Posts by Klems

[Blocked Image: http://image.noelshack.com/minis/2015/09/1424837140-2015-02-25-00001.png][Blocked Image: http://image.noelshack.com/minis/2015/09/1424837140-2015-02-25-00002.png][Blocked Image: http://image.noelshack.com/minis/2015/09/1424837141-2015-02-25-00003.png]

http://image.noelshack.com/fichiers/2015/…02-24-00013.jpg

I think there should be a window here, so you can peek B from it. You need to go through middle to go here anyway so it's not that unbalanced.

http://image.noelshack.com/fichiers/2015/…02-24-00014.jpg

These doors are way too narrow. When you go through the first door, it's not obvious that there's a second door to your right. This can be fixed by changing the door placement a bit so they're not so close to each other.

http://image.noelshack.com/fichiers/2015/…02-24-00017.jpg

This whole area is a bit too complex for no good reasons. Maybe because it's too dark, or the pillars block sight, I'm not sure.

http://image.noelshack.com/fichiers/2015/…02-24-00018.jpg

I know it's temporary but you can see through these props. Add black brushes inside your props. Same thing for the cars in the parking area.

Also there's a really OP position right here if you jump on the first stack of cocaine blue meth.

http://image.noelshack.com/fichiers/2015/…02-24-00020.jpg

I feels like this area should be walkable.

http://image.noelshack.com/fichiers/2015/…02-24-00022.jpg

The prop to the left isn't suitable for cover because there's so many weird little bumps on it. The stairs to the right are a nightmare to shoot through, and also are somewhat pointless. I suppose it's here to provide cover from middle, but you're already covered from middle if you jump on the prop.

http://image.noelshack.com/fichiers/2015/…02-24-00023.jpg

I feels like you should remove the cactus or change it in some way so you can peek B more easily from middle.

Quote from Gambit

Just have an open mind about the suggestion/idea because I feel ANY suggestion about changing nuke would garner than the same reaction you just gave. In 1.6 it was a LOT better because of the wallbanging, but without that in csgo, it makes it a really stale map.

Everyone on reddit wants to change nuke in some way, and I think Valve will change it at some point. Sure, it makes succesful T strats interesting to watch, but of course when most other rounds end in a slaughter I don't see the point.

To be honest I think it's just a matter of changing chokepoints and a few positions. Make the entrance to hut taller so you can peek rafters, change the geometry of radio to give more room for terrorists. Remove a half of rafters (the one over main and squeaky/hut), add a ladder rather than a trickjump to silo, etc. I think outside is fine, you could make garage brighter and remove a few crates on CT side but that's it.

Then of course wallbangs should be a lot more powerful.

I dislike your idea because it change the map too much. But I wonder if you really have to make these huge changes to make a map more T sided. Do we have data (round outcomes) on the new train?

Also check this http://steamcommunity.com/sharedfiles/fi…s/?id=385634872

Quote from Guni

The reason everyone complains in competitive about Dust 2 is that its too simple and balanced, theres no challenge. Its why its the most popular casual map, cause its not hard to play.

The reason everyone complains when they play dust2 is because the map is balanced? What kind of reasonning is that? Or I didn't understand you properly. I see more whining during T side nuke. Unbalanced map are more challenging only one half of the time.

Aztec or dust1 isn't hard to play when you're CT. Neither when you're T when I think about it, you only have a couple of entrances and all you can do is flash your way in and hope for the best.

Nuke should stay unbalanced, I agree it does make it interesting. But you can't deny there are some issues, especially in A site. It's difficult to flash correctly and when you go through squeaky, hut or main you can be shot from a dozen of different positions.

Quote from JekasG

- Generating the Grid to do the Pathfinding ( Main Problem )I know for normal Pathfinding, Normals Linear Grids are used. But i am certain that for my situation i cant use that.

So how can i generate my Graph ?

Please include some algorithms

We need an exemple of one of your levels and a size estimation (level size and maximum path length), otherwise we can't really help you.

Is the level made on a grid? Or with polygons? You can use octrees though.

It shouldn't be that hard though. There's no need for a navmesh, pointgraph is good enough for a platformer game, especially if this game is on a grid.

Generate enough points for each platforms and connect these points (= points are connected by walking). Once this is done, you connect adjacents platforms by finding out each pairs of points that are connected by jumping.

Quote from JekasG

- How is jumping done ? ( Major - Minor Problem ) Am i to program some code for the computer to input some buttons so the AI will jump ?

