Forum:Why are Broughy1322's speed tests inaccurate?

Hello, folks. I've decided to start a forum about this and see what people have to say, but I am still sticking by my side the Broughy1322's "speed tests" are inaccurate.

Firstly, I'd like to know how he determines when the vehicle has reached its top speed - not how he finds out the numerical value of the speed, but how he knows the vehicle has actually reached maximum potential. He claims the vehicle has reached its maximum potential when it has reached its maximum revs. All morning and afternoon today I have being driving cars for as long and straight as possible without traffic/weather/anything to stop me, and I can honestly say that the car's revs NEVER come to a full stop. The revs are always increasing, if you listen to the pitch of the revolutions, they gradually get higher even when you think the car has reached top speed - this proves the car cannot possibly have reached its maximum. For the car to have reached its maximum, either the revolutions have to reach the redline (rarely), or they have to remain the same. There is no way in-game that can prove revs have stopped increasing. The tachometer in first person does not prove this either, as I have noticed how jumpy it can behave, or sometimes it doesn't go any further through the scale even though the pitch is still increasing. This first point proves the cars are impossible to tell when they've reached their maximum potential. This can also be proven with speedo mods. In both GTA IV and GTA V, the most accurate (as in, the most amount of decimal places) speedos fluctuate heavily at high speed, showing the inconsistent amount of speed the car is pulling. Speed cannot be determined by this minuscule amount of acceleration.

My second point questions the accuracy of distance. While I'm not entirely accusing his distances to be inaccurate, I am questioning what exactly he is doing to gain distances. In the creator, when creating the race, you can see the distance of the race so far. It is accurate to 1.1 km (or miles, depending on system). If you look closely, when moving the last checkpoint further away from the previous one, you can see that moving it the tiniest amount doesn't change the value - because the in-game measuring system is rounded up to the first decimal place. This means that "1.1 mile" could be a minimum of 1.05 miles, or a maximum of 1.15 miles. This might seem like a small amount, but assuming a vehicle has taken 20 seconds to pass a distance of "1.1 miles", you could say that 1.1 mile/20 seconds = 198.0000 mph, but, if taking the lower bound of 1.05 miles, the speed may have been 189.0000 mph, and if taking the upper bound, the speed could have been 207.0000 mph. As you can see, that's a total variation of 18mph. Sure, you could add these up and divide by 2, but that doesn't take into consideration the actual distance and this is not what Broughy does.

Now, I am not arguing here and am not up to argue, but I want people to prove me wrong here, and if they can't, then I am not going to change my mind. I'd also like to add that I Broughy himself claims that "we don't want acceleration to take the result", however in all his tests, he reaches the 4th checkpoint (the checkpoint at which he begins timing and using the distance attained) where the Senora Freeway reaches Los Santos and slowly declines, which can also be heard in the revolutions; the revolutions of the engine can be heard increasing in pitch, thus acceleration is still occurring and is still affecting the result.

Please do comment and answer/solve any of my comments above. Thank you. Monk Talk 14:14, June 29, 2016 (UTC)


 * I'm not too sure how accurate they are, I've seen Broughy's videos before and they make sense. I did have a watch of this just now, though.

He is right. I had a look at it for myself and yeah, the speedos are pretty much bullshitting. The left image is stock Z-Type. The right image is fully modded Z-Type. The MPH/KMH are part of a trainer, yes, but they show the car's speed directly - just a simple "get current speed" type of thing.


 * On that note, the stock Z-Type seems to have a top speed of 120MPH because that speed never went up until I drove down a slope. Then again, I tried Adder afterwards and it was also locked to 120MPH. It's all a giant clusterfuck. v-michael-trunk-mini.jpg V-franklin-trunk-mini.jpg V-trevor-trunk-mini.jpg 16:02, June 29, 2016 (UTC)


 * Ok, so it seems he is half-correct. Some speedometers do lie, but some are accurate. Drove an 811 and the speeds did match. Gonna try other speedometer types to see which ones can actually be relied on. v-michael-trunk-mini.jpg V-franklin-trunk-mini.jpg V-trevor-trunk-mini.jpg 16:18, June 29, 2016 (UTC)


