It definitely would affect rocket jumping! If you are changing the geometry of the projectile impact box you are bound to affect several types of jumps (wall climbs being one easy example).

Change it to an octagonal prism like warsow did.

Isn't this something that could be modified by Valve for the release of matchmaking?

I mean most things in the game could be modified by Valve for the release of matchmaking. It's whether it's a big problem and how easy it is to do. And this scores fairly low on priorities I imagine.

Or rockets going through another soldiers butt and not registering.

Most likely not a simple easy thing to do. Axis aligned bounding boxes are for simplified collision detection, so they would have to redo that

Honestly, how hard is it to make everything behave as the best possible scenario?
Stairs being a simple ramp in the eyes of projectiles/explosions.
Hitboxes being cylinders scaled to the actual size of the mercenary.
am I missing anything?

Stairs/ramps/cracks combined with splash damage gives me internal bleeding when I think about them

Very hard.

a basic understanding of the situation and critical thinking skills

im sure u hit ur power button on ur alienware and connect to the internet and assume theres magic data fairies making sure all the packets get to the right place at the right time

http://tf2maps.net/threads/tip-splash-damage-bug-what-it-is-how-to-avoid-it.20749/
Yep...

I guess I simply don't know enough about the subject. My basic understanding sees "stair prop" that has jagged edges, and changes the prop's box or whatever to be a ramp or inclined plane, whatever the need.

I understand it goes past that. I understand TF2's explosions already behave strangely on sloped surfaces that aren't jagged at all.

But no, I'm not data-fairies level.

thinking about this, doesn't this mean that there's points where pyro's airblast gets easier or harder, entirely depending on the angle the projectile is coming from?

http://i.imgur.com/DRbJZG3.png

(t-that's supposed to be a demoman grenade, I swear)

http://www.reflexfps.net/forums/index.php?/topic/2953-devs-stance-on-enemy-model-animations-hitboxes-ragdolls/#comment-22391

Reflex has switched to a capsule. The dev talks about it in the forums post

http://i.imgur.com/JLDvKYl.png
http://i.imgur.com/UFI4nDW.png

damn that's actually really cool, i'd love to see something like that in tf2, although i really have no idea how the nuances of the source engine would change...

Considering the new CSGO hitboxes look alot like individual capsules on every limb, I wouldnt imagine it would be TOO hard, but thats wishful thinking I suppose.

the hitscan hitboxes are really similar to that already though, admittedly they could use an update in the same vein of the new cs ones. the projectile/collision hitbox could never work like that though, and im sure the cs collision hitboxes are nearly identical to the tf2 ones

I might be misremembering but i think 2007 tf2 draws eplosions as aligned squares rather than spheres, but it was patched out soon after release.

point being valve is capable of making this change easily

edit: [September 15, 2009 Patch] Fixed explosive radius detection using an axially aligned box instead of a sphere

i think the problem its not if its possible to do, but how will this affect the game afterwards

the end difference for the explosion radius change was that the most significant loss of detection/damage was where the corner used to be. If I'm interpreting the numbers right, 2007 tf2 had rockets be map aligned squares w/ 242 as the perimeters, and then changed it to be a radius of 143.

Rockets became 18% bigger when aligned along the map's axes, while the diagonal of the square went from ~171 to 143, so the rockets became ~17,3% smaller along the diagonals.

actual rocket area mildly increased, so assuming they use the same method to do bounding boxes, you would end hitting slightly more shots because of the change than you would lose because of it, assuming that map alignment is random in such a way that you'd end up facing the diagonal and the perpendicular sides relatively consistently. Since most maps flow along the grid, you'll probably see a slight buff to projectiles as the hitbox extends slightly further in the direction most of the combat tends to orient itself around

or im wrong idk

It's not quite that simple. You can't (easily, reliably, and at high performance, pick any 2) determine whether a single point is part of a staircase.

It's a bit like asking "why doesn't facial recognition software just look for 2 eyes and a mouth"

i would like to remind you all that when there were updates in the past that changed sticky or rocket jumping a lot of people lost their minds.

this will do it on a completely new level if you actually insist on them changing the bounding boxes

be careful what you wish for.

This can be fixed by closing your blinds.

a console command that jump servers can auto-set on older maps to use rectangular bounding boxes solves the one justified niche it would otherwise have

if ppl actively leave beyond that, they were already leaving when matchmaking inevitably changes a million other things anyways

if rocket jumping is changed it'll affect wall jumping as it is now
thanks for the correction capnfapn

if rocket jumping is changed it will affect rocket jumping as it is now.

So does this explain why half the time when jumping of the wall my rocket gets eaten by the fucking wall? (Fuck snakewater)

No that's not it

if you only change the collision boxes for rockets and not for moving then nothing is going to break and nobody is going to notice
valve are just lazy

also

move the explosion up by the max step height. it's basically free and works well enough in practice
but again valve are lazy

How you fix stairs is give mappers something that blocks rockets but not hitscan so artmappers are willing to clip their staircases with it and no, displacements flagged to block projectiles do in fact let rockets through because rockets are actually just short hitscan in the physics engine.