(Eric Eddy) #1

Some how manipulating the data, extra speed magnet or no trainer resistance, etc.

Maintaining 11 - 15 W/kg for full laps. No world class international pro can sustain those numbers for more than 60 seconds, let alone for 5km.

(Christian Wiedmann [X] 50) #2

It could be unintentional. I’d give benefit of the doubt before calling somebody a cheater. They probably experienced some kind of bug in their setup or in Zwift.

Plus I don’t think you can ever prevent someone from cheating in a less detectable way. I suspect I’m “cheating” because I’m using a KICKR without a power meter, which by all reports is generous on the power reporting on longer rides.

I think the only solution is to primarily compare yourself to riders you know and trust. Ultimately, I think there is only a social solution to the problem of cheating, not a technical one.

(Mark Williams) #3

I’d be a little reluctant to call “cheat” when we’re using a beta. There’s always the chance of bugs here and that would be my first guess - mainly because the idea of anyone wanted to cheat in a closed beta where all the numbers are qualified anyway pretty much defines the term “loser”. Seriously, I’m hoping / assuming these may be bugs.

As more constructive feedback for Zwift, I was thinking about this recently. On my first lap recently I kind of went for it a bit. I must have lucked out (all the decent riders were offline) because I came in second for the lap. Well, that was until I noticed that the leader had a time of 37 seconds :slight_smile:

I wonder if, beyond just continuing to fix bugs, Zwift might consider setting some sort of check for any one leaderboard/segment, whereby any times logged that are realistically too low to have been achieved by a human are just thrown out. A 37 second lap should be pretty easy to detect as “questionable” :slight_smile:

Just a thought.

(Mark Williams) #4

Pretty sure no-one is saying these numbers are realistic so no “proof” really needed there. The point is less about whether the numbers are realistic and more about whether they could simply be down to bugs, which seems more likely to me, at least, than intentional cheating.

I’d just prefer to give people the benefit of the doubt.

By the way, if it is cheating then I just feel sorry for whoever goes to that trouble. Talk about missing the point! :slight_smile: Doesn’t bother me beyond that, personally.

(Clint Westhoff) #5

@Eric: “A Kurt Kintetic is about 30% more generous on power that a CycleOps PowerTap hub.” No it isn’t - it can be, but your absolutist statement is not correct:

There are obvious examples of things being amiss that could be: 1) conscious cheating; 2) user error; 3) bugs; 4) someone just experimenting for fun not realizing it offends you; 5) something else.

But, I personally would be reluctant to accuse someone of “cheating” without some clear evidence that it was a conscious intent to “win” (what exactly is someone taking away from you by winning by cheating in this environment when we don’t even have organized races yet).

Also, just from watching power estimates from trainers, TrainerRoad, TourDeGiro, etc. grow up, I have seen many “cheater” allegations turn out to be just faster people. I also see a lot of people who equate virtual power with inaccuracy even though it can be just as accurate as a power meter.

(Eric C. (Zwift HQ)) #6

So hey, let’s talk a bit about gaming numbers:

We’ve seen a lot of ways people can game numbers, especially with the introduction of zPower. This will always be a factor because we rely on the user to honestly tell us and the program which trainer they are using. We are also asking them to inflate their tires to a maximum of 100psi (and no less) and to tighten their trainer properly.

These are all factors in making sure Zwift is as open and welcoming to every person who is interested in having a less tedious indoor ride experience. It’s how we address potential bugs, user errors, and, yes, cheaters that will pave the way forward.

For example, someone with rollers whose power meter is being repaired (actual situation) but doesn’t want to miss their indoor ride might pick “trainer unknown” and the zoom around the island with an average 800watts. They may not even keep that data but want to have their legs moving for an hour for exercise.

Is this person cheating? Should they be allowed to ride? Or should they suck it up and stick to another program?

The idea is find a way to figure out when numbers are unrealistic or to make these kinds of numbers irrelevant. Ideas have been floating around the office, including have a threshold time - but what’s the threshold? And who makes it? And do we update it if/when it’s beaten and we think it’s feasible?

We’re working on answering these questions and trying to make an open and welcoming platform for everyone. Thank goodness we’re in Beta! :slight_smile:

In the meantime, we’re going to play it conservatively when it comes to calling people “cheaters” and we may eventually have a way to flag suspicious rides for review. It’s an idea and we’ll all be moving forward towards a solution together.

This is a learning experience and, yes, we have access to all that data as well. It’s not like we’re ignoring it but we’ve got lots of fish to fry in trying to get Zwift up and running :slight_smile:

(Karl Litterer) #7

I witnessed a rider putting out 16.4W/Kg for several minutes last night with “Virtual power” (no lightning bolt). Got a few screenshots of it and then I called them out on it with a “M- Artificially high Power username???” when they got close to me. Magically their numbers came back in line after they stopped for a second or two. Guessing that was when they turned the resistance back on.

