Zwift Racing Development – The Latest Chapter

January 18, 2022

LAST UPDATED January 19, 2022

In my past articles looking at the development of racing on Zwift, we’ve considered matchmaking, looked at what developments to the platform as a whole could enhance the experience, and considered what fairness means and how it can be improved.

It’s been over 6 months since my last update, so let’s take a look at the latest developments!

There is a thread on the Zwift Forums containing a huge amount of discussion (2700+ posts) on what developments would really help improve racing, and how these developments should be prioritised to bring the biggest rewards for the smallest efforts. The discussion is sometimes heated, sometimes off-topic, and sometimes repetitive, but as time has gone on, a general consensus has been achieved as to what the 3 key development steps should be.

This is my take on the consensus. Please leave a comment if you feel this misses the mark!

Step 1: Pen Enforcement First

The first change is to enforce the existing categories. This means that the Zwift system will understand your category, and stop you from entering a pen for a category lower than you are assigned.

This immediately tackles the issue of sandbaggers – riders that race below their assigned category in order to do well. Whilst there are obvious flaws with the existing categorisation system, this fundamental step will drastically improve the racing experience for many riders.

Step 2: Race Organiser Toolkit

The reason step 1 is named pen enforcement and not category enforcement, is that there can be many ways to determine which riders should go in which pen – not just the traditionally-accepted ZwiftPower w/kg categories. By giving race organisers a toolkit to determine how the race pens should be divided, they have a platform to innovate – something CEO Eric Min has repeatedly mentioned he wants to see more of. Races with the most interesting formats, or races which are deemed ‘fair’, will naturally do well and be successful.

The Clubs functionality is soon to be rolled out wider, and this brings event creation to the masses. There can then be a natural progression of great race formats and leagues from the Clubs platform to the public Zwift event calendar – all surfaced via the new homescreen of course!

Here is an example of the configuration options that could be available to an event organiser when creating a race and determining the boundaries for each pen:

Watts (5s, 15s, 3m, 5m, 20m)
W/kg (5s, 15s, 3m, 5m, 20m)
Autocat (does the work for you via an algorithm of different metrics)
Age
Ranking

Separately, the organiser can set the rules on pen enforcement:

Enforced / Not enforced / Can only enter prescribed pen or higher

After determining the metrics to be used, the race organizer can set the boundaries – so there is no longer a consistent category system. What this means is that pens are dynamic and not static, which ends the problem today of ‘Cruisers’ – riders that deliberately manage their power input to stay in a particular category where they do well.

Step 3: Global Ranking System

You may have noticed that in step 2, one of the inputs that could be used by a race organiser is ranking. A global ranking system would be the primary racing reward system, as you try to improve your ranking in every race.

Thanks to the new dynamic race formats from step 2, in some races you may be one of the strong riders, some races mid-pack, and sometimes at the bottom. This differs from today, where riders tend to consistently be either strong or weak within a category. The ranking system will mean that even if you are dropped early in a race, you may benefit from a ranking improvement if you beat the riders around you.

There is already a ranking system on ZwiftPower, as Eric has written about recently. However, it is quite basic, constrained by the existing categories, and does not deal with new riders well. There are a number of solutions available which are used in the wider gaming industry that could be used, including Microsoft’s TrueSkill or the iRating system used by iRacing.

This ranking, as the primary progression metric, should be attached to a user’s profile, and can also be used as a way to sort pens if the race organiser chooses to.

Zwift HQ Interactions

Zwift have openly recognised the need for development in this area, and have agreed with the community that pen enforcement is the first step.

A recent post from Flint on the forum has explained what is coming and when:

We’ve got some dates now. I’ll have more details for you all in the coming weeks, but here is what I’ve got for you so far:

When: week of 21 Feb (still tentative, but looking good for now)
What it will be: 3-7 days of single events, with relatively easy courses with the explicit goal of testing the beginning of broader race rules starting with category enforcement. In other words, putting a floor on categories you can join below a set minimum, based on power curve data.
What it won’t be: Autocategorization or matchmaking. It is also not a GC event.
What we’ll need from you: Feedback! I’ll have a form made up and ready to go as well as an in-depth post explaining the finer details.

Two things I want to call out specifically here:

Simply enforcing a floor on categories may seem like very little after so much time has passed. The reason for this is called out in the above bullets: this is the foundation for broader rules that can be applied to events when needed. We’re not just building category enforcement.

The values we will use and the ways we can enforce categories are not permanent, these are merely what we feel will be the most beneficial ways to implement this in an early version to produce meaningful feedback, data, and – most importantly – race experiences.

It’s great news that pen enforcement is on the agenda, and we have a date for when we will see something we can test. It’s also great that clearly they don’t see the categories as permanent or static. Something that I am passionate about personally, is that race organisers should not be constrained – allow for flexibility and configuration, and you will be amazed at what the community can create.

Zwift also has an open position for someone to head up racing development, and there are some interesting points in the job description:

“We are looking for a highly accomplished Lead Game Designer to expand & evolve Zwift’s racing & competitive features. The ideal candidate has extensive experience designing PvP and Team v Team game systems including stats tracking, matchmaking, rewards progression and leaderboards, and has tuned these systems through ongoing live operations. We’re looking for someone with proven leadership skills and a highly collaborative approach in partnering with all disciplines.”

Is all of this moving too slow? It’s certainly been a long time since we have seen any real improvements to the racing experience on the Zwift client itself, but at least something is happening, and we have a date for when we can see the first fruits of some labour.