Competition & AI Tech Update: July/August 2022 Recap

Hey Zwifters!

Welcome back to another edition of our Competition & AI Tech roadmap update. This will cover the progress we have made in July and August.

The goal is to provide insight as to what we are working on surrounding competition and what’s on deck in the coming months.

The post will cover these sections:

  • What can you expect in the next release?
  • What is Zwift working on now?
  • What is the status of feedback the community provided?

As always, we welcome your feedback on this post! To make sure we capture it all in this moment in time, we will be auto-closing the post 7 days from the original post date.

Let’s get into it!

What can you expect from the next release?

Release version 1.29 includes updates from across Zwift, but does not include anything directly relevant to this roadmap or area of work. We do still have some news for you below, however.

What is Zwift working on now?

  • Improving fairness in racing through:

    • Implementing scoring and categorization based on race results. While we are still in the extremely early stages of this work, it’s important to bring it to light. We don’t have all the details right now, as we are still exploring how to best implement this and are only in the early phases of work and experimentation. For now, it is sufficient to say that we will likely use scoring and categorization structures based on existing models, e.g. ZwiftPower and USA cycling.
  • HoloReplay: compete against personal data

    • HoloReplay is a brand new feature coming 5 Oct that expands upon the Pace Partners idea by bringing Zwifters ghosts of their best segment times. You will be able to race against your segment PRs set from 16 Aug onward on a 90 day rolling timeframe, against the most recent full segment completion during a free ride, or both at the same time on the same segment. More details will be coming so look out for those!
  • Pace Partners:

    • As many of you may know, we’ve been hard at work throughout the summer testing many variations and making tweaks to the Pace Partners, including a new UI, new routes, and new paces. That work has culminated in a new route schedule across two worlds with 9 new paces. You can now also join your favorite Pace Partners as ride leaders in events.
  • Competitive experiences:

    • Team Trial Improvements - You can race with the new TTT experience with WTRL in their TTT series as well as in ZRL. Check it out at! Among the improvements for TTTs are an updated paddock experience, allowing teams to join in the correct order according to their start time, the ability for all riders to pedal with their paired hardware in the paddock as they would any other event, and racers can now participate in a TTT on any course in Zwift (at the race organizer’s discretion). The goal is to roll this out to all event organizers in the future.

    • Individual Time Trial Improvements - Just like the TTT, race organizers can now host an ITT on any course in Zwift.

What is the status of feedback the community provided?

  • Backlog : It’s on our radar and some work has been done
  • Discovery : Early in the development process
  • In Progress : It’s actively being worked on
  • Done : Released

Improving Fairness in Competition

  • Category Enforcement: improve the category limits to more closely match ZwiftPower - Done :white_check_mark:
  • Hardware Requirement: implement ruleset for events requiring the use of specific hardware to participate - In Progress (testing completed, needs UI work before release)
  • Hardware Requirement: implement UI informing when hardware is required and what types are required - In Progress
  • Hardware Requirement: allow community race organizers to have required hardware in events - In Progress
  • Category Enforcement: implement UI informing why certain categories are visible - In Progress
  • Category Enforcement: implement UI showing when an event is using CE - In Progress
  • Scoring & Categorization: allow race results to have a score - In Discovery
  • Scoring & Categorization: allow scored results to influence category - In Discovery
  • Scoring & Categorization: increase the number of available race categories - In Discovery
  • Category Enforcement as an option in clubs - Backlog
  • Custom category limits when using Category Enforcement - Backlog

AI Tech

  • Pace Partners: finalize pace selection - Done :white_check_mark:
  • Pace Partners: finalize route selection & schedule - Done :white_check_mark:
  • Pace Partners home screen UI improvements - Done :white_check_mark:
  • HoloReplay: implement playback feature based off personal data - In Progress
  • More nuance and dynamism with Pace Partner behavior - Backlog

Competitive Experiences

  • Time Trials: improve in-pen pre-event experience for TTT - Done :white_check_mark:
  • Time Trials: improve event join experience for TTT - Done :white_check_mark:
  • Time Trials: allow pedaling in pen ahead of TT start - Done :white_check_mark:
  • Time Trials: expand route selection for TTT/ITT - Done :white_check_mark:


A note for bugs: this update is from a limited perspective, and will not always be representative of everything affecting events, segment results, etc. These bodies of work focus most directly on competitive events, Pace Partners, and the future of competition at Zwift.

  • Crash: Zwift can crash when viewing results in a meetup - Done :white_check_mark: (1.29 release)
  • Crash: Zwift can crash when completing certain laps in free ride - Done :white_check_mark: (1.29 release)
  • Pace Partners: red squares can appear over the avatar model - Done :white_check_mark:
  • ZwiftPower: using “days to show” in filters for series breaks - Backlog
  • ZwiftPower: heart rate can show in profile page results but not specific race - Backlog

You’ve got custom category limits with Cat Enforcement in backlog…

Can I emphasize that this is very important to improving racing experience? Look at the explosion of ½ category series using closed events.

