Map assessment and pathing enhancements

[Nelsona]

Yes, all sort of more or less lucky shots, not uber accurate, these can be seen in various SLV maps with a complex geometry... but maybe this simple map should be a sort of guiding for improving SLV attached navigation, I don't think anyone is working at this chapter. Either way in maps over charged with added specific nodes SLV might crash because of iterations - SLV in my opinion it's not really ready but worth playing in simple and less charged maps.

[evilgrins]

I'm actually thinking of asking you to remove SLV pathing from a different map, one you previously edited, but time enough for that later.... the flapjacks one.

I've got some other simpler SLV to check out later.

[Nelsona]

FlapJack has indeed native air paths... and removing them it's simple.
- UnDefine Paths;
- deleting nodes claimed UnReachable/Aerial Placement - shown by MapGarbage (except teleporters and other similar ones);
- saving a copy-map and closing Editor;
- Re-Opening everything;
- Sending Paths Define command.
It's all about 5 minutes of work.
Now enjoy crashes in walls mainly in all RX/SLV mods - guaranteed. RX09 mod has a bit of CTF, tested in map with native air support, as for others... say a pray.

[evilgrins]

This is something simple:
CTF-UTDM-Cannons-TwinValley2_rS2189.7z

Slight hole in the cloud.

Outside the base, under the ramp, that circled doorway? It's on both Blue & Red sides and if a UTDMS gets in there it never leaves. I'm assuming that's because there's no pathing in there at all for it to follow to leave.

[Nelsona]

I see... let me look at specs data... it's a bit of hunting because I cannot get all data in Editor, only in run-time, and then I have to record nodes names from area - perhaps it's the entry/door. Pawn hunting was moving inside but reachspecs are too small... Sometimes Editor does this funky stuff... reachspecs from A to B is larger than from B to A due to walls calculation, definitely movement should be the same in plain fields, lol...

I'm assuming that's because there's no pathing in there at all for it to follow to leave.

Nah, Editor did paths there but are too small:

BottleNeck_Specs.PNG (1.27 MiB) Viewed 2864 times

In image it's about PathNode65 and PathNode67 - they are linked Each-Other but let's see collision accepted:

Code: Select all

ReachSpec: 306 : Start = PathNode67 End = PathNode65 ColRadius = 70 ColHeight = 40 ReachFlag = 1 Dist = 547 Pruned = 0.
...
ReachSpec: 1156 : Start = PathNode65 End = PathNode67 ColRadius = 70 ColHeight = 40 ReachFlag = 1 Dist = 547 Pruned = 0.

Technically Skaarj was fighting and charging at enemy discarding navigation, but later... it's too big and reachspec doesn't accept it.
And on the other side it's Blue Base with PathNode152 and PathNode153:

Code: Select all

ReachSpec: 561 : Start = PathNode152 End = PathNode153 ColRadius = 64 ColHeight = 40 ReachFlag = 1 Dist = 510 Pruned = 0.
...
ReachSpec: 563 : Start = PathNode153 End = PathNode152 ColRadius = 64 ColHeight = 40 ReachFlag = 1 Dist = 510 Pruned = 0.

This second one is even smaller than the one from RedBase, lol. These were added in Goblin Style but using a low Range Scan of 810 if I well recall. I would like to edit manually these 4 reachSpecs but I believe it won't load in 227 because of UT specific assets used. I need to export UT stuff first...

First measure would be aligning these two nodes perfectly at grid attempting to gain maximum for collision data... If this solution won't help, both Skaarj should be rescaled to a lower DrawScale in order to allow them to adjust their body themselves.
Another way but... I don't know more about results, would be using forced links with fake teleporters which are causing entrance of a 150 × 150 creature.

Let me try grid-align solution first.

Edit:
GridAligning nodes returned a 70 radius × 44 height allowed. I believe rescaling Skaarj to 0.7 would help as long as there is not much different data by aligning nodes. If this won't help, it's needed replacing nodes with blind Teleporters...
This is a version with said nodes aligned at grid and Skaarj having 0.72 DrawScale. For me it was operational.

CTF-UTDM-Cannons-TwinValley2_rS2204.7z

(731.3 KiB) Downloaded 121 times

[evilgrins]

I don't want this edited at all, but you can for your own purposes if you want to. Just after we finalized those 2 versions of this map I recalled there was a 3rd and thought maybe you'd want to take a look at it. I had this map a ways back but I didn't keep it, but it popped into my head just a sec ago.

There's downloads at the start of the thread:
http://www.unrealkillers.com/showthread ... lley-Siege

I suspect a number of the reasons I didn't keep it may go a long way to complaints you'd have. This guy edited the map for Seige but he added things to it that seemed to confuse the bots in certain areas.

[Nelsona]

Yep... not a single time I could see good a CTF map turned into trash for the sake of Siege - I even know one guy coding some primitive-mesh things (not finished) good to ruin games with monsters for this SG game which I don't play and I think even others are not really fascinated by that thing especially where some idiots are very "cute" and delivering a nice "welcome" to new comers in servers - "impressive" community - maybe that thing would be more welcomed generally without those "south" assholes "playing" in servers... those are anti-social in true terms not others as claimed on Discord.

Ignoring all Siege specific locations, let me take a look at what was destroyed in said map....
Checking...

Edit:
Yeah... both FlagBase actors were moved away behind that panel having ZERO paths point to/from them - to not forget a few useless PathNodes grouped and isolated from Navigation network. It can be turned back to CTF if anyone else wants a CTF here. For me is enough what I have.

Edit2:
With regard to chapter map checking (it's another map on the table) - I had to complete more reporting toward AlternatePath detector from MapGarbage - feature available in next release:

Code: Select all

CTFReport: --- Reporting AlternatePaths setup: ---
CTFReport: For Team 0 = 116 pieces.
CTFReport: For Team 1 = 112 pieces.
NotSoUsed: 25 AlternatePath actors are having connecting problems with the rest of Navigation Network.

25 such AlternatePaths (from that insanity) are mainly useless because they are not having enough paths linking them properly... Else there are other ones more or less usable. Either way that map has 10664 registered reachSpecs - "Mapper" doing this mess was that one getting used to spread ruined textures in multiple versions being busily engaged in a never ending mismatch versions quest, for the sake of making UT files... I'm curious about the day when will get back for keeping to ruin assets and claiming that he is loving UT...

[evilgrins]

All that, and... as you know the flag area has 2 tunnels that 1 goes into the mountain and the other goes forward to a ramp that leads out to the front of the base. Even the original map was pathed up well enough for that... but this 3rd the guy sealed the tunnel but didn't alter the pathing, so bots practically kept running into that new wall on both sides.

[Nelsona]

That is what I want to see, I figured modifications but not that wall... or it was another download from MediaFire, maybe a rebuild one.

I think I remember a Deck16 lousy version having all nodes in original but geometry changed and making a mess called "edited map", I don't get these "masters" and their habits. What are for these files after all ?

[evilgrins]

Are you sure it was still a "16" because there's a lot of Deck maps, all variations on the same theme.

For example, Deck-32...