Forgive the derisive title, it may be a little inflammatory but nevertheless, it has become a stock response for our support team when they are asked questions like…

“why do I stop pedalling when I go around corners”

“why do I brake so I don’t hit other riders”

“why do I brake when descending down the Stelvio Pass”

We know, we all get it, we really do – No one likes change, it’s scary, but in this case, change is essential for cycling eSports to progress. 

In reality, cycle racing is exciting because the courses offer riders different challenges based on the routes. Up until recently, cycling eSports has failed to capture this. It has been (at least for the most part) a threshold effort with a sprint at the beginning and end, and not much else.

It’s boring, it’s for testers (time trialists) and the guys who’s physiological systems are better placed to deal with those physical demands.

This isn’t good enough and it’s making the sport boring to watch and even more boring to participate in.


In the minds of the team at RGT Cycling all of the above outlined the need for change. So what are we doing differently to change the game?.


Firstly, we worked hard from the beginning to create a physics algorithm that was based on real-world research. It took us a long time and 1000’s of iterations and we’re still not done with this, but we’re getting closer to a hyper accurate representation of the real world speeds and movement.


Obviously, in the absence of steering we don’t want you stuck in the stickiest of drafts unable to get away from your opponents in the sprint. However, we do want to simulate the need to move up to the front before long climbs, or big sprint finishes etc.

We also want you to be able to sit on the wheel of your team mate without constantly overlapping them.

To simulate this we have added braking to avoid you hitting the back of other riders. This has been crudely implemented elsewhere, and is sometimes referred to as sticky draft. 

On RGT you can see it happening as your power value turns red indicating the brakes are being pulled.

You can overcome this by pushing significantly higher power numbers than your opponent or by approaching them at a significantly higher speed. 


This is one of the most essential developments we have made in providing a more realistic racing experience

To recreate the concertina effect experienced in real life racing when cornering, we have simulated breaking into corners. Sometimes this might seem slower than you would like but it’s based on angles, forces, the line the rider takes, and pedal strike considerations to workout how slow the rider should go.

Corners, as with real life racing can become tactical opportunities to launch attacks from.


We know from our own personal experience as riders that our normalized power values are more realistic when racing on RGT, when compared with racing on other platforms and this is a good indication that we are on the right road to more exciting racing for virtual cyclists.

All things considered, we’re working in a complex area, delivering on our commitment to realistic and engaging eSports for cyclists but we’re a long way from finished and have features in development that add even more to the experience.



The RGT Screen App Drafting interface is broken down below:

Drift Back – This appears in the power HUD when power gets below the lower required amount to sit on the wheel. You will start to get dropped if you’re seeing this.

Pull –  This appears in the power HUD when power gets above the upper level required to sit on the wheel. You will start to come around the rider in front of you at this point.

Saving – This appears in the power HUD and shows the amount of power you are saving by sitting in the wheels.

Arrows >>> – The power value shown here is the value you need to maintain to stay in the draft.

Arrows <<<  – The power value shown here is the value you need to reach to come around the riders in front of you.

If you want to stay in the draft, stay between the two numbers next to each arrow, e.g 

Taking the image as an example 170 watts (the halfway point between 100 and 240) would be the optimum power to aim for.

The power turns red when automated breaking gets applied.


You will not get pulled into the draft at speeds lower than 16 km/h. (If there’s no room to overtake then drafting is forced regardless of speed).

Hopefully this is helpful and goes someway to giving you an understanding of why we’re doing, what we’re doing.

Got a comment on the physics or want to make a suggestion for the future? Please let us know in the comments below

41 thoughts on “PHYSICS ON RGT

  1. Nikolaus says:

    overall i think you have made a great job in rebuilding drafting physics!
    However there is one thing where I experience differences on RGT vs real racing. If you have a large bunch of riders you experience a significantly larger and distance wise longer lasting drafting effect than compared to a small group or a single rider. Correct me if I am wrong but I did not experience this on RGT so far. My impression is that the algorithms picks the rider were the speeds are matching best and just takes his/her slipstream into account. And if the algorithm chooses to change the rider to draft in a big bunch you see a massive drop in W savings for this changing period which again you would not see in a big bunch in real life.
    Keep on your good work and keep on improving! 🙂

    • adminjames says:

      Thanks for the reply 🙂
      The number of riders in the group does make a difference to the overall saving of any rider in the draft. However, we certainly agree that this can be improved.
      We’re working on our drafting and refactoring the code at the moment. We hope to have improvements shipped sometime in the new year so keep your eyes and ears open.

