XFAI Feedback

Coming soon…


Let me know your thoughts about the Event Readiness concept. This should be coming out soon and would like your opinion on it and if it works well for your needs.

1 Like

Not everyone has a key event. Some just want to train to get a higher TP or be better in Zwift races.


Choose Goal.

This looks right on. Will it guide the AI plan to generate something that gets you to that level of readiness (or warns you that you can’t make it)?

1 Like

Concept is clear, but not sure what it means in practice e.g. whether all ‘systems’ need to recover in the time frame, or only the low TL… and how is ‘recovered’ defined (form greater than X)… I’m not sure it matters if you are still yellow a few days later… though does depend on the event… :slight_smile:

Even 4 days for ‘poor” readiness doesn’t sound that bad to me… so if level 5 is 5 days I’m not sure it should be strongly recommended against? As above, I don’t really have a feel for what it means in practice though.

What would be useful is an indicative TL for a given target event XSS for each level? E.g level 1 requires TL = target… level 4 requires TL = 20% of target etc

Is there or should there be consideration of target difficulty / pacing? There’s a big difference riding a 360 XSS Gran Fondo with an XSSR / difficulty of 90 and of 60… the former is tough racing and the latter is ‘just’ a long endurance ride… which of those you choose depends whether you just want to finish or finish at the pointy end… and ability to complete is presumably linked to TL?

That’s the intent… but it is just a starter point. If you change availability or increase your recovery demands, for example, the plan may become unachievable and you’ll need to reduce your readiness level for the event.

Unfortunately, there are many variables in the entire process and it’s difficult to predetermine what can be achieved.

Need to look at this from the perspective of a beginner and not someone that’s familiar with Xert or the concepts. Need a simple way to define a target and get a reasonable plan. We could show more of the underlying reasoning and calculations during the process but that will likely just confuse. (Maybe the help document needs to be simplified even).

There are a lot of things that need consideration and we have a bigger plan to address these more directly than in the event set up guide. It’ll come after we get the goals/events working well enough. Details forthcoming…

Find an example in your history where you were red status for 5 days afterwards.

OK sure, if the definition of ‘recovered’ is ‘not red’ then that’s fine. But if your definition is (somewhat) positive form and no longer yellow i.e. to be literally be ready to perform well in an event (which is kind of implied), 5 days wouldn’t be the end of the world.

Ramblings and random thoughts –

I’m more interested in a plan for an A event that pushes me to the limit on the way there based on what I’ve done in the past plus a reasonable improvement factor, tapers for the event, and allows me to max out at the event. Post-event recovery load be damned. :smiley:

The problem will be catering to those with ABC events year-round, but I don’t think there should be more than one A event in a seasonal progression.
Weekend warriors may be best off going with XATA and an empty calendar if they don’t have any particular event/goal in mind other than a weekend group drop ride/race.

Competitive would mean maximum training effect for the event.
Recreational would mean sufficient training to complete the event; no podium dreams.
A recreational rider especially doesn’t really need to ride their event distance while training.
For example, prepping for a century doesn’t require more than half that distance in training. A 100-mile event might include rides up to 50 miles (around 3 hours for most rec riders). Half that time/distance would suffice for much of the training required.
Of course, to be competitive at an event is an entirely different matter but it’s not necessarily more distance that gets you there. :smiley: With low/high/peak strain management you can hone things to a T. FAI looks to be on the right track,
Since Complete vs Compete have different goals and requirements what about a 3-position slider between recreational and competitive for the Setup Guide? Pick event type then quantify it with a slider position ranging between Complete It ↔ Middle of the Pack ↔ and Podium Position.
It appears Event Readiness will help handle this but could also be a pre-check factor for an event forecast.

Starting and ending TL for a forecast progression --.
When initially testing Goals I ran with it for a week which raised my avg XSS/day.
Then I cleared the forecast, played around with some other options including Event then returned to Goal and replicated initial Goal settings.
Forecast ran successfully but the days forward had a bump in XSS. More than I want to commit to at this time of the year.
What about an option where you define TL you want to start out with and the max TL you’re willing to commit to with the Goal date you define. For example, say I’m at 50 TL now and want to start a progression at 30 and cap it at 90 with goal date 115 days out from now. The FAI would complete the in-between with those parameters.
Essentially ignore my recent history and start me at this level. Start low(er), aim high.
In this scenario the potential improvement would be based on TL values. The target date improvement figure will be calculated for you.
On a related note, some users are at 4-5 stars and trying to create unrealistic forecasts. If you are already riding 20 hours/week, how reasonable is it to go to 30? You are either stuck with a lateral move (little or no gain) or the most appropriate response would be to forecast a taper/detraining period. I.e. you need to downgrade and take a break for X weeks if you want to progress and peak for your target date.

