Phmlspd demos (Boom, -complevel 9)

BFG UV speed in 3:59.

Seeing as I enter the final room at the second minute in the other demo, going without the BFG should be faster if done well, but getting it obviously makes the whole thing much, much easier.

ph07-359.zip

Map07 TAS UV Speed in 1:16

My first attempt at a tool-assisted demo. Pretty fun, playing in slow motion feels like dancing with demons. Watching the demo playback at normal speed almost makes me look like good, too ! :p

Strategy was to grab nothing but the SSG and green armor from the start, rush to the keys ASAP, get the cybie to infight the archviles, step back from the spider MM after activating the ten second W1 lift, go up in the final room and reach the exit with a suicide.

I planned to do this as a legit demo first, but after seeing the trouble I had even playing at 50% speed (about 12 attempts up to the cybie infighting AVs, and 3 more attempts to the exit)... Yeah, right.

ph07-116.zip

I had a blast discovering TAS this evening - beat all my UV speed times within two hours.

TAS UV Speed
Map01 in 0:50
Map02 in 1:35
Map03 in 1:09
Map04 in 0:26
Map05 in 0:36
Map06 in 2:51

Most interesting demos are in my opinion Map04, using a completely new route ; and Map06, as the "skill" displayed there is so ridiculously outside of my abilities I can't help but laugh at the result.

phtasspeed.zip

It kind of feels odd to go back to normal recording and ditch demos who went perfectly well until, well, they didn't.

Anyway, managed another UV max for Map07 in 9:49. This one looks much better and is more or less how I intended the map to be maxed when making it.

ph07-949.zip

Another map07 UV max in 9:10, picking the invul sphere to kill the mancubii faster ; though I was surprised the time wasn't much better as the run was a lot smoother than the previous one. Replaying it, it looks like I lost time killing the first two cybies, lost time waiting for the yellow key cybie to infight cacos and didn't manage to get into the first ammo cache while doing that, and probably should have killed the remains of the caco swarm with BFG rather than rockets. I'm thinking I should also kill the barons guarding the exit with the invul sphere, it's a PITA to get them otherwise as the fast lifts make it hard to use RL or BFG efficiently.

ph07-910.zip

TAS UV Max in 6:06 for Map07. It's mostly alright until the third minute then gets a bit random, I'm not really sure how to approach the final part optimally (tried a few different ways and so far always ended up with a final time over six minutes).

ph07-606.zip

More TAS UV Max goodness

Map01 in 2:07
Map02 in 3:49
Map03 in 3:26
Map04 in 5:59
Map05 in 2:56
Map06 in 6:41

phtasmax.zip

Missed 3 enemies but 7:40 on 7. I exited w/o killing the last few enemies becasue I was lazy or something and thought the time would have been worse.
For the double door, complevel 9 makes it hard to open the 2nd door ('use' action line in front blocks the use line in back).
Had 3 unused cell packs at end.

http://www.speedyshare.com/files/20307528/phmlspd-07-740-missed3enemies-ggg.lmp

For the double door, complevel 9 makes it hard to open the 2nd door ('use' action line in front blocks the use line in back).

Yeah, I agree. I'd lie if I told you it was planned, but after seeing it in action I thought it made for a nice trap, with the pressure there is at that spot.

I understand it could be seen as more frustrating than fun, being essentially an unobvious engine quirk, and using something like that definitely go against my general design philosophy which is to try to label things as clearly as possible, so I'll have to avoid doing that again in future maps.

Nice demo, it's always impressive to see you aiming your BFG shots perfectly like you're at the shooting range rather than running around at top speed in a map filled with monsters. ;)

If it was intentional then do whatever you want of course (just mentioned in case you didn't know). edit: oh, re-read that you didn't plan it.

Tried leaving the back 2 cybers to infight the cacos. 7:44 uvmax.. so about the same time.
http://www.speedyshare.com/files/20317954/phmlspd-07-uvmax-cl9-744-ggg.lmp

Another thing to watch for is walk over lines. Didn't look in doombuilder but assume there are some for the warping ledge reverents and sometimes you can pass the line without triggering it so they don't warp. I sometimes put like 4 extra walk over lines close together to prevent that possibility. Still TOD managed to wall run past 4 parallel walk over lines in one of my maps, so not sure if diagonal/non parallel lines work better.

Yeah, I noticed that too especially in Map06. I tried putting more lines close together but, as you also experienced, even with multiple lines it sometimes fail. Spending time to work around bugs gets old fast (especially while on a time constraint) so I just accept it - all in all, it can mix up things a bit adding an element of randomness, and I'm not against that, especially as you can eventually manipulate it for speedrunning.

It's weird because some lines seem to fail pretty constantly, some part of the time, and some never. That's comparing all parallel lines.

Thinking about it though... I started noticing it since I switched to Boom configuration (and generalized linedefs), and on the other hand I've hardly noticed this problem if at all on vanilla maps. Admittedly, most of my dooming prior to starting this wad was with GZDoom, and I don't think linedefs ever fail on ZDoom engines.

Have your experiences with lines failing been on Boom maps ?

Lines can fail especially when strafe-running over them. Maybe the lines that tended to fail were in areas you crossed quickly? This is sometimes exploited in short runs (or at the beginning of longer ones) and in TAS to avoid effects or quicken routes.

Right ; what I find confusing though is there's areas I'm crossing at top speed where the linedef never fails and other areas where I can cross at slightly less than max speed and they'll fail often.