      We still stand by RGT Physics as the most realistic.



  2. Andrew Fraser says:

    On the flat will an 80kg rider doing 4w/kg, i.e. 320w go faster than a 70kg rider doing 4w/kg, i.e 280w and if so by how much. Our club created races on an essentially flat circuit and it appears that everyone goes the same speed for the same w/kg which in reality is nowhere near correct on the flat.

    • adminjames says:

      Thanks for the reply 🙂
      There is a difference between the actual wattage produced and the speed of the rider. W/KG also influences this.
      We’re working on our drafting and refactoring the code at the moment. We hope to have improvements shipped sometime in the new year.

  3. John Martin says:

    I have found when riding a route on RGT that I have ridden in the real world the elapsed time is very similar. In some respects it can feel harder but is often helped by being able to draft a Bot. The net result seems pretty accurate to me. If you adjust the drafting physics as mentioned above I fear that the overall difficulty may then become overstated. Great work guys. I am really enjoying using RGT.

  4. Gabriel says:

    Sorry for not agreeing the principles of automatic braking. The rider can’t feel any physical stimulus that could trigger braking (or at least stop pedaling) during a virtual cornering or descent, like he feels in real world. There is no feel of centrifugal force, no bike tilt, no slide slip and no crashing in virtual ride. Virtual automatic brake just make the rider loose his pedaling effort because he has no physical stimulus to tell him stop pedaling, like in real life. I think it should be removed, releasing resources for other computing challenges.

    • adminjames says:

      Hello Gabriel
      Thank you for your message.

      We disagree entirely – But that’s okay. We’re not always going to agree with everyone!

      There is no rider stimulus such as forward momentum or acceleration provided by pedaling, but we think this is essential too.

      We’re fully committed to this feature and will continue to enhance it.


    • Gerald Hines says:

      I completely disagree. I stop pedaling and slow down ahead of corners in the real world because I SEE (physical stimulus) the corner I’m approaching and know I can’t just go flying through it without crashing. I stop pedaling ahead of corners in RGT because I SEE the corner I’m approaching and know that, due to automatic braking, I’m going to waste watts if I keep pedaling.
      It has the same effect, and I prefer it to just letting me keep all of my momentum straight through a corner. That’s not just realistic. And RGT’s physics results in me having to put in more effort to get back up to speed out of the corner, just like I have to do in the real world instead of getting the “free energy” ride through a corner like I get in other apps.
      I want my virtual ride to be an actual replacement for my real rides, which is why I don’t care for Zwift. All of those free “power ups” and whatnot aren’t what I’m looking for in a workout.

      • adminjames says:

        Hey Gerald
        You’re right “RGT’s physics results in me having to put in more effort to get back up to speed out of the corner”, but that’s the point.
        It’s that simple. RGT is a supplement to real-world cycling, a racing simulator, and a training tool. If you want to improve the same physiological systems you use when riding outside you need to exercise them. It’s that simple, and that’s why you need to pedal harder to get back up to speed.


    • ChrisH says:

      Agreed, Gabriel. The algorithm that controls positioning seems arbitrary and often makes counter-intuitive decisions as well. It can be frustrating to have elements of the program that directly affect speed and position within a race be so important but also be entirely out of your control. I’ve raced a couple of times now on RGT, and while I like the idea of making it as realistic as possible, the execution leaves a lot to be desired. I think for the idea to work (well), there has to be a way for the rider to control their own position. The in game mechanics simply don’t make race-intelligent decisions. There is nothing more frustrating than to put on power and watch the algorithm try to slot you in behind another rider you are trying to pass, then hit the brakes while there was plenty of room to have passed. It is useable and I suppose in some regards better than other platforms, but still has a long way to go approach “realistic” in my opinion.

      • adminjames says:

        Hi Chris
        Obviously sorry you’re feeling this way and we’re taking this on-board. We may allow organizers to select to turn collision avoidance on or off. Making this easier for those who are not interested in learning how to ride.

        The in-game mechanics make the same decisions for everyone. Those who learn this are able to position themselves accordingly and make the most out of their power in a race.

        It’s because of this races are won and lost on race craft and athletism, not a random power-up or boost after a standard TT Effort.

        We’re certainly making this better but we’re never going to stop braking for corners. It’s like saying “Arhh guys, I don’t like gravity, can you switch it off?”