 * Mate, I'm talking about his D/T = S speeds, not the FPV Speedometers. He uses distance and time and I've discussed whether it's accurate or not. But yes, the speedos are shit and totally messed up, totally unreliable. None match at all xD Monk Talk 16:42, June 29, 2016 (UTC)


 * Yeah, I've always wondered this. D/T is speed in a vacuum, and yeah, everything's just a massive cluster of mess. Thing is, as you said, the distances can't be accurately measured, what you said about 1.05 and 1.15, this is bounds in mathematics, and basically summarises how much variety a rounded figure has and how much it effects the result of a sum. So yeah, I think his speeds are pretty controversial. As for Wild, yeah, I've noticed this too, and again, nothing matches xD it's just a mess, but hey, it's still quite specific in terms of each vehicle differing in each form of data speed, so it can still be documented :P Mr. Ferrari (talk ) 17:08, June 29, 2016 (UTC)

Sedan Interior 1: Match. gentaxi (Buffalo, Dilettante, copcars, taxi, etc): Match. cavalcade (SUVs, Panto): Match. dukes (muscle cars, Sultan RS/Banshee900r): Mismatch. Sports Interior 1: Mismatch. Van Interior 1: Mismatch. Truck Interior 1 (digital): Match. In fact, the number works exactly the same the trainer display. banshee (Banshee, Vacca, X80, etc. Comet-based falls here too (T20, comet, etc)): Match. Except for FMJ - that one is hilariously off. feltzer (Feltzer, Verlierer, etc.): Mismatch. futo (Futo, Blista Compact): Mismatch. sultan (Kuruma, Sultan, Khamelion, etc): Mismatch.
 * I know. I just saw the video and remembered that speedometer info was put in tables a while back and he had a video on the speedometers. I did some tests with some of the speedometers and these are the results:

As for the top speeds and the way he calculates them, I tried some of the fully-upgraded cars in his spreadsheet. Zentorno was 122-126 MPH, stayed mostly at 122-123MPH. His results say 122MPH. Exact same result for T20. On Banshee 900R I got 134-140MPH, stayed mostly at 136MPH. His results say 130MPH. Used the same road as in his videos. I guess his results aren't precisely accurate, but they're really close. 17:11, June 29, 2016 (UTC)
 * It's really difficult to summize xD Precise? Relatively. Accurate? Hum, it really depends on what's "correct", since we don't know the actual intended speeds (hence these kinds of tests), we can't say it's accurate. The variety in results can be big or small depdnding on the vehicle, as evidence by wild's results - some really precise variations, not so much with others. Thing is this kind of test can give a different result every time, and that's not even the end of it, because it all depends on the quality and style of test done xD with different results everytime, and even more different results between people, it's impossible to nail down 1, 2 or 3 speeds xD it means that this kind of data is highly controversial and really wouldn't work in the overview tables xD Mr. Ferrari (talk ) 17:20, June 29, 2016 (UTC)

I just wish the roads would be more flat like the runway. The physics are just broken on any other road. Driving over potholes? Here's a bonus 3MPH for you! lol   17:57, June 29, 2016 (UTC)
 * I had a go on some other vehicles. All fully upgraded, max suspension, armor, cosmetic upgrades, engine, brakes, BP tires, etc. X80 Proto, Pisswasser Dominator, Schafter V12, and Monroe matched. Carbonizzare was off by 2MPH. I haven't a clue how the hell he calculated 811 (132.5MPH) because I've been getting up to 160MPH. Ok, I was incorrect. I didn't watch his video to the end so I missed the info about the framerate bug. Getting 132MPH now when I get down to 30fps (or go above 100fps).


 * Yes! Haha xD and how jolting the steering 1 fucking nanometer drops the revs xD it's messed up. I think we should stick to stats that dont vary. Ie fps speedos have a limit, as do the website stats and handling files (fixed). Too much possibility with D/T.


 * Check out the video description on any of his videos on the top speed series. He's using a more accurate process by shortening checkpoint distances and repeating tests, then finding an average. I might apply this method myself. Monk Talk 19:11, June 29, 2016 (UTC)