Forecast AI planner feedback

@ManofSteele @xertedbrain
Hello, some feedback on AI planner after a couple of weeks, I’m a little mixed after 2 weeks, but all-in-all it’s a great new tool, so i hope this helps refine further!

  • I feel that previously the gauge was giving a much more direct (and motivational) feedback: it told you what xss you were needing, and the gauge told you immediately if you were ahead or behind. Now I 'm seeing green (and in my case too many orange and red) stickers. but that’s basically thumb up vs thumb down, so much more crude. with the AI planner you set a target, and AI tells you if its achieveable or not, but not how hard it will be to achieve, so I have had to stop several intense workouts midway, and adjust the target downwards several times concluding my target was way to ambitious. And even though the AI planner sucesfully built a plan for my target (and it has my history and actuals to work from), it was just too hard to complete. I’m therefore getting red and orange scores across my planner, this is a lot less motivational (dare i say borderline demotivating).

  • i’m 100% convinced now that the previous gauge was missing the low/mid/high side of the coin, and adding this is an absolute leap forward. Yes, the focus setting you chose would point you to recommended workouts, but in the end, the gauge evaluated on total/overall XSS. AI approach makes much more sense, (provided the quantities of low/mid/high is works from are the right ones)

  • the AI planner might need some extra settings for a user. Eg, currently i’m getting weeks with 3 high intense workouts in a week, and/or intense workouts on 2 consecutive days (with no rest inbetween), this doesn’t feel right, both from a 80/20 perspective, and from a "120 days before a target date you lay aerobic base, the hard stuff follows after). I should be able to set a max nr of high intenste WO in a week - my current polarization is at default 4:1. AI can then tell me if a target is out of reach if I set it a certain way. (back to the AI planner only telling me Y/N target is achieveable, but not how hard it will be to achieve)

  • I’m essentially preparing for a 130Km ride with 1300 Alt meters in May, being several 5-8min climbs, one or 2 longer ones, and a lot of flattish miles inbetween hills. the focus guide tell me that’s GC (8m ower focus). However, it seems AI is very focussed on the 8min power side of the profile, and much less on the endurance side: back to the: 4/1 balance and +120 days pre event base focus

to counter this, I’ve planned several endurance rides in the next weeks, before lauching the AI planner, so the planner works around those ('factors those in") and this method seems to work, but it feels like I’m bending the system this way?
Also I’ve considered moving to the continues improvement but I’m a fan of the AI concept, so that doesnt seem right either.

I hope this feedback is of any value to continue this exciting journey!

ini summary
1/ consider finding a “'achiemenent”/“well done” type of gauge . I know this is stupid, but I seem to miss the joy of seting the improvemet rate a bit more aggressive for a couple of weeks because I’m doing well.
2/ i’m a little doubtful that the AI is effectively preparing me for my ride, not sure if its the AI, or me choosing the wrong focus with GC, but I need a mix of 8min power and solid endurance. the plan doest really tell me why it picked what it picked


Try with the latest updates.

I think overall it’s really cool to see that development. I now started using xert for running with focus on long distance. If I put in a target event, the AI gives me basically two sessions per week with a duration of around 4h each… Not really feasible. I understand why it says so but some more options to provide some boundaries would be cool

Hey guys! I’m a longtime Xert user but had fallen into a routine of using Xert mainly for “tracking my fitness/performance.” But I got re-energized with the AI planner and have it set for a gravel race (Barry-Roubaix on April 20). I agree with @Blouies desires (and the weirdness of the plan: mine is prescribing 14hrs next week and then a series of low levels weeks while my event is 15 weeks out :).

Another suggestion and great-to-have: I’ve done this race before and have the files. Rather than letting the user only chose data/type of event/length of event from which the AI tool decides needed XSS etc. I would be great if I could just link to the race I’ve done before so it has a much better plan on the actual “signature” of the event (and what I was able to achieve in the past in terms of XSS etc)


@xertedbrain great new functionality!

  • having experimeted, i have the impressoin that the dropdown '(friendy groupride, race, gravel, competitive grravel…) mailnly infuences the XSS?

  • with event (vs goal) menu, it seems you can’t use the “more/less polar” slider nor the focus type (GC, rouleur,…) slider, is there a reason for that?

an extra addition could be a 3rd option where you can enter 3 datapoints from a ride you did before as a guide. what i mean:last year I did a ride on a cycle weekend, and i can see from activities that it was 4.5h in duration, my XSS for that ride was 325 and xert categorized that ride focus as Mixed GC specialist. It could be a nice extra if I could enter these 3 values: duration, XSS, focus. Or even simpler if i could flag that ride as a reference ride for AI calculation.

but it 's a great new feature for sure!