Are anti-sandbagging measures actually turned on in the anti-sandbagging races?

I did my first ZHQ Futureworks Anti-sandbagging race a few days ago (as usually they weren’t scheduled at a convenient time for me but it looks like Zwift has changed the schedule and fixed up the naming).

I didn’t see any green cones of shame. In my cat B race the results show 3 riders at 4.1w/kg, 2 at 4.2w/kg and one at 4.3w/kg and another at 4.5w/kg. I guess the riders at 4.1 and 4.2w/kg fall within the 95% limit for 20 minutes but I don’t see how the 4.5w/kg rider avoided any sanction.

In C cat there were 2 riders doing 3.5w/kg and 1 rider at 3.6w/kg.

In D cat there were 3 riders at 3.1w/kg and 2 each at 2.9w/kg and 2.8w/kg.

I don’t know, maybe I’m missing something but surely the 3 riders in D cat riding at the top of C cat limits shouldn’t be finishing in the top 10?

Is the green cone still active in these races or does the anti-sandbagging measures now consist solely of a warning message if you sign up to a lower cat than your FTP would suggest?

I was trying to search for examples of anti-sandbagging false positives and came across your post. A possible explanation is this statement by Jordan Rapp in a Zwift Insider article. Rapp said:

Jordan is still going to devise a trigger point that will prompt action. It will be a secret blend of both 1-minute and 5-minute power. Each of the B, C, and D categories will obviously have a different trigger point. Zwift holds this data on each racer and that’s what will be used a source.

The trigger point will be biased more toward 5-minute power as Jordan explains: ”5-minute power is a really good proxy for FTP. It’s pretty clear from 5-minute power the difference between an A or a B or a C and a D.” And the trigger point will stay secret so that racers don’t manipulate their performances to remain in a lower category. “People staying nought point zero zero one under the limit is not really what it’s meant for.”

When riders select a race category at signup, their 1-minute and 5-minute power will be used to decide if Zwift pops up a warning indicating that they risk being flagged. At least initially, Zwift will not prevent the actual signup. But they will present a warning.

I’m not saying that the system functioned well. I’m merely saying that at least as of April, 2020, the anti-sandbagging system prioritized 1- and 5-minute power. If those riders managed their power well, they may have avoided sanction.

Edit: then, browsing this forum, I came across a mention of a September post on Slowtwitch.com by Jordan Rapp. There, he said that:

But there are some significant limitations with it which is why it hasn’t been rolled out more broadly. I originally built it as an interim step to a larger system that was shelved; so now we’re doing the work to figure out what needs to be done to actually make this a more complete feature.

And now Jordan Rapp is working at Respawn, not Zwift.