Would anybody else like to see the “Recents Only (past 3 months)” Filter Setting be “1 year”? If not replace the 3 months option, at least add a 1 year option or let us pick how far back to go.
Here’s why it’s an issue for me. I’m trying to do more outdoor workouts this year or at least make my outdoor riding more productive using the Buckets app on MTB rides. Generally, those are more focused on fun simply because it’s difficult to maintain structure on a MTB on a trail in the woods. The Buckets app should help a lot with this.
However, I need to be able to quickly figure out what to do for a given day. The new Training Advice text and Forecast AI stuff (clicking on that training day) is a LOT more helpful than the old method. So when I click on the day I’m interested in training to see what’s ahead, it shows potential Workout Suggestions - which is AWESOME! BUT, it would be even more useful if, when I filter them by Outdoor, I can see suggestions of rides from the past year (rather than 3 months) because riding outdoors is very seasonal. I simply cannot do the rides I did 3 months ago due to those trails being covered in snow!
Currently, if I uncheck the past 3-month option, I get suggestions from long ago, like before I had a power meter (years ago) and rides with HR derived metrics. So obviously those aren’t very useful suggestions.
For Outdoor ride suggestions, given the innate seasonal nature of them, it makes sense to go back 1 year rather than 3 months.Maybe even prioritize suggestions from around the same time of year (assuming they first meet the metrics criteria)!
I like his idea too because it enables you to look back on a time frame that is the same season, same amount of daylight, kids in school, etc. So seeing rides that I did “last fall” helps me plan for rides this fall.
I also like the feature where Xert is proposing rides that match you target for the day based on previous rides. I also agree that 3 months might be a bit short because the season is only so long. So in the start of the season you really have no rides (because you were only using the trainer) and it will eventually grow. As 12 months might also be too short because it doesn’t fully cover what you this the same season last year.
It begs the question, why not use all of the activities? There might be some reasons for that:
You are in the wrong location. You have moved or did the ride on a vacation or similar.
The ride can not be replicated. You fitness has significantly increased or decreased so the XXS you will generate for such a ride is no longer correct or the time it will take is way off. Maybe you did not ride with a power meter, but now you have one.
A road has been closed/changed.
Seasonal variation. E.g. certain gravel rides cannot be done until later in the season when all the snow has melted.
We can give and for workouts. Maybe that could be added for activities as well? That way we can easily remove unrelevant suggestions? So if Xert proposes a ride we know we can no longer replicate we can just hit tombs down?
I also think it would be nice if the default time period was more like two years.
Also, maybe some kind of tagging of activities could help to find activities for the current location or season. Then you could do searches like «-gravel -mtb» or «spring gravel» or «mallorca».
I can see the reasoning behind the 3 months option to both speed the process (reduce server load processing prior workouts) and also to pick more relevant workouts (recent workouts are simply likely to be more relevant if you’re into a training plan already). I can also appreciate the extra thought needed due to training objectives that are not seasonal at all (events, challenge, etc.). I’m not claiming to have given this a whole ton of thought yet but here’s another idea:
How about adding a Seasonal check-box that adds weight (increases preference) for suggested prior workouts from 1 year ago? Maybe stick with a 3 months wide window centered on the date 1 year prior and eliminate the 3-month check-box entirely?
This can easily morph into a “suggested workout preference function” being derived and applied to the suggested workout search that could increase processing time for sure but there are reasonable ways to keep it simple yet still get what I’m after while also preserving the line of thought that’s already there (I think).
and 
for workouts. Maybe that could be added for activities as well? That way we can easily remove unrelevant suggestions? So if Xert proposes a ride we know we can no longer replicate we can just hit tombs down?