Hey everyone, I hope you’re all enjoying your training!
I’m very curious to see how users are using Xert’s Forecast AI training plans. I think that sharing our experiences with one another can provide valuable insights and help each other maximize our training!
With that in mind, I thought I’d share how I’ve been using Forecast AI. Maybe it’ll provide some insights or help you learn something new!
Power Goal Program: Following a year of doing mostly long & easy endurance riding in preparation for long course triathlon (which is now completed!), I’ve now set a target to improve my 6-minute power by the end of the year. I’m looking forward to having more ‘punching power’ for group rides & will hope to start targeting some Strava segments in my area next season!
Since my High & Peak training loads are relatively low, I needed to use the Polar Specificity option to achieve my power increase, which will put more emphasis on my Low Intensity (Threshold) system, rather than overloading my relatively untrained High & Peak systems.
Adjusting for Availability: I’ve enjoyed how XFAI adjusts my training based on my preferred schedule, including my preference for a Sunday rest day. Also, being able to pin activities & adapt my forecast around them has been great - I’ve used this feature to plan around my weekly Xert Surplus Seeker group ride on Zwift… +1 for the new Zwift integration!
Program Settings: I’ve used the default polarization setting (4:1), which looks to be giving me roughly ~2 High Intensity workouts each week. I also slightly increased my XSSR preference to 120, so I can make the most of my time for high intensity workouts!
Looking at the rest of my Forecast chart, it looks like XFAI has given me several rest days. I’ll probably plan to ride easy for an hour or so on those days, rather than completely taking the day off.
Final Results: I’ve been deciding how I could/should test my fitness at the conclusion of the program and see how the results align with the initial projections… maybe I have to do a 6 minute effort? I’d be curious to hear what others would like to see!
To be honest.
Ive not had any success with it.
I either get not achievable or I get a decrease.
Ive just done something that you mentioned.
Initially it said 4 watts increase was achievable.
Do you have restrictions on your availability? What’s your maximum weekly hours? Did you try adjusting the program difficulty setting at all? All of those factors will have an impact on program achievability.
Try playing around with some settings - maybe adjust the Focus - and share what you find. Screenshots can likely be helpful if you’re looking for suggestions from the community.
Hi Scott.
I have changed hours and focus on numerous occasions it always is green then programme runs and either goes backwards on TP or says not achievable.
.
I get a bit disheartened when this keeps happening so I just dont bothet.
Next time I do one I will show you.
Thanks.
I’ve been using the AI based training plans since April, and they’ve been awesome. Following Xert’s Forecast AI training plan, my fitness has improved massively (although my starting fitness was fairly low).
For the most part, I generate the plans with GC Specialist or Climber as athlete type, since I tend to favor slightly longer efforts. I prefer to plan for a roughly six-week period, then go for a breakthrough at the end.
Like others here on the forum have commented, understanding how to use Forecast AI is not as easy as I’d like it to be:
Xert seems to be fairly conservative in terms of the gains it thinks you can achieve, and if it doesn’t think you can get to a given goal, it refuses to generate a training plan. I’ve spent a lot of time tweaking availability, recovery demands, max weekly hours etc. to get Xert to create a plan that I like. Personally, I usually beat Xert’s expectation when I go for a breakthrough towards the end of a training block.
One particularly vexing issue I’ve run into is that the Xert AI doesn’t seem deal well with me taking time off the bike. After the last breakthrough, I blocked off a week as unavailable, thinking I was going to just rest. During that week I tried in vain to get Xert to generate a new training plan starting after the rest week was completed, but no matter how much tweaking of parameters I did Xert never came up with a plan that would gain me more than a watt or two over my usual six-week planning period. In the end, I couldn’t stay off the bike for an entire week, so I just did some unstructured rides and got a little training load in. When I came back to generate a plan again, Xert came up with something much better. The only explanation I can think of is that the system drops my training load precipitously if I don’t do anything for a week, and then wants me to spend a week or two ramping up the training load again. Personally, after a week completely off I might see a tiny dip in fitness, but I’ll be fresh and ready to pick things up where I left it. It doesn’t make sense to me…
Usually, at some point during the training block Xert will decide that my chosen goal is completely unrealistic even though I’ve completed the prescribed training and none of my parameters like availability have changed. I’ve basically learned to ignore Xert’s suggestion to adapt the forecast unless I really need extra time off due to illness, injury or other work/life events. Normally, I take some time to generate a training program I’m happy with it and leave it as is, although I might switch a hard and an easy day depending on the weather, how I feel etc.