- How can i perfectly do the Trajectories ? ( Major - Minor Problem )

You gotta remember your physics/maths classes. You can deduce the general trajectory of any solid from his initial velocity/position + gravity. http://en.wikipedia.org/wiki/Trajectory_of_a_projectile

This trajectory can slightly change depending on how well you wrote your physic engine and on which numerical integration you used.

Basically, for any pair of nodes in your point graph that could be connected by jumping, you try to compute the optimal initial velocity to make the jump (some trajectories are faster than others). This velocity can't be higher than the base running/jumping speed of your IAs, obviously. You also need to trace along to check for any obstacles. Dropping down is a special kind of trajectory where the initial vertical speed is 0.

Once you know which points are connected and have the initial velocity needed to reach it, then there's nothing else to do. Once an IA reaches this point and wants to jump, then apply this velocity and that's it.

It will probably be too efficient at first, you can balance that by limiting the velocity, adding a cooldown/animation between each jumps by the IA, etc.

Quote from JekasG

- How are cost / weight calculated ? ( Minor Problem )

Which ones? If you're talking about weights in a point graph then it's just a matter of being more or less proportionnal to the time it takes to traverse an edge.

Maybe this can be of some use

I had the same problem about sensing speed on a flat texture. The answer isn't hard though: don't use flat textures. Lighting and reflection can do wonder. Another solution is to add some grain to the texture, either in a shader or on the diffuse. You can also try changing the fov depending on the speed, or add a looping sound (whoosh like) and change the pitch/volume according to the speed.

About being able to see edges, you could add black/white outlines to the geometry, or make a shader which read the depth buffer and do something at the edges. The problem is that each surface is completely flat, so the eye doesn't catch the edges as there's no normal/light change. You should smooth the edges in the geometry, so you can either add lighting or a rimlight effect.

Use dithering if you have color banding issues.

Quote from Spherix

The screenoverlay is something I've been thinking about, but as textures are going to be stretched to fit the screen, I'd either have to create massive textures that would eat a lot of (max.) bsp size, or settle for ugly stretched textures. The latter is going to happen anyway with different aspect ratios ofcourse. Any examples of how you did it are more than welcome ofcourse.

As you said, the way I did it was with huge and horribly stretched textures. It was fine for what I did (a simple prototype) but not for your purpose.

There are ways to have an almost vector-like graphics in source, with the use of $distancealpha https://developer.valvesoftware.com/wiki/$distancealpha

This fix the issue of having a massive texture (you can have fine results with even a 256^2 texture), but doesn't solve the aspect ratios problem.

Do you use vscript? There's probably a way to add a custom HUD with it. Otherwise, I was able to add a entity-based bar on the hud with an env_screenoverlay.

You should definitely use vscript if that's not the case already, I heard it's the way to go when you're making new gamemodes.

You can show the capture progress inside the game world, by adding a huge spotlight oriented toward the sky changing color, or by adding animated supplies drop in the 3D skybox, or by erecting a flag ala Battlefield.

As for the gamemode itself, it's limited by several factors (how do you acquire money, ammos, health, do you respawn infinitely, do dropped weapons disappear, etc).

The way I see it, the buyzones should be at spawn. The spawn is protected from spawn-camping, ala TF2. Players spawn with base equipment (a pistol and kevlar/no kevlar, depending on settings). Weapons drops but disappear after a few seconds. Everyone gain money each 5 seconds, but the team controlling the point gain more money. You get money by capturing the point/killing ennemies. There's a buyzone near the point, accessible only by the controlling team, so you can restock on ammos/armor while defending. The point itself is designed to provide a strategic advantage to the controlling team, either by opening doors so that respawning players may access the point faster, or/and by simply adding windows/covers looking over the ennemy spawn main entrance (a flank route should be added if that's the case).

Money is limited to 10000/16000 and you can't heal while defending the point, so it balance things a bit.

Otherwise you can change the spawn time of players, or change spawn positions, depending on whether or not they control the point.

Quote from jackophant

Has anyone got any stats for how lightmaps affect compile times?

The time is proportionnal to the total number of lightmap pixels (luxels). Which is, the surface of your lightmapped surfaces divided by the lightmap size squared.

If you divide the lightmap size by 2 (16 to 8 ), it takes rougly 4 times the compile time.

In theory you should also multiply by the number of bounces, in practice the light decrease with each bounce so the factor may be more around 2/3/4.

You can also divide by the number of cores (1 core per luxel ideally).