One way to deal with flagrant abuses on virtual power, take all the ride data being stored on the server, and create a generous rider maximum W/Kg/time curve. If a rider exceeded this curve, the user could be notified that power was unrealistic and power could be reduced by a % based on how far over the “maximum” curve they were for the remainder of the ride.

(Mark Duncombe) #8

I have a fear that people might start taking Zwift to seriously getting obsessed with the numbers and things start getting toxic.

If people want to cheat then I don’t see how you can stop them completely and keep Zwift accessible/affordable to the majority. Heck, I am pretty sure I could spoof the ant+ data coming from the sensors, making it look realistic while sat on my sofa drinking a beer and eating pizza but as another poster said that’s missing the point.

At the end of the day for me at least Zwift is all about making winter training more enjoyable. so I will be using the following strategies

  1. Comparing my metrics only with my previous metrics to monitor my progress.
  2. If I do “compete” I will only care about the results of those I know and trust.
  3. Not take it too seriously
  4. There will always be someone faster than me, cheating or not, so I am not going to let it get to me.

Regrading #2. In Strava, I only care about my times and have some friendly competition with the people I follow. So one way to avoid “cheaters” is to have some sort of group/friends/followers structure in Zwift to filter out the randoms.

(Chris Ashley B) #9

I saw a 11-14 W KG for the entire duration of the hill the other night… looking at Andrew Coggans W/KG charts now, world class top of the list is 11.5 W KG for 1 minute… Oh well… do what ya want I guess…

(Chris O'Hearn Masters (E)) #10

+1 to Mark D’s reply.

It’s a game.

Maybe there will be Chinese Zwift factories pumping out miles and getting points like in Minecraft or other MMOs. Whatever - at the end of the day it’s about what your own legs can do.

I think in a competitive sense Zwift will not work at an all-player level. It will be great for riding with people, socialising, training etc. but for any competition there will always be loopholes and ways to game the system.

The only way I can see competition working is if there is an ability to set up mini-leagues or groups where entries can be controlled and someone is responsible for vetting riders and results.

It would be up to people to come up with their own categories - it could be age, weight, geography, discipline, race category etc. and the ones that work best will be the ones that people have most trust in.

Maybe they can be verified off-line through email/photos/power files etc or something like that. I would suggest that the Zwift developers need to give the community the tools to do that not try to centralise it themselves.

(anon18154799) #11
  1. It’s a game. People will try to cheat/game the system. Why? Who knows. Fundamental problems with his/her self-esteem, likes messing with people for no reason, etc. The developers have said that we will likely have rooms/events where you can designate verified power only or smart trainers.

  2. It’s a training tool with some fun perks. Sure, I like getting a jersey. However, it doesn’t ruin my day if I get beat. M. Desrossier has been doing that in the Sprint every time I get on. I do my workout and see where I can stack up against the other confirmed power folks.

  3. Relax, Francis. The good people at Zwift have given us the opportunity to view and play on the system before it goes live. And, we get to do it for free! How awesome is that? Let’s keep things in perspective. The bugs/“cheaters” may annoy us, but it’s not costing me a paycheck. Worst case, I can’t brag to my Strava buddies that I was the fastest Zwifter today because a virtual power person tweaked his/her system. Does that matter anyway?

(Kelly Latham AHDR (C)) #12

This discussion in general is one of the reasons i think it would be more useful have a setting/option of showing your own PB’s/records. Instead of worrying about what other people are doing you can concentrate on beating your own numbers. I’ve seen some incredibly unrealistic figures being achieved by riders almost ever time i’ve logged on, and it annoys me too.

(Craig Martin) #13

It’s an interesting challenge to take data from all the combination of power meters and trainers you support (and some you don’t) and try and resolve that to game play that mirrors real life. Some riders have spoken about using a couple of different set-ups, how does your virtual performance and numbers vary across these. It’s fun to take the jerseys but, as with Strava, reality intervenes when the big dogs appear and my KOMs don’t last long. That’s pretty true to life :slight_smile:

(D W - ltb LMDGD) #14

Disclaimer: If you see me riding around the island at a virtual power level that doesn’t match my Zwift 40k Holiday Challenge time that’s because I don’t have the option to specify rear wheel size.

I have to be honest that it irks me that every time I ride the KOM’s are quickly populated by someone in the handful of virtual power riders with times like a hill climb in the 40’s, a sprint time nearing 7 seconds flat, and a lap time much faster than Jens. Fortunately, we’re in Beta and with additional trainer profiles, bike settings, and whatnot, honest people can separate themselves from the not so honest people. But what do we do about the 26 lb guy doing full laps at 11 W/Kg and sprinting for extended periods at 1999 Watts? Banish them to Cheater Island?

