Racing Score updates [July 8 2024]

Racing
,
1

Huge thanks to everyone who participated in the Racing Score test events and shared their feedback! After a great week of racing and diving into the data, we’re excited to bring you an update on how things are shaping up. Your feedback has been incredibly helpful and positive, and we’re super excited to keep improving the experience!

Here are the main points of feedback we’ve received and our plans to address them:

Issues with Seed Score. Particularly when we don’t have recent data or outlier data in our modified compound score calculation.
In the short term, we plan to make some minor adjustments to our prediction formula to better accommodate outliers and Zwifters with limited recent data.

Score decay seems too aggressive.
Today, we slowed down the rate at which scores decay and are considering implementing a grace period of 15 days without racing before the decay kicks in. We may tweak this value as we gather additional data.

Scores don’t move enough after a race.
We’re continuing to review the data, and as we analyze more race results, we’re ready to tweak configurations to stabilize score volatility if it turns out to be a recurring issue.

Improvements to how we handle DNFs.
As we collect more race results, we’ll keep a close eye on trends and consider potential changes, such as implementing penalties for DNFs after a certain percentage of the event is completed.

Provide a better view of score history and how it changes over time.
We are planning to introduce a view for viewing your score history. We’re already tracking full history in our backend, it’s just a matter of implementing a view of this.

Visibility of score distributions for community race organizers.
We will make this available via existing community support channels so that organizers have the data they need to set up their events.

In addition to these major points we have made some minor bug fixes:

  • Fix to race result score delta arrows
  • Update to infotip on ZwiftPower
  • Improvements to score decay

Thank you again for all the incredible feedback! Keep the racing and feedback going! We’re loving all of the enthusiastic participation and discussion. We’re going to auto-close this discussion in 14 days to keep your feedback relevant to this next phase of work.

17 Likes
3

Looking forward to the next phase.

1 Like
4

Great update, looking forward to seeing how it plays out.
Is the adjustment to the score decay applied retroactively? My score has gone up slightly when I haven’t done a test event for a week. Don’t believe any of my efforts would have impacted the seed score.

5

Today rather than race I joined a robopacer group and did a nice 30 min warmup. Then I did an all-out 30 sec sprint to establish a new 90-day max. This new 30 sec value is a blistering 6.86 w/kg which is up from my previous 6.36 w/kg :rocket:

I haven’t yet done 5 score-based races yet but this effort increased my racing score from 254 to 290 which was fun to see (for science, of course!).

6 Likes
6

i worry about pen distribution. the other day i was intending to join a prime time RGV test event, saw that my pen was empty on the sign up sheet (high B i think), then decided not to since it isn’t really a race without people to race. i wonder if other people had the same thought. the other categories were relatively well attended, i think.

i appreciate that it’s early days though

7

Somebody has to be the first one to sign up. Perhaps there were 20 people all sharing the same thought process.

2 Likes
8

A bit of a carrot for signing up early would be nice. As simple as pen position based on sign up time rather than entering the event.
If the dream of dynamically allocated pens is to come true, people need to be in the pen early enough to let it work.

3 Likes
9

i really wouldn’t be surprised, this is a social platform at the end of the day. there are probably a million reasons why many ce A pens are empty nowadays but the one that bothers me personally the most is that they are all empty to begin with… i just think “guess i’ll find a group ride or do something else”, or, controversially, “i guess i’ll find another category”

10

Not sure what the info tip is but can we have the racing score visible on results (the way current ZP ranking can be shown on one of the columns). The current racing score and also the increase/decrease would be ideal.

When are community organisers likely to have access?

IMG_1263
IMG_12632731×1271 371 KB

3 Likes
11

I always said that the best thing that could happen is to find myself in a middle of any category or a band which is exactly what I could experience yesterday with racing score event so :+1::+1::+1:.

The one feedback I would have is to maybe do a design sprint for an event with flexible assignment of a pen/band. What I mean is:

  1. Have people sign up for an event into a single pen, say “flexible”
  2. Five or ten minutes before start, sort racers per score and create 4 or 5 groups of similar size
  3. Assign the racers to given pen but these can all be in the same pen on map, just show only your group
  4. Whoever signs or joins after the seeding, would join given band. There are always many late people to sign up including me :smiling_face:.
  5. Race would start at the same time but each group would only see themselves.

The main benefit of the above is that one day you will be in the top of your band and the other day it may be mid to low-end making the race unique every time.

4 Likes
12

IRL is can sign up early and still be a the back at the start. As they say not first your last haha

13

will racing score only be using race data after a certain point of testing? It is called racing score afterall

14

I would think it would be all data so it give people the ability to race with they like.

15

if you do that you get many people who are unable to put out the power numbers in races due to fatigue from holding tempo/threshold pace and then not having the legs than if they were just noodling around with a pace partner then putting an effort in, race score should be just that, races

1 Like
16

for the most part i think it’s designed that way, but if you pull some huge effort out of nowhere after like 20 races i think it will probably account for that also. just from what i see, anyway. not really in a position to be pulling huge efforts out of nowhere personally right now

17

but you need people to get a score to be in the right pen

18

initially maybe, that’s why I said after a certain point of testing but power shouldn’t really be coming into it after a certain point if its a scoring based system, real life racing doesn’t take power into account, you get points for winning then you get moved up a cat

1 Like
19

I said that on here before be people say it would not work due to points inflation.

20

But I think Zwift can stop getting point when the racers earn enough points to keep the cat for the next season.