I don’t really trust Xert to generate a training plan for a period longer than 8-10 weeks, since like I mentioned it appears to change its mind regarding the feasibility of the plan that I asked it to make. I actually enjoy making new training plans, so it’s not really a problem. Besides, I think the programming is more effective when you refresh your fitness signature on a somewhat frequent basis anyway.
Unrelated to XFAI, lately I’ve noticed that while Xert seems to have a very good idea of my aerobic capacity, it doesn’t seem to have much of a clue about what happens when I go much above (F)TP into VO2max and anaerobic efforts. I feel like in the past, Xert’s MPA metric used to track pretty well with how I felt, but after the last breakthrough I’ve struggled a bit to make sense of it. In the last 2-3 weeks, Xert has prescribed some high intensity sessions that have left me feeling that my lungs are on fire and my heart is about to explode. When I come to look at Xert’s analysis of the session, it’s rated as “moderate” with just some fairly minor dips in MPA. My personal recollection is that I was pretty much maxed out. I can only guess that since I’ve done only minor work above TP in the past, these 1-6 minute efforts are a weakness of mine.
All that being said, I think Xert is a wonderful system. It took a while to understand how it works, reading blogs, forums and watching videos, but I feel like the time I invested in learning has paid off hugely. XFAI has carefully ramped up my training load to a far higher level than I’ve ever sustained before, and I’m nowhere near burnout. The training sessions are enjoyable, the gains are steady and while I realize that I won’t be improving at this rate forever, I am sure I’ll keep using Xert.
I’m relatively untrained at the moment hence the long time scales and XERT has me doing Low intensity until December before any hard efforts are programmed in, so I’m currently in base. I’m fine with that, consistency over the long haul always works I find, but the 171 TL seems a bit high and wasn’t that high when I did the ‘event’ originally.
But, would it be better to split this long-term goal into shorter time-period chunks to achieve the same goal?
Thanks for sharing your thoughts! I took a quick look at your account and have a suggestion that might enhance your experience with the Xert Forecast AI programs.
Currently, XFAI sets your availability based on your usual activity time, which means the programs are designed around your typical riding schedule. It seems like you’ve reached the limits of what can be achieved within your current training load.
To see more significant improvements, consider adjusting your availability. By increasing the ‘Weekly Availability’ settings, Xert can schedule longer training sessions, giving you the opportunity to push your fitness further. Remember, the principle of progressive overload is key - if you continue to do the same workouts/same rides, etc., your progress will plateau.
You might also think about selecting a shorter focus duration. For example, aiming to increase your 20-minute power can be quite time-intensive, so experimenting with a different focus could yield better results. In my example at the start, I’ve selected a focus of 6 min power, but I still expect that my 20 min power will increase as well!
Hope that helps/makes sense. Send me a PM if you want some help with that!
It looks like your program forecasted a final of 224 W. As you drag the circle forward & back, the empty circle will show the Forecasted signature/TL on that date, while the filled in circle shows your current signature/TL values (or completed, if viewing dates in the past).
I think configuration is not bad but not optimized.
Why it’s not possible to add a .gpx (without power data) as reference ?
I know that it’s possible to other system’s.
And suggest the power reference during the race on this .gpx.
The picture shows an old programme that I did 2 weeks ago. Obviously I’ve updated my programme as it was completed in one big ride when it should have taken me 8 months…
That’s exactly one of the ways that we anticipated/hoped that people would do it - nice!
The final target TL will depend on the XSS of the event & your Event Readiness. Obviously you can complete that ride with a TL much lower than 171, which is why we added the ‘ER’ metric. There seems to be some discussion in cycling ‘Durability’, and this is how we can indirectly account for that.
If you want to read more to better understand this, you can expand below...