        • MichaelG says:

          I love the attempt to provide a more realistic ride experience. Although I believe the secret to achieving this goal is in the immediacy of the feedback to the rider.

          Having to look away from the course and racers to see if the brakes are being applied or figure out what wattage is optimum to stay in the draft or determine if the upcoming corner will require braking or not hinders the riders ability to make good split second decisions.

          In the real world you can feel your speed and more easily visualize the approaching corner and determine if you’ll have to let off the gas and by how much prior to reaching the corner. In RGT you have a much more limited feel of the ride and depend on the on screen data to help you make decisions.

          With that being the case the ability to make split second decisions comes down to optimal visualization of the data. To achieve that it would be useful to see braking visualized on your avatar (red glow or something) or if an upcoming corner will require you to brake then provide some visual clue on the road ahead where the braking will occur. That way the rider can make decisions without looking away from the course or the other riders.

          The same thing with drafting if the rider could tell whether they needed to increase or decrease power to stay in an optimal draft without having to look away from the racers it would allow them to make better drafting decisions. One thing I don’t see often in RGT is a strung out line of riders drafting or a pace line trying to pull back a break. The chase usually yo-yos itself along as rider exert too much effort then let off too much. In the real world it’s pretty easy to pace line a chase. You could also implement Team Time Trials if riders could better maintain a consistent draft. Don’t forget to add in the elbow flick gesture to let riders know you’re done pulling and to come around.

          The beauty of all of this is that all the data already exists it just comes down to providing a different way to visualize it. If you need incentive to implement modifications like these then only make them available to paid subscriptions. Which may increase your subscriber base.

          Keep up the great work…

          • Tim Sinclair says:

            Have you tried riding with the ‘first person’ camera view – it helps a lot with visualization 🙂

  5. Scott Frey says:

    Thanks for this helpful information and for the platform, which I am really enjoying. The community spirit of RGT users is truly outstanding. One the topic of physics, are you modeling air resistance as a function of riders’ velocity?
    Happy holidays

    • adminjames says:

      Hi Scott
      Thanks for your kind words. The team will be thrilled to see more feedback from another happy users.
      I think what your asking is, does the drag and draft change based on speed and the answer to that is yes.
      Thanks again

  6. Gabriel says:

    I’ve done some more riding, solo or by drafting a bot. I really liked the idea: higher wattage with the bot in your draft, then “rest” by drafting the bot at lower wattage – the perfect team :). But I noticed that if I try to corner just around maximum allowed power (power figure continuously switching from red to white and back), the bot slightly tends to drop me (he knows the game algorithm better than me, doesn’t he?). So, in the game, the rider has only 2 alternatives:
    1) Corner with lower power than maximum allowed and ride below maximum possible speed, or
    2) Corner with higher power than maximum allowed, which means preserving the maximum possible speed, but lose power in braking.
    In real life, alternative 1) is possible, but 2) does not exist: it means just hard crashing. But in real life there is also the third alternative that the rider can never apply in virtual riding, because he has no stimulus to reach it: take (almost) the best position and power for cornering at maximum speed.
    Thus, I am not sure that such feature in the game can reasonably mimic reality, even in a perfect application.

  7. Gabriel says:

    P.S.: Overall, I think RGT became a very good simulator, maybe missing (just for the moment, I hope,) the critical mass in number of riders. Btw, I have just leaved Zwift, mainly due to a very poor implementation of drafting and, probably, poor synchronization of the real position of the rider on the screen.

  8. Andrew says:

    The values that are displayed next to the “drift back” and “pull” arrows within the drafting interface are very hard to read owing to them being a faint shade of grey.

    Just a suggestion, but by having the values as green (drift back) and red (pull) would make them easier to read.

  9. Sylvain says:

    Dear RGT Dev team,
    I really appreciate the work you’ve done to have a more realistic plateform in terms of physics, braking and drafting. I definitely think it’s the way to go for eSport competitions.
    However, I’m not yet entirely satisfied with ho the braking is applied in some corners.
    It looks like that the braking is from time to time to strong or shouldn’t be applied at all in regards to the speed of the avatar and the curvature or the turn.This is, for exemple, the case on the last turns of Borrego Springs before the finish line.
    This is maybe due to the fact that the trajectory of the avatar is not really realistic in these turns also. The avatar is like fixed on a rail. We would expect the avatar to try to cut a bit in these corners on a large road for a more realistic feeling and I imagine that it would allow for less braking as well.
    I hope you keep up working on these aspects that are for me really important if you want the riders to understand and adhere to your strategic choice towards more realistic virtual riding.
    Keep up the good work!

