Zwift (wrong) ANT+ protocol for Power and Cadence sensors (not FE-C)

Dear Support,

I found a possible bug in the protocol what Zwift is using for recording the sensor datas. So as i know, there are 2 types of the ANT+ protocols, a „display” and a „record” one and i assume the Zwift is using the “display” protocol for data recording (instead of the „record” one). The problem is with this “display” protocol that if the data value is suddenly dropping to 0 (for example when a rider is stopping the pedaling) this “display” protocol is correcting data (as error correction) with the last known value (the “record” protocol is not doing error correcting) for 2-3secs and the recorded data will be fake.
So if a powermeter is paired to Zwift via ANT+ as power source and/or cadence source there will be fake values for 2-3secs and especially with power will cause totally different avg results (and the racing is pointless becasue it is an accidentaly cheating). If the trainer is used via FE-C protocol (so the power, cadence and controllable are the same) this problem is not exist.

Here is a few comparisons which helps to understand the problem: (Neo2T’s data recorded via a Garmin head unit as power sensor, Quarq’s data recorded with Zwift PC client, Quarq as power source in Zwift, Neo2T as controllable)

proving that the Quarq’s data is good here is a comparison when the Quarq’s data also recorded with a Garmin head unit , clearly see that the Zwift’s recorded data via ANT+ is totally fake due to this wrongly used protocol:

the situation is the same with cadence too (ok, that is not a marginal problem, the user can’t cheat during a race with fake cadence data ) here both cadence datas are coming from the same source (P2M NG via ANT+), under the name with P2M NG is a Garmin FIT file, under the name with Tacx Neo is a Zwift FIT file with P2M NG cadence paired as cadence sensor in Zwift, so only the recording method is different, clearly see that the Zwift is modify the data as “error correction” with this “display” protocol.

just to be clear, i found this problem with many powermeters, Quarqs, Assiomas (there are lots of comparisons on ZwiftPower luckily) so not a single case :slight_smile: i dont know how is the situation with BT protocol, i never used it .

Can somebody explain it from Zwift developer team?
i’m also sending my problem via email to the support .



ibb . co/wBt1ZgN

delete the blank spaces from the link (behind and in front of the “dot” ) and will be good