according to the watch, you feel indeed tired, then you'll be happy with DSW. However, I don't feel like my watch is very accurate in tracking my sleep so it was annoying to see that the watch changed my workout for that day with respect to it was recommending the day before. 

Thanks for the update! I enjoyed reading it.

M

Thanks for sharing! 

Yes, an alternative way is to work with the speed instead of the pace. For instance, a 5:08 min/km translates into 11.7 km/h, then I would multiply the total number of minutes at a particular speed and divide it by 60, so for the 48 minutes workout I mentioned would be (11.7 x 48)/60 = 9.36k and this also works for interval sessions etc. 

I second your comment about the DSW to be a good tool. I was able to beat the estimated race time by ~14 minutes and I've read over the internet cases in which the estimated race times are over/under estimated, but I think they're a good guideline to set a realistic goal.

Sure there can be improvments to the DSW.

But to be fair I think you can do an aproximation pretty fast, and estimate the run to be 9-10km long. Do you really need to be more precise? Pick out your favorite calculator and put 48÷(5+8÷60).

Btw I have a question. If you do the run a tiny bit faster than 5.08/km will your run be longer (still 48min) or will it be less than 48min run (still 9.36km)?

M

Yes, I agree that one can do an approximation relatively fast, but this is more straightforward to do if it's a base run (i.e. constant pace for the whole workout). If it's a threshold/ VO2 max/Tempo run with intervals, the computation is a little bit more tedious. I guess all I'm saying is that it would take a programmer to write couple of lines of code to display the distance for the convenience of the inexperienced runner like me :). 

As for your question, I personally tried to meet the time goal whenever I couldn't specifically keep the suggested pace.

I found that it didn't do a great job with training for an Ultra 50 trail run.

Garmin had the course in the calendar so knew the distance, terrain and accumulated height but it seems that it maybe doesn't take that data in to account, maybe it just uses the distance and that's all? It left me under prepared. I think for a normal marathon without hills it would probably do a decent job.

Interesting, my "general marathon" prediction time from my watch was ~7 minutes lower than the race prediction time for this specific course, considering that the general marathon prediction time is set under "perfect conditions", I thought that Garmin indeed took into account the elevation gains. But on the other hand, I never saw any hill workout during my training from DSW so, your assessment is probably correct.