Availability –
I get why the forecast runs first without constraints but that can look strange with what appears to be too many rest days especially when you’re accustomed to shorter workouts and riding 5-6 days/week. However, when you view XSS stats by week on Planner plus view the XPMC chart with the forecast progression shown, the training pattern does look reasonable.
Once you curtail longer workouts by defining daily hours (example, 2 hours max for me during indoor season). the weekly XSS distribution changes with less rest days.
Even so I’m not sure you should forecast without defining some level of availability beforehand.
What if you ask for maximums as a pre-forecast requirement? Example, max workout duration/day, max days/week, and max hours/week that you are willing to commit to at this time. Then a pre-check routine would halt the process if your schedule and TL max would prohibit the forecast from running successfully.

90 minutes indoors is equivalent to 2+ outdoors and 2 hours is the most I want to do indoors but the forecast doesn’t know that. Not sure how to best handle this other than indicate a range rather than exact hr/min duration?
May not be practical, but could also be a forecast option to select indoor only or outdoor only while the default is indoor/outdoor with a duration range shown. Indoor/outdoor range being 45-60 min,1-1.5 hr, 2-2.75 hr, 3-4 hr, etc.
Bottom line: All-indoor training for 3 months is significantly less time than all-outdoors at the same TL level.

Once you do define a granular schedule, I’d like to see a Schedule tab appear under Planner Settings where you can view the current week’s schedule in table format and scroll forward week by week to see future weeks on the forecast.
Take that a step further and the current weekly schedule could remain defined even when a forecast is cleared. Then it could also be applied to XATA recommendations without a forecast on the Planner.
IOW a scrolling weekly schedule if a forecast is in play or a single weekly template if not.

Another issue is many users are selecting event dates and targets they already qualify for.
Some aren’t taking a close look at the Setup Guide table below the sliders before running the forecast. Then they are dismayed and mystified that TP goes down but TL declined with the settings they selected. Matchsticks shown next to current/target numbers might help if simply to indicate relative increase in orange/red. Or stop them from forecasting when conditions don’t warrant one.
A pop-up could say your settings will result in a decline in training/fitness or no gain. Are you sure you want to continue? They’ll probably be cases where someone may want to proceed and create a scheduled taper period of de-training.
Or the routine could also say something like “Your current TL and fitness are sufficient for the event you entered. Would you like to switch to a Focus Power goal improvement instead?”
If someone enters a goal/event date too far into the future, pop a message explaining why it’s better to start a progression X days before the goal/event date.

Rest days –
While rest days are also recovery days where fitness gains occur, seeing a bunch of them in a week is alarming if you’re not used to it. Plus, two or three in a row? :thinking: I’ve done that on occasion this past summer and it worked well for me, but I’m sure many are not expecting that on their Planner. While the weekly XSS stats appear to be valid, it’s the distribution I question. I think that could be handled better with some pre-forecast maximums.
I would think training 2x/week is a minimum requirement. 3-5x is likely the norm. 5/6 needs a careful balance (not enough rest/recovery if too much intensity). Plus a few will claim 7 days/week doesn’t hurt them, but they’ll likely never know so maybe a forecast is a good thing. :slight_smile:

Adapt Forecast –
Looks like this is getting better with each update so I removed many of my notes on this.
Needs to be a balance as to when you may not need to run it such as swapping days in a week.
Currently a single activity can trigger the red dot alert. Should the logic be to wait for two days or more or show orange if starting to go off track then clear if X days later you are back on track or go red if not. Some may want to shuffle their weekly load and wait until end of week to decide. I suppose it works that way now as long as you are aware of the option but a gradual approach over X days could clear the dot automatically. I.e. you’re still on schedule. No need to adapt.
Or if drag and drop is added in future :smiley: support the ability to rearrange the week’s forecast without consequence.

Autogen –
A future Autogen function could shorten/extend or otherwise adjust existing workouts on recommended lists to match the XSS and strain ratio goal for the day.
IOW Autogen on steroids that uses your selection as a template. :slight_smile: It’s still an AI Generated workout for the day but the entire workout library becomes a resource for templates. Include a reality check to prevent abuse by comparing the numbers first. “Hey, this workout already meets today’s goals. No mods required.”