Even if this was somewhat less complex than the default categories using CP only, I think it could lower the barrier to entry of these popular races (and other formats) that are currently invites through larger teams.

This changes who has a chance to win from the standard categories, and really improves engagement for people who are solidly in the bottom half of a category.

1 Like

Agree with Craig on that one, if there is anything you can do to bump custom cats up the list. This is far simpler than a ranking system so could be a quick win.

Can you also add in there somewhere to look at wattage floors. This needs to be variable dependant on weight. Again this could be a quick win to make racing fairer.

and I suspect i know the answer but is TTT mode in the backlog for use by other organisers or will than remain a WTRL exclusive feature?

1 Like

I do wonder who sets the priorities and what is the goal behind them.

For me as a user, hardware restrictions would come much further down the line than cat enforcement changes - Now bearing in mind only the race community benefit from both changes, I don’t see any world where racers would want hardware changes prior to cat system that works or a ranking solution.

Priorities don’t seem to be customer or end user focussed.


In general A-category racers want hardware enforcement, with the “zPower Heros” blowing races to pieces.

The hardware enforcement is important for Zwift for elite level competitions as well.

Fair point, you can tell I dont race in A.

Though, even looking at that at the most basic level, the Elite level & A riders are a quarter of users, B,C, D make up the other 75% and they would benefit from an improved Cat system and race rankings over hardware restrictions.

It seems like the number of zPower hero’s has significantly diminished since the early days of zwift - I know there are plenty of spin bikes and dumb trainers still knocking about but my feeling is that has reduced significantly compared just pre-pandemic times.

I don’t have the numbers, but do those groups represent a majority of the racer base?

+1 to move up custom CE…i would put it on nr. 1 on to do list and 2nd minimum hardware requirements.

For Watt floors i know James has seen it and hopefully that will also be on the list soon.

Higher than results based categorisation?

That has always confused me why the simple things are not done. Knock off the quick wins that have big impact.

Thing like custom cats, we can already set the cats per event so all that would need done is tie that into the CE calculation system. If you told me a dev says that would take longer than a couple weeks I’d be amazed. Along with other simple stuff like get ZP to show you CE cat. Simple DB query and show on a webpage is not rocket science.

Maybe i’m over simplifying things but from working in IT i just can’t understand why things take so long

Yes absolutely as it should be easy to implement compared to a results based cat system.

Custom cats should just be tweaking CE calculation formula to read from the event the limits rather than the hard coded existing Cat limits

1 Like

The answer is dependent upon internal findings we will never know.

If it takes 1 week to introduce Custom cats and 6 months to do results based categorisation go with the quick win.
If they are of a similar length, go with the one that gives the most benefit long term.

But, you should have a product owner & BA both making those calls as to what gives the most value to the user.

1 Like

It was a reasonable question.

Flint is the PM.

This post was flagged by the community and is temporarily hidden.

As @Gordon_Rhino-Racing replied it looks like it could be done faster…but if you tell me we will get results based next month i would not bother with CE🤷‍♂️

Custom Results-based Categorization?

Agree - as someone who cant improve(bounce between the 2 cat levels on both systems) due to life/age/beer etc, racing is pointless at the bottom of a cat with no one to play with. Give me some targets half a Cat ahead of me and I’ll get interested again.


You’re right: custom category limits has been sitting as “Backlog” for a while now. It definitely gives the impression we’re not listening to what you’re asking for. That one is directly on me both within the scope of my job and as the author of these posts.

There are a couple reasons for that, and I’ll do my best to give insight to the thinking here:

  • The problem that custom category or split category races solves presently is - as I understand it - to give racers tighter competitive bands in a race by eliminating the massive discrepancies in power that can occur from the bottom power limit to the top in most categories. This is the same problem we’re hoping to solve for with a results-based system, as part of that work also includes significantly expanding the number of categories in Zwift.

  • The goal for us is to get a results based system out as quickly and smoothly as possible that would make power-based categories obsolete. To do that, any continued work on power-based categories beyond the current category enforcement will not be done, as that would not provide long term benefit against a results-based category system as it strictly relates to racing. This does, of course, mean that there is now the obvious need to ship something sooner rather than later.

  • Category enforcement is currently missing critical UI and is pretty confusing for most Zwifters as a result. We’re close to having that UI released (before the end of October barring no major issues :crossed_fingers:). The demands that custom category limits would add to UI work - design, implementation, etc. - isn’t justifiable right now given that we’re looking to move to this results-based scoring and cat system as soon as possible.

Again, the hope is to make power-based categorization in racing obsolete in favor of race placement, just like you’d get racing in real life. That isn’t to say power goes away for all time on Zwift, just that when it comes to racing specifically, your category is not determined or limited by your ability to produce power.





If the move to results categories is the decided way forward why is custom cats on the backlog and shared as a potential feature?

Especially when considered alongside the news that wider cat range comes with the results based solution so would be obsolete once that changes is brought about along with power no longer defining categories.