    • adminjames says:

      We’re very happy with how the product has progressed but we agree entirely, we’ve still got a long way to go!
      We’re working on all aspects of the physics and hope to improve this with a new release we have planned this year.
      #ridefree 👍

  10. Nate Fulcher says:

    So I have a slight issue with the physics….I’m heavy weight which means in real life on the flat and downhill my weight and watts are an advantage which don’t really seem to come into effect here, especially if trying to get away from the light weights downhill pretty rubbish after you kill yourself uphill to stay with the bunch hope for a little rest on the downhill and then get dropped but a guy half my weight…. can we please have our gravity back!

    • adminjames says:

      Hi Nate
      Thank you for the message. You do have your gravity but the overall savings for big groups on descents is too high meaning you will get brought back.
      We know about this and hope to improve it with the next update to physics.
      #ridefree 👍

  11. Tim Davies says:

    Today I noticed huge drafting watts on the fast descent in race conditions, over 1200w at times. The punching thro’ when riding in a fast bunch I can understand created this high ‘demand’, hence the high draft. I cannot actually exceed 800w myself, so a draft of 1,2kw is quite meaningless to me. The other effect is a near instantaneous being dropped if you a distanced by a few metres, and huge power demands to get back on. Real world descending is not so extreme, and the drafting extent not a fixed value, i.e. as speed increases, so does draft. Also we descend in lines behind each other, in RGT it tends to parallel up, not sure why.

    • adminjames says:

      Hi Ian
      The savings calculation includes everything at the moment, including the savings from a descent. We know we need to change this.
      Thanks for the reminder.
      #ridefree 👍

  12. Richard Neil says:

    Newbie to RGT & loving it! I was reading ‘Drift Back’ & ‘Pull’ as instructions TO me in order to maximise drafting. I thought that when ‘Drift Back’ appeared I needed to ease off & when ‘Pull’ appeared I needed to apply more power. Reading the above I believe I have got this the wrong way round . So the game is telling me what’s happening – EG ‘Drift Back’ is when I need to add power to stay in the draft (as I’m drifting back) and ‘Pull’ is when I’m adding more power than needed to stay in the draft? Have I got this right?

    • adminjames says:

      Hi Ian
      Yes! This has been mentioned before.
      When you are being dropped it should say “Pull” and when you are pulling around the riders in front it should say “Ease Off”.
      I will add this to the list.
      #ridefree 👍

  13. Peter says:

    Moved to RGT just after Christmas and I’m definitely a fan. Question. I’m 62kg and it seems I have to work much harder then I used to on another platform! Am I right? Not complaining but often my W/kg is so much higher even when I’m paying close attention to the draft etc. Are the physics tilted against lighter riders more so than on other places??

  14. Gazzinho says:

    Hi I’m loving all the work you’ve put in and I really appreciate how much more realistic this programme is too any other out there. I did have a comment about elimination rides with bots on the Herne Hill velodrome. It seems to me that the braking dynamic works against this set up as everyone is together making it impossible not to hit each other and apply braking, it would be nice to have the option to turn off the collision for this type of ride. I can accept it on all other rides when not running at a small loop as there is literally no where else to go than into riders. I think the braking works well in all other aspects other than the bots having the upper hand into corners. I noticed on the Herne Hill track everytime you get 180m from the finish you slow down for 2/3 seconds it’s really odd.

  15. Gazzinho says:

    Hi again,
    How do you get 20 riders to sign up in order to get your ride to go live? I struggle to get 3/4 at best . I have some fantastic ideas for rides and feel like I can contribute to the varied selection of rides you offer. Admitently sometimes I feel you could have alot more variety on offer.

  16. Steve Reynolds says:

    New to RGT after 5 years of the Orange One, and straight away the absence of sticky draft is a huge step forwards. One other feature that deviates from reality is the halving of the effective downhill gradient in that other app, what’s the physics in RGT?

  17. Mathieu Drujon says:

    Hello ,
    Instead of watt min>>> and <<<max I would suggest to add a visual gauge.
    Like a car speedometer with a needle, a pointer which would change in real time regarding the watts needed to stay in the draft. We could divide this gauge in 4 quarters, 1st on left, red : watts to low
    2nd quarter, green : watt min to stay in the draft by saving maximum energy,
    3rd quarter orange : still in the draft but you spoil a bit of energy for nothing,
    4th quarter: i.e purple : you overtake the rider in front of you.
    It would be easier to follow without to have the eyes really on it.
    Don’t hesitate if you need more information on my idea. Best regards.