Okay, that’s enough for today. :rofl:


Looks good, though I would personally add some text in to include resuming training rather than event/race especially if this is used for a goal rather than an event/race. What would also be useful is to have a figure such as Level 2 - an event/goal providing 3x times on the day TL and you should be able to return to normal training without getting sick with a day of rest/recovery, Level 3 an event/goal providing 3.5-4 x on the day TL but the chance of getting sick is high/or the return to normal training will require 2 days rest/recovery. The latter is more difficult as everyones recovery demands are different and vary with the amount of riding you do, the more riding then the more resilient you become (in general).

It is how I have used TL before. I had an event where normally I would aim for a TL of 90-100, but that year I was at a TL of 75. So I knew that to complete the event and be in the top three in my age class I would almost certainly get ill as the XSS for the event in previous years to achieve that was over 4 x my TL at that time. So I could either just use it as a training stimulus and get round, enjoy the event and carry on training as normal , or go for it and take the consequences. I did the latter, ending up third in my age class and…got an upper respiratory tract infection and couldn’t ride my bike for at least 4 days afterwards. You make your choices!

1 Like

I think “Event Readiness” is easy to understand and helpful as an additional parameter. I especially like that you avoided describing wanting to win a race as an example since there are people who will never win anything and still might want to do their very best and “win against themselves”, do a personal best effort at an event, and thus might go for a readiness of 1 or 2 if they can.
Or even just want to comfortably go on a bike holiday with added weight from tent and stuff and try to prepare for that by estimating their daily XSS for that and setting that as a multi day event with no racing at all, just wanting the fitness to have fun on their holiday and not struggle to reach the camping grounds each evening.
Just another thought. I know that Xert is mostly for racing but it does work for other uses too.

1 Like

Thanks for all the great feedback. So many things we can add and directions we can go. If I could get you to boil things down into 1 or 2 things that you think everyone will benefit from the most, including beginners, that would put you into my head space at the moment in terms of priorities.

Remember that you can run Forecast AI many times. You can do “What if” scenarios. The set up guide should be seen as just a first step towards a greater understanding of what it takes to reach a goal / prepare for an event and then adjusting and then finally settling on a plan.


How can I tell it that I have a race scheduled every Tuesday?

Add the race activity (from your history) as a Planned activity using the + button on the first Tuesday.
Repeat for each Tuesday in the coming weeks that you plan to race on, then Adapt Forecast.
To remove the forecast activities from those Tuesdays, click the clock icon and set to Unavailable, then Adapt Forecast.
The red dot warning should appear on the Adapt Forecast button as you do this but if not refresh your browser page. The current beta has an issue with refreshing all components on the page when making changes.
When setting Availability, the Default is for all Tuesdays. If you elect to do that you can reset back to Available after the last Tuesday race occurs.

1 Like

Is it possible to see the gauge with the training surplus or deficit with AI beta? For me this was xert’s charm.


That is what I missed the most when I had forecast a plan.
Due to a number of things I didn’t like I have removed forecasts now and will stick with continuous improvement and be able to see the dial

1 Like

To restore the Pacer Needle you’d need to select Continuous or 120 Day Program and go with a blank calendar guided by XATA.

I don’t think the Training Pacer applies to an FAI plan. At least not in its current form.
The Pacer needle shows your surplus/deficit pts against the 12 o’clock position when compared to weighted moving average XSS over a 7-day window at your selected ramp rate.
A moving target is at play that varies based on what you chose to do within XATA guidelines (or not) compared to what you have done in recent past on same day of the week.

An FAI plan is pre-populated for the weeks ahead and includes a variable ramp rate.
Assuming you stick close to the plan you’ll remain “on track”. Essentially hovering around the Noon position if a needle were shown.
If you veer off plan (needle too low or too high) the Adapt Forecast red dot appears. Every time adapt is run the needle is essentially back to Noon.

I think most FAI users will try to stick to the plan, especially if they’ve configured Availability and have a restrictive schedule.
If wavering off course is going to be a common occurrence, another way to handle it would be an escalating red dot warning. Not sure it’s needed but…
First dot would appear at 9 o’clock position on the button icon for the first off track infraction.
Add another dot at Noon position if you move further from plan over the next day or two.
Then a third dot at 3 o’clock is the final warning. Adapt the forecast!
The idea would be to accomodate weekly changes until they exceed a limit.
You might, for example, switch up the week’s low/high/rest days and by end-of-week (or two days later) be back on track. In this case any red dot warnings disappear. No adaptation required.


Nice explanation.