For an athlete with a TL of 25 (1-star) trying to do a multi-hour hard ride is going to leave them fatigued pretty quickly and likely red status for multiple days after. However, if you train more and reach a higher training status (say 50 or 100 TL), then you have more fitness to ‘buffer’ you from the effects of fatigue, if that makes sense. I’ve tried to sketch out what I mean below…
Vertical axis is form over time (as you ride), X-axis could represent time (or XSS) of your event. The longer the ride is (or more XSS it accumulates), it will cause your form to decrease. The higher your starting form is, the larger your ‘buffer’ is before you reach ‘tired/very tired’ status.
Also important to note that this concept applies to all 3 energy systems: Low, High, & Peak. If your event will put a lot of strain on your Peak Power system, then it would be advantageous to have a well-trained Peak system that can tolerate the repetitive strain - and in fact, could be to your advantage if other competitors don’t have High/Peak systems that are as well trained
Your example can be explained by the numbers not shown in your screenshot.
Namely start/end TL table deltas.
Here’s an example of an 8 watt increase to my 8-minute power before running a forecast:
The other contributing factor is hours and availability. If too restrictive and you are already training at or near your max hours/week that will squash estimated deltas.
If you can run forecasts on a laptop or tablet you can more easily view all the factors before you run the forecast and see how changes to Focus and Specificity affect possibilities and projected deltas. For example, try Polar GC Specialist or Breakaway Specialist and note the difference in possible watt increases before Achievability text changes from green.
I’m not seeing what you describe in any of my forecast tests, but I’m only defining a rest day and most days have no restrictions.
Bottom line: If a forecast results in a flatlined TL graph, there is only so much improvement you can squeeze out of it. One thing you can try is increasing intensity on HIT days by adjusting XSSR (Difficulty) under General Setting. The question then becomes whether you can handle that increase.
That isn’t what you should assume from the projections which are based on your current and recent TL history.
There is nothing stopping you from leap frogging along the way and exceeding a projection.
This is especially likely to happen when starting TL is low.
It’s also a good idea to validate your current signature before running a forecast. Simplest way to do that is ride the original fitness test workout in Slope mode. Over and done in 30 minutes with less than 10 minutes of hard intervals. Even if nothing changes you confirm where you’re at (no latent BT).
Thanks for sharing! Looks like this program will significantly increase your training load from 1.5 stars up to 3-star status, which is fantastic! Further, your increase in training load is applied to all 3 of your training loads (Low, High, & Peak), which will help prepare you for the duration of the race, as well as giving you a nice boost in your power for those short/punchy climbs you’ll likely encounter in Belgium!
Simply having GPS/elevation data is nice, but one could ride on that course in a number of different ways… for example, you could attack hard up each of the climbs, or you could drop to your easiest gear and spin up them. The course is the same, but how it’s ridden will dictate the demands of the event. That’s why Xert’s analysis is based on power data! If you’ve done the event before (or know someone that’s done it), you could use that activity file with the Race AI option to plan your training!
That makes complete sense. I’ve always found that if my event XSS is 2-3 times my current training load then I’m fine with spinning the cranks the day after and return to training. But if it is going to be 4-5 times more, then I can either take it easier and still be able to ride the following day. But if I go for it then invariably I’ll get ill and have to have 3-4 days off the bike. Training load for me at least is like a sponge, or how much effort I can absorb without getting an upper respiratory tract infection
A brand new Xert user here (since 22nd of October or so) - Had no clue what I was doing, so I started with a short plan - everything just… quite randomly set:
Things look good, had one unexpected rest day due to knee issues which immediately pushed me backwards a bit, but I think I’m still on the track to reach the short term goal. I think I’ll try some longer plan after this, with more ‘blank’ days (I enjoy the workouts, but I like doing some Zwift racing as well, and they don’t always match the planned workouts that well + after a race I’m usually exhausted enough that I wont be doing another ride just to reach the TLs)
Good stuff, thanks for sharing! Wow - you have XSS per Hour preference maxed out, so Xert is going to give you some tough targets on your high intensity days! Be sure you pay attention to how you feel - with the recovery demands turned all the way down, Xert has your permission to pile up some serious training!
If Xert gives you a rest day, you can still ride easy. It just means you don’t have to train on that day to achieve your targeted increase in fitness.
With XFAI and the XSS Buckets data field for Garmin (if you have one), you could race on a prescribed high-intensity day and then aim to top up your buckets afterwards, if needed. I think it’s something I’m going to be doing myself this winter training season
