Dodges explained. Power & more - Rocket Science #14 

Rocket Science is the diagonal flipcancel better than the diagonal dodge with boost?

When you do your wave dash episode, I would like to show you something that could be very helpful in your findings, which debunks the 4 wheels on the ground to get a flip. I can get 2 side wave dashes in one jump. I posted it on reddit and didn't get muich attention and I think it would be more useful for you. I can make a video about it.

I've been doing the diagonal cancel with my kickoff for a year. The secret is now out!

@Erkin, not sure if you mean flipcancel + boost or just the flipcancel. The flipcancel changes nothing about the impulse but because the boost exhaust is always facing exactly backwards the boost will be used more efficiently -> cancelling is better if you do it right

@SAN FRAN www.reddit.com/r/RocketLeague/comments/7so6m8/flip_reset_conditions/dt76pzs/?context=3

yeah, kudos to musty for finding this and making it practical (actually, I hate him for it because I can't do it).

I was wondering if anyone else noticed. He straight up described the diagonal cancel of a speedflip a year before it was popularized.

I was about to comment about this when a saw this comment. Halfwaydead discovered the speed flip a year and a half before Musty.

Thanks for creating awesome music! and of course for watching

Oh, hello Protostar!

You're the least expected person to see here

MORE ADS!

>>

Honestly it feels like batmobile has insane power on flicks. I spent so much time in freeplay and now that I've switched to octane, I still can't get the power that the bat brought. But maybe it's placebo idk

@@Exsulator2 Longer and wider hitboxes do indeed make it easier to get more touches on the ball during the flick. With practice, you can get plenty of power with any of the hitboxes, but yeah it's technically easier with plank than octane

@@JohnJillky that's true but I think (I don't know, but I THINK) that since the hitbox is longer, and all cars rotate at the same speed, then it's logical that the longer radius, the more speed it technically has. Imagine spinning a 1 meter plank vs spinning a 3 meter plank. If they have the same RPM, then the 3 meter plank's ends actually moves a lot quicker than the 1 meter plank's ends. So imo a longer hitbox can literally generate MORE power than a shorter hitbox, granted that you use the very ends of the hitboxes for the hits

@@Exsulator2 Halfwaydead actually did a video covering flicks. One thing to note is that in game physics often don't work exactly like irl physics. Hard to say how much that affects things especially with how much error went into the way he tested things. Ever since I learned about RLbot, I've wanted to know if it could be possible to code a bot to do things like flicks perfectly each time. Wonder if they could do a better test for this kinda thing these days (I mean, I've been 1v3ing Reliefbot recently, that bot is way better than I remember some time ago lmao) The link is time stamped: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-DRO8-784eSk.htmlm48s

Thank you very much :) It wouldn't be non-linear, just more sensitive. The best way to try if you'd actually like it less would be through external software like DS4Windows or the Steam controller configuration (in cross mode + lowering deadzone outer). They will modify the raw controller values and therefore affect everything, also the deadzone and dodge deadzone as a result. If you use DS4Windows with 1.5 sensitivity you'd make your deadzone and dodge deadzone 50% bigger to compensate for that. Then the only difference to 1.5 in-game sensitivity is the flip canceling.

Ok, maybe 'leniar' and 'non-leniar' isnt the exact wording. I meant 50% on the controller should remain 50% on the dodge cancellation. And psyonix prolly wants it such that, the aerial sensitivity being higher doesn't make its so that the dodge is affected faster. I was using modified sensitivity through DS4Windows till about 2 weeks ago.(used for about a year) When I realised it was affecting the dodge too, and since I never liked the fact that it was affecting my turning on ground(just dealt with it for the aerial benefit); i switched to having aerial sensitivity in-game. It took a day of adjusting, but my overall performance has increased(not placebo cuz my friends also noticed the improvement in my play, without me saying that I had done anything). Best case scenario would be to get a different setting for it. But that might not happen since it might crowd the already extensive settings. Only thing I miss from the increased sensitivity on dodge cancellation is that I cant do my usual celebration anymore, lol

QaleQ I have my controls set up to allow both. Square circle is air roll left/right, boost as r1, air roll/ powerslide as l1, jump as X. Once you get used to these controls I am convinced that they are the best.

@@AbusedPopTart I switched to these after having boost on square and only air roll right on R1, definitely find it exceptionally unlikely I'll ever switch my binds again. Hybridization is the future

Not to my knowledge. At least not constantly. Haven't had any time to play myself lately so I haven't used it in-game really even though I've gotten good at it in freeplay.

Rocket Science first and foremost, the rocket league community is grateful for all your hard work and time spent doing these kinds of videos. Eventually you’ll get the recognition you deserve. Hard work is always noticed. Secondly, What’s the best method (e-mail, Skype, etc) to get into contact with you? There are things I’d like to discuss with you whenever you have the time.