(Chris Ashley B) #15

About 3 weeks ago I was having a little battle back and forth on the green jersey sprint with another rider… we traded off a few times and then I got a high 8 second sprint… his next sprint was in the 7s… I was thinking WTF… next thing I got a message say “I used the aero boost sorry” I could really care less… but if they want to have actual “races” on here… the boosts need to go, and a resolution to the weight needs to go as well… that will be difficult…

(Patrick Tan.Ascenders (PTz-ZSG90)) #16

I do not suppose we need to take those numbers seriously as we ride to our capacity and known data with that level of personal performance.

Cheating? Having experienced a certain degree of spiking in results from segment toppers in Strava, I recommend we look beyond these “cheaters” since they can live with the fake ego boosts, while we just get along with a closer to real world experience.

Let’s just enjoy the positives from the efforts of the ZWIFT Team. Pity the cheap thrills…

(Joe Gerbrick (Gettysburg, PA)) #17

Thank you for providing that official answer. Consider this - there are millions of cyclists throughout the world who ride everyday in their own environment based on their capabilities. No two environments will be exactly the same, even on group rides. Thinking of Zwift Island as an extension of my own unique environment, capabilities, and goals helps keep results in perspective.
The competitive nature of many cyclists will magnify result variations, especially on measured sections of road. Zwift, by far, provides the most enjoyable indoor experience I have seen. In my limited experience, I look for cool ways to motivate me such as closing the gaps. This is a virtual motivational environment that I do not take too seriously, with primary focus on bridging the winter gap. I’m confident that the anomaly in results will be sorted out in the data over time. In the meantime, let’s have fun and be thankful for the opportunity to participate in the beta. There are still many folks on the wait list.

(Jeff Doughty PVC WBR (B)) #18

Definitely having mixed feelings about this subject.

I was a bit bummed today when I broke the 8 second barrier on the green jersey (powertap, correct weight entered)… I was beat by .45 seconds. No biggie, there are certainly faster riders than me!

A few minutes later when I began the polka dot jersey this guy flies by me at 16 watt/kg and holds it for the duration of the climb. He get’s the polka dot jersey at 42 seconds… Low and behold its the same guy who won the green jersey I nearly missed. At first I am thinking yikes this guy is strong. Then as he stops the hill climb he holds 6 watts/kg or so for a lap until he is off my rider screen.

Can I say this guy was cheating for sure? I don’t know, it certainly seemed more than suspect. Especially considering he was using zPower.

It was certainly a bummer not to get the Green Jersey… Considering most of the time I don’t go for these things…

Now, ultimately I don’t care. My ride was my ride, and my numbers don’t lie.

What I really want to see are private rides. Rides could have requirements like Powermeter only, Only riders the creator allows, or Rider weights are shown in their profile and the ride creator can flag a rider or something if cheating is suspected.

As far as the BETA is concerned… let the cheaters cheat. There will always be ways, be it entering false weights, powermeter calibration, using inappropriate trainers… etc…

Luckily strava tends to separate the boys from the men (or girls from women) If you don’t upload I don’t trust you much, and if you do and your outdoor rides are weak, or your power is super low I just remember you name and laugh as you ride by.

I especially like the 20 watt/kg lap done by a guy on my ride Monday night. Settle down Francis!

Happy Zwifting all!

(Chris O'Hearn Masters (E)) #19


Same thing happened to me. Did 400w up the hill then got beaten in nearly half the time… some bloke called E.Min.

Got his name flagged for sure :wink:

(Mark Duncombe) #20

be aware of the following

what appears to be some one cheating could be

  1. a bug
  2. someone frustrated with hearing how great Zwift is an using an profiled dumb trainer just to see what the fuss is about 3.Someone who accidentally has set things up wrong

lets not get hung up on this and end up with a power meter using elite master race looking down on those who cant afford/justify the significant cost of a power meter.

and BTW, having a power meter does not mean you cant cheat. Ant+ is a relatively simple protocol and you could fairly easily hook up something like an Arduino and a RF module to broadcast Ant+ power data. Take it a stage further and you could receive ant+ power data and rebroadcast it suitably pumped up.

As more dumb trainers are supported I expect #2 to decline as it could just be people trying it out with unsupported trainers. it wouldn’t surprise me if much of the so called cheating was as a result of #2 rather than intentional cheating.

if/when BLE is supported by Zwift then other hacking options become available particularly as the BLE community is much larger than Ant+. I’m a BLE engineer BTW and could easily massage my BLE sensor data if I put my mind to it to create fast but believable data.

So if someone wants to cheat on zwift then you cant stop them, zpower or real power meter but they are not cheating you/others only themselves.

I also suspect Zwift needs the dumb trainer crowd to make Zwift commercially viable.

Chill out, ignore the occasional crazy power readings you see from others, build a community, make it social and have fun.