Carter glitch on q2dm1?

[Electfried]

Re: Carter glitch on q2dm1?

Yes, I agree. The Carter administration was such a huge fuck up that it's horrendous effects were felt 20 years later in video games. I certainly didn't vote for him. But I would definitely elect him to be fried. I know that's a serious alligation.

SOME ONE GAVE A SHIT ABOUT TITLE POSTS

[haunted]

Jay Dolan:

It's mostly noticeable with clearing/djing those 64 unit crate boxes on certain maps relative to the origin.. i think me and adren were testing it and completed it with as little as 65-66 fps or something(after multiple tries) on dm8 one time contrary to what you read in n00k1e's guide.

It's interesting to think about the other effects the origin has on movement per map and how it's played too in a vague sense.

[[BTF] Reflex]

Nudges..  sounds like the solution they had to 'sticky wall syndrome'

I wonder if this relates to being able to go faster while strafing against a wall.

[Jay Dolan]

Possibly. I'm not actually aware of being able to strafe jump noticeably faster along walls, but the "overbounce" velocity I mentioned earlier might enable that. The other thing that likely helps there is the clip against an immediate wall doesn't cost any pmove time: your movements are each time-boxed like everything in Quake. But time and distance traveled are treated as one in the same in the pmove code, so if you don't actually travel any distance, you don't burn any time. Thus, clipping against an immediate wall and picking up that overbounce is essentially free. Here's what I'm talking about:

http://jdolan.dyndns.org/trac/browser/quake2/trunk/src/qcommon/pmove.c#L119

Notice how time_left is decreased only by distance traveled -- it is not impacted by changes in your velocity vector as a result of clipping.

-----------------

To break it down a little further:

When you run into a wall, pmove attempts to clip your velocity along that plane and scale it back up to your intended speed. Here's the code that calculates how to clip along edges and applies overbounce:

http://jdolan.dyndns.org/trac/browser/quake2/trunk/src/qcommon/pmove.c#L63

And here's where that clipping is invoked (PM_StepSlideMove, the heart of Quake physics):

http://jdolan.dyndns.org/trac/browser/quake2/trunk/src/qcommon/pmove.c#L162

It's this logic (the "go along the crease" bit) that likely enables what you're describing, but it's also what hangs you up when jumping up stairs sometimes. I spent literally weeks rewriting all of this logic to try to simplify clipping and produce more fluid interactions with stairs and crates and things. This is why double-jumping is incredibly easy in Quake2World. But I only had limited success in making stairs easier to jump up:

http://jdolan.dyndns.org/trac/browser/quake2world/trunk/src/pmove.c#L133

Someday I'll revisit that and get it totally right

[[BTF] Reflex]

Just to clarify, its not while jumping.

Only way I can think to show it is walking off the ssg platform on dm3 over to the armor.
If you straf and slide the wall you can make it without jumping..  if you just walk off normally you cant.

Something Luci showed me..  can use it anywhere to get a TINY bit more speed and not make noise jumping.
Links to code not required (I never made it past Turbo Basic)

Thinking about it,  there is probably no relation.

[haunted]

The only relation is that it can be similarly broken down into a vector against the plane(wall) you're sliding against. Forward+strafe would be the horizontal and vertical components, and they are dotted into each other, equating to a single vector that's a combination of force applied perpendicular+parallel to the plane, like a sweet spot of sorts or maximum point in which you'll go the fastest.

You can usually only take advantage of physics like this in the air. So adding an additional plane to the mix(floor + wall) obviously changes things, but no it's probably not related precisely how you're thinking reflex.

[Jay Dolan]

The only relation is that it can be similarly broken down into a vector against the plane(wall) you're sliding against. Forward+strafe would be the horizontal and vertical components, and they are dotted into each other, equating to a single vector that's a combination of force applied perpendicular+parallel to the plane, like a sweet spot of sorts or maximum point in which you'll go the fastest.

You can usually only take advantage of physics like this in the air. So adding an additional plane to the mix(floor + wall) obviously changes things, but no it's probably not related precisely how you're thinking reflex.

Not horizontal and vertical, but X and Y. And not dot-product, but added. But yes, you are on the right path.

Ref, I wasn't linking to code to bore you to death. I know that quadz, Qwazy and some of the others are equally interested in the code as I am, that's all.

It'd be easy enough to add some instrumentation to the pmove code and confirm at least some of these theories. Basically, record the length of the player's velocity vector pre- and post-clipping, and print whenever clipping results in an increase. If anyone is particularly interested, I can do that.

[haunted]

ah, yeah i meant 2 dimensional tip to tail / vector sum; not dot product. It was early when I typed that.

edit: or late... something like that.