Thanks. Tbh I feel like I get recognition. Can't complain, going towards 10k subs and people keep watching. If you would've told me that when I uploaded the first video I would've called you crazy. It's never going to have the amount of views that "entertainment" content like Jon Sandman has anyway but if it keeps growing I'll be happy. Best method to contact me: Join my discord. If it's a public question just post it in one of the channels else DM me. If you don't use discord, twitter DMs or halfwaydeadrl@gmail.com

Thank you. It should definitely be possible to create a timer bar with bakkesmod but I've never really worked with the GUI so it would take me a long time to figure it out. But you could ask bakkes. He's currently busy with uni I believe but when he has time he's usually very helpful. I talked about the extra speed in Episode 11. Basically you would first just apply the impulse which would give you sqrt(2300^2 + 950^2) = 2488 total velocity. Then you would just normalize the speed to 2300 but keep the direction. Did you see my reply to your arduino comment btw? I believe I didn't get a notification to your last comment and I just answered yesterday

I'll have to try out Bakkesmod, it looks pretty cool. I assume that if this feature were implemented, it could be limited to training and maybe unranked games, right? I wouldn't want to add a cheat to ranked play. Ah ok, I just watched 11 (don't think i'd seen it before) and the behavior is just what I would expect. So a diagonal or sideways dodge at full speed will reduce your forwards speed. Yes, I just saw it. So you've got an arduino setup now? That should work well i think.

Almost all of BakkesMods features are offline only. Arduino is working great, too great :D Can determine the input lag so accurately that Windows running anything in the background shows up as a difference of like 0.1-0.2ms. I should try at lower FPS. The variance would be higher making randomness more of a factor but it wouldn't need the full power of the CPU. I can easily collect 500 samples now.

0.1 ms? That doesn't really make sense to me. If the game is running at a fixed tic rate of 120hz, then each time frame is ~8ms meaning your results should have a variance of at least 8 ms even if your measurement system is perfect - simply because you might hit the button at any time within that 8ms window. Or are you talking about looking at an average of a whole lot of samples?

First of all, I wasn't talking about car physics. Inputs get checked every frame and camera swivel isn't tied to the tickrate. Nevertheless, I get about a 4ms range of values at 1000FPS due to 1000Hz polling, inputs barely missing frames and performance inconsistencies. 120FPS would as you suggested create an 8ms range which would require more samples to make sure randomness doesn't affect the measurement. And yes the 0.1ms difference is noticeable on the average but also on the maximum values that i got. In the test with 15.3ms average I had 17.4ms max and in the test with 15.1ms average it was only 17.1ms. I'm taking the average of 500 samples which gives me an estimated standard error of 0.03ms so even a 0.1ms change due to variance is highly unlikely.

Yunis Rajab - AFAIK supersonic is not full speed but rather something like 95% of full speed. (Consistently boosting brings you to 100% for example) So it should still work and make you faster for the moment. Not sure but you might lose speed and thus lose supersonic when landing so it might not always be worth it. You also can't really steer during the flip.

with worst documentation :p

can't argue that lol

Thank you! With kickoffs it's not always about being there first although it certainly has it's advantages. If you're touching the ball while you're still mid air and the opponent gets a solid 50/50 he's going to win. My point is that it is more important to come at the ball from the right angle and jump off the ground at the right time. Diagonal dodges on controller also have the "disadvantage" that they can be at basically any angle. So you might try to do a 45° dodge but you'll end up doing a 55° dodge by accident. A straight forward dodge has some angle snapping and it's affected by the deadzone which makes it a lot easier to get the 0° angle dodge every time you go for it. I discussed how it works exactly in Episode 7 about dodges. Consistency is key, especially in 1s. In 3s teams often "lose" kickoffs to their side on purpose which also changes the entire approach you want to take.

Yeah all things regarding jumps and a lot of other car physics will be in their own video but you're right, a mention would've been good and the timer is the same, yes.

Not sure if I do get what you mean. A side dodge will always give you sideways momentum but depending on where the camera and car are facing and your current velocity it might look like it's just stabilising. For example, if you use the side dodge opposite to the direction where your car is currently flying in then it will slow you down and not speed you up. If you're talking about the stall thing I mentioned at the end then that is not really a side dodge. It requires the specific bindings, although I'm not sure how to explain it better in a comment than the video + squishy's tutorial that I mentioned.

That's probably it.

Can you show a video of this happening? It might be the perspective that makes it look like that. In my testing, what you described never happened.

atR NytroDryft ikr

momentum.

If you're talking about what I think you're talking about then yes. Dodging cancels all your vertical momentum and the clipping happens when you do a dodge really quick after your jump because you're barely off the ground and 0.15s later your vertical momentum will be reduced to 0. If you're too close to the ground some parts of the car will clip into the ground when rotating. I hope that explains what you were talking about.

On the controller you would be using the analog stick to steer left and then an airroll right button for example. You can also steer right and airroll left. On keyboard you're just pressing 2 buttons which is very easy but on controller if you don't move the stick perfectly left you won't do a stall. So if you move your stick a bit up and left by accident then the game will just to a forward flip. Because left and right cancel each other out but forward remains. You need to make sure there is no forward/backward component to the dodge.

Now I get it. Thanks!

Yes, it is used by all highly skilled players

Speed flipping is a real mechanic

v_forward = v * cos (angle between the forward direction of the car and the direction of v); // can skip this angle with some vector math x = 500 * sin(dodge direction angle) * ( 1 + 0.9 * v_forward / 2300); //sideways dodge y = 500 * cos(dodge direction angle); // forward dodge arctan(x/y) = angle of the impulse sqrt(x^2 + y^2) = magnitude If this isn't self-explanatory then contact me on discord or twitter :P that's why I said that in the first place.

Yeah just sent you a message, German as well, probably easier for me to understand then