  18. Gabriel says:

    Hi, I have a cadence question: I tried a workout on RGT that activates “erg” mode. But no matter what gear combination I try, it is impossible to meet both power and cadence targets. I need much higher cadence than targeted, to reach the power. The gap is bigger at lower power: 105 rpm cadence vs target of 85. How should I handle it?

    • Tim Sinclair says:

      Generally speaking if you want to work at a certain cadence for a particular interval then you should adjust your cadence slowly to allow the ERG control to maintain the target resistance. Sometimes it can be better to complete a workout with EFR off and the intensity slider under the Slope tab turned to zero. That will allow you to self select your cadence/gear combination to allow you to reach achieve your targets. ERG doesn’t suit all types of training requirements. At lower power the software may have no control over your trainer at all, due to the natural resistance of the trainer that has to be overcome.

      • Gabriel says:

        Thanks for the answer, Tim.
        I couldn’t find a slider for turning “ERG (or ERF?) on/off”. It is activated by default by RGT during workouts.
        The power of the workout (step1-120W and step2-280W) is well above the natural resistance of my Tacx Genius. The application does control the trainer, but if I slowly adjust the cadence from approx. 105 toward 90 rpm (I set this target on the workout), the power also decreases until I fail the interval. It is like RGT ignores my 90 rpm target and obliges me to spin at around 105-110 rpm, no matter what.

        • Tim Sinclair says:

          Sorry you’re still having problems with this 🙁 . The ERG switch can be found under the slope tab of the phone app when in a workout (it doesn’t appear at any other time). I would certainly try turning that off, and reducing the slider under the same tab (feedback intensity) to zero. The road profile will have no effect on your resistance. You will still see your intervals on screen and be able to use them simply as prompts. Adjust your gearing and cadence to give you the combination that best suits your target. ERG can be quite restrictive when you want to modify the cadence.
          Best wishes

  19. Andy says:

    Hi, I’m new to RGT and enjoying it so far. I did a race last night and on several occasions, I and the other rider I was with did a number of u-turns which was rather disconcerting. Also near the end of the race, my screen app showed zero power and my avatar stopped but the control app was still showing the power from my turbo. I managed to get going again by stopping and re-starting the control app but lost about 20 seconds and 2 places. Hopefully these are just teething problems.

    • Tim Sinclair says:

      This can happen when there is a connection interruption between your Phone App and the Screen App. It can often be caused by a reduced quality WiFi signal (quality = without droputs, and is not speed related) . RGT is working in real time, second by second. No caching takes place. The avatar data does not get updated so things start to behave strangely. Sorry to ask the usual obvious q’s but is your Screen Device hard wired to Ethernet? Are you close to your WiFi Router with the phone? If then phone app is still showing power, then restart the Screen App so that you don’t have to pair your devices again. The phone App and screen App don’t need to be on the same network so you could also try using your phones data instead of WiFi to se if things work any better that way. RGT is not data intensive so it shouldn’t eat all of your phone credit 😉

  20. Peter says:

    Hi RGT team, I quite like most of the “Physics” aspects and mostly agree that they make sense and bring a virtual ride closer to real ride. However, I have noticed that it is VERY easy to get dropped from the group, compared to a real peloton. Specifically, rider grouping tends to be lined out across the road parallel vs. real life “teardrop” shape. Such group would then reshape itself into “dual head” formation where a rider (or few) can get “stranded” – i.e. essentially there’s a 2 groups, main and small one, parallel across the road. At that point main group would simply ride away from you unless you put in an effort. If you’re lucky your avatar would steer toward the main pack providing you with shelter. If you’re not so lucky the gap will open and you will get shelled out the group, where just a minute ago you had good position in. Can this grouping be addressed please? Or at least explain what causes riders to align parallel across so one can avoid that at all cost. Personally I’ve never had such a hard time holding a group on the road no matter what. This is just my two cents regarding physics aspect which seems furthest from reality.

    • Tim Sinclair says:

      In short – yes 🙂 It’s very subtle (as in real life). If you work hard to catch a rider and then move into their draft then that is when you will notice it most. You see the power you are saving in green under the Power / Cadence / HR panel at the top right of the screen.
      (In RGT Terminology ‘trainer difficulty’ is something specific – see the SENSORS Page and I assume you aren’t referring to that)

Leave a Reply

Your email address will not be published. Required fields are marked *