This is a (very) rough estimation, I'm not counting the overhead, vertex lighting, bsp complexity, lightmap parametrization, textures reflectance, etc.

Needless to say I'm waiting for a GPU based implementation.

Probably Well from QWTF or TFC.

[Blocked Image: http://www.vossey.com/gallery2/d/93513-1/Well+TFC.jpg]

The remake from TF2 is great, especially in CP (TF2 doesn't play well in CTF to be honest) but the flag room doesn't allow for impressive plays anymore. I really liked the small window, anyone who watched TF Done Extreme will know why.

Also Dustbowl and Goldrush from TF2. Especially with their uses of translucent surfaces, 1-way doors, flank routes, etc.

Life of a Party from Thief 2 is a classic.

[Blocked Image: http://screenrobot.com/wp-content/uploads/2014/01/1.jpg]

Also if anyone remember the game NoX I think the wizard guild is the best part from the warrior campaign, along with the final boss. It almost plays like an arena FPS sometimes.

Quote from Deh0lise

I've also removed the stars from the 2d skybox, made it less lowres and placed different planes of stars to give it more depth.

But why? This has to be one of the weirdest decision I have ever seen regarding skydome. You know stars are BY FAR the furthest objects in any scene, right?

I made a skydome with stars once, but it was to give an impression of rotation to the entire map. The parallax effect is even weirder when you noclip through the map.

You should made the clouds as a skydome, not the stars. Right now it looks like the entire map is surrounded by fireflies.

Maybe you should block the visibility under the tank train in B? I think the whole train won't be used because of that.

Some planks are a bit too thick, especially in the area from T spawn to A.

Anyway, it's a really great map, one of my favorite in the workshop. I'd love to try it in a 5v5.

http://image.noelshack.com/fichiers/2014/…09-23-00001.jpg

From cobblestone, near B main.

Reflex is a Quake clone, if you're denying that you're delusional. Call it a continuation of CPMA if you will.

Not that it's inherently a bad thing, I liked CPMA too. The point isn't to be original anyway, it's to provide modern features to an old and proved game.

But again, Quake Live already provides these features. Where are all the players eager to play an arena FPS? What does Reflex add to the genre, exactly?

At least games like Telos or Project Nex try to do something new, and even then I see a lot of similarities with already existing FPS.

Everywhere I look it's the same stuff. People jumping around in narrow, closed and vertical maps, shooting rockets and circle straffing while spamming their weapons.

It's hard to do something new and fresh with this genre, we'll see which game will be succesful in the end.

Quote from cityy

More movement freedom eventually just means that the game removes a depth aspect from the level design (increasing the # of possible routes and connections to take for the player). What they really should do is make the already existing movement visible to new players.

Are you me?

Strafe jumping as a mechanic is really neat, especially in games like Defrag. The skill gap is almost infinite, which is always nice for a race game.

But how are you supposed to explain something so abstract to new players? Item timing is bad enough as it is.

I don't like the physics in CPMA, they are waaay too permissive. There's no feeling of weight, there's no penalty for turning/steering. You can't easily predict the trajectory of an opponent. It's even worse in Warsow and Painkiller. At least you had to make some kind of decision on your positionning in vanilla Quake 3.

If you can go from anywhere to anywhere under 2 seconds, what's the point of even having movements or maps in the first place? Just play the game in noclip mode. It's the same thing, only less skilled. A bit extreme I admit.

Quote from SpronyvanJohnson

In fact, it's so damn good that Id Software themselves copied the CPMA physics and ruleset for Quake Live.

As far as I'm concerned, most Quake Live pro matches doesn't use CPMA physics and ruleset at all.

Quote from musly

i know this map was made to capture the flag but there's a capture the flag mode for cs go, anyway there are many solutions for the map be playable.

You started the map before knowing which game mode you'll be using? That's kinda weird, although not unheard of.

The hostage mode is already a one-way CTF, most CS maps are designed as CTF maps. You should consider it before using a custom gamemode, players (and volvo) prefers gamemode they're accustomed to.

You should use instances to rotate some of the buildings too.

First post on mapcore o/

I'm working on a cs:go map:

[Blocked Image: http://i.imgur.com/RMTgOiws.jpg] [Blocked Image: http://i.imgur.com/ihH0MFns.jpg]

[Blocked Image: http://i.imgur.com/NAdlTaIs.jpg] [Blocked Image: http://i.imgur.com/oait0izs.jpg]