AdamW

uploaded a 4:44 speedrun of map23, just for fun.

oh, just to respond to the point raised earlier - in map23, the cacos at the window before you go down the lift to the final area are the SAME cacos flying around and bothering you while you're in the final area. so it's fine to leave some alive as you go down the lift. would actually have been faster to kill more of them upstairs, I now realise, but never mind. :)

ok, did map23 in 14:11...sorry erik :). This is still messy, it's a messy map, and my ammo use at the end is terrible (should've saved some BFG shots), so there's another three or four minutes in this I guess. The yellow key area is particularly annoying to do quickly. I made the start a lot faster, that's where most of the time comes I think.

think i'm happy with mine...i've been trying to improve h223 lately, but i don't mind if i don't get that done.

ok did map21 again, now I think I'm happy with it...this one looks like I actually know where I'm going :D

little upgrade for h221 to 5:56...only six seconds but hey, it's a slightly different route and gets it under 6 mins.

there isn't one afaict vince. sorry!

think 7:46 is correct - that's the time on the demo and the zipfile, and I do those first...I only wrote 7:43 in this thread, I guess that was a typo.

op: what's with the time on hr03 on your page? Did I report it wrong? oh, and the map03 run is doom2.exe not prboom. also the map01 pacifist run. actually i'd rather you pretend that one doesn't exist, since it looks so bad - i just wanted to record a demo with doom2.exe under linux back when my emulation was still pretty bad quality (I have a way to get it really nice now, hence the decent MAP03 run).

ok, uploaded a new h203 in 7:43, looks a bit more businesslike :). this was recorded with doom2.exe, as well - I've got it running very nicely under Linux with dosemu now...

source port demos don't qualify for compet-n.

Hmm. On your page, you say "Also, i do belive that these are one of the very first speed demos for doom2 ever to be showed on the internet." Well, not really. COMPET-N was founded long before 1996.

btw, i'm fairly sure my map07 jump makes the level possible...pacifist :). it'll be pretty hairy once the blue key archies get released, but I think still possible. I'll work on it tomorrow. Bet that wasn't one the team expected to see :D

just uploaded a snazzy little trickjump i found for MAP07...don't know if it was known before, was this map in the beta or anything? I like it anyway. could help a speedrun. hr207trik-adamw.zip is the file.

and 3:03 for map05. yeesh, when do I get to the maps with some frickin' health and ammo in them?! :)

ok, 8:36 first exit for map03 uploaded. Ugly as hell, but this is a horrible map again, so it'll do for now.

ok, 2:28 for MAP02. Horrible little map. Occasionally you can telefrag the baron at the start, which helps a lot.

ok, uploaded an ugly 1:46 I can't beat at the moment. onto 02 I guess...

ok, uploaded 1:58 for MAP01...still nowhere near optimal but a bit more businesslike :). want a go, altima? all my demos will be prboom 2.2.4 complevel 0, btw - running Linux. well, i HOPE i had the complevel set up right, anyway...

just did 0:12 for MAP01 pacifist...under DOS emulation in Linux, with screensize down two notches and low detail enabled. fuzzzy! :D

hmm, probably the least commercial commercial pack ever made...was designed by a bunch of enthusiasts (none of us made a bean on it, btw, the project leader lost nearly five grand) and only sold about 15 copies *ever*. hehe.

oh btw, i posted this in the first day thread but seems no-one else is reading that, so can i point out that memento mori 1 should be in 1995, not 1996? it came out december 1995, I remember playing it on Christmas Eve. oh, and I always find the best way of flicking perspective in demos is just to play 'em in prboom - then the F12 trick works and also switches the status bar and sound effects. yay for prboom.

haha yeah, I did once figure out roughly how many attempts I've made at episode 1 overall, it was a very scary number :) yeah there's nothing wrong with TAS runs at all, just that they're not supposed to be sent to COMPET-N :). I love TAS runs actually, you can do lots of stuff you'd never manage in a real run - the 18 minute nightmare run of Doom2 is one of my favourites, as is Sedlo's sadly unfinished TAS run of Plutonia... and yeah for doing MAP07 speed now you don't need to kill ANY arachnotrons at all, hehe. that's an absolutely sick run, btw - it was known that the trick was theoretically possible for a long time beforehand, and me and adamh both (and some others) spent *ages* trying it, but of course it was sedlo who managed it...but what completely blew me away is that he did it in 8 seconds and decided that wasn't good enough to upload. only sedlo...:D. of course, it's so hard no-one would use it in the episode run. i don't think anyone but sedlo has done it yet, actually.

cycloid: "blimey!"? aw, thanks :D you'll enjoy 30nm4949. it's definitely worth watching henning's most recent run too, you'll see how much has changed to cut the time down by 15 minutes. as for the cheatedness of the girlich / winterfeldt stuff, well it's a while since I watched 30nm6520 so I can't give you specific stuff from that, but I can mention some others...for instance, Uwe released a run of e1m5 in 0:59 in 1996. The closest legitimate time was 1:08, and it took nearly four years for someone (well, er, me) to replicate 0:59 without cheating. If you watch Uwe's demo, he hardly even straferuns at all, and to a speedrunner (I don't know how obvious it would be to someone who doesn't run) it's obvious at certain points that he's playing in slo-mo - he pauses for tiny split seconds at crucial points, not long enough to notice something in real time but definitely long enough to notice something (an enemy movement, normally) in slo-mo. Much more obvious - by