Case in point, in Map06 : the linedef rising the BFG never failed me, even using SR50 or rocket-blasting myself from a pillar. On the other hand, the linedef lowering the floor to the yellow key fails much more often (I'd say roughly 1 time out of 3), with SR40 and sometimes not even at top speed.

I don't know, but in hr2 32, the warp lines often fail. And it seemed to especially occur where there's tons of enemies infighting all around and stuff (so maybe all that enemy computation screws up the trigger too, and also yeah going fast/ strafe running/wall running/sr50 seems to help a lot).

(doomguy moving straight down from here)

...............



.
 .
  .
   .
    .
     .
      .

Just a complete guess, but looking at the above, if doomguy crosses the first line (if you think of lines as a string of pixels anyway), he basically crosses the line once if moving perpendicular to it. But the 2nd line, seems like doomguy would cross the first pixel THEN the 2nd pixel and third etc thus triggering the line more than once sorta (or it might not work like that for all I know, if this even made sense).

Your explanation makes perfect sense to me. It'd seem logic for it to work that way.

I've made diagonal lines for my new Map08 and so far haven't been able to break these, though I haven't got much time to test.

FDA of 8 - failed of course.

abphmlspd81st.zip

Just recorded a 12 minute UV max in which I waited 1 minute for the last mancubus to TP in - I screwed up and forgot to put "block monsters" lines in a large mancubus teleporter. I will update a fixed version ASAP.

Edit : Updated now, in the Wads&Mods thread.

Map08 UV Max in 10:15

Rough around the edges, but a good starting point.

ph08-1015.zip

Map08 UV Speed in 2:34

I'm happy with that one, especially the start.

ph08-234.zip

Map08 UV Max in 9:48

After spending pretty much the whole evening on it I got a... 30 second improvement. :/

I've got quite a few ideas for more efficient routes but can't manage to do it. Most likely going to have to TAS it up.

ph08-948.zip

I gave it a shot again today and even managed to come close multiple times, but simply couldn't do it in the end, so went with TAS.

Map08 TAS UV max in 6:32

My regular attempt who died closest to the exit was at 9:14 with 6 cybies left, without TAS this route probably wouldn't be much of a time improvement as I'm not good enough to handle the 16 archviles and cyberdemons efficiently.

ph08-632.zip

Map08 TAS UV Speed in 1:58

ph08-158.zip

This time was ass plus I died on the last 2 cybers (8).
http://www.speedyshare.com/files/20375126/phmlspd-08-ggg.lmp

Damn. Your run looked really great for the first five, six minutes. The final cyberdemons can be a PITA, though.

Did a map 8 UV-fast yesterday for laughs mostly. Considering I can survive the very start a fair amount of the time, I'm just happy to live to the end mainly.

Time for the run is a sluggish 13:42. I could certainly improve this some more, just don't have the time right this moment. Maybe in a day or two I'll take another crack at it if nobody else tries.

ph08f1342.zip

That was quite a feat, and nice to watch. I'm also surprized you could manage to survive the start that often.

Improvement for map 08 UV-fast. 12:57 is the time now.

Basically, the txtfile says all that needs to be said about how this run went.

Think I'm going to leave it alone for the moment and await the next level to be released :P.

ph08f1257.zip

Nice. If you don't mind me asking, I'm wondering what was the use of the first invul sphere. I thought you were going to flip the three key switch and deal with all the archviles, but that you changed your mind midway, but maybe I'm missing something here. It seems to me you'd probably manage somewhere around 11 minutes with that invul against the second cybie wave (and no sneaky revenant hiding).

Phml said:

Nice. If you don't mind me asking, I'm wondering what was the use of the first invul sphere. I thought you were going to flip the three key switch and deal with all the archviles, but that you changed your mind midway, but maybe I'm missing something here. It seems to me you'd probably manage somewhere around 11 minutes with that invul against the second cybie wave (and no sneaky revenant hiding).

Yeah I'm gonna admit that invul pickup was really dumb. Not like I can't strafe around and kill everything in the center room at that point without it.

That and I was kinda on edge after a real BS death in that crusher room. Go me trying to find the last few enemies, forgot about those revs in the crusher room, and managed to rambo charge in at the last possible second I could fit under it and get crushed :P.

About the archviles...part of me wonders if UV-fast would make an invul. basically useless being you'll likely be blasted by half of them or more just trying to get near.

I do basically want one invul for the first half of the final room. Otherwise that cybie guarding the switch to release the exit room cybies would take forever to kill I think.

It'll probably take a lot of patience to get a good run out of this level UV-fast style, given the luck I need for basically the first half or so.

For now, I might as well post this other UV-fast I did around the time map 6 was released, just never got around to making a txtfile for it.

It's for map 6 in 13:21. Could easily be improved.

...And no, I have no idea why I just sat there after taking cover in the RSK room before grabbing the invul. :P. Probably a lovely whoopsie moment.

ph06f1321.zip

About the archviles...part of me wonders if UV-fast would make an invul. basically useless being you'll likely be blasted by half of them or more just trying to get near.

Seeing how hard it is with all 16 AVs on normal UV, I think you're most likely right ; though I wonder if it'd be possible to make the HKs/barons infight enough AVs to make the situation manageable.


That Map06 -fast demo was sick, especially the way you manage health so well at times ; the only bit I found slow is the end, as you used RL against the last few archviles rather than pick up the BFG charges scattered in the room, but it's obviously much easier to say that watching the demo than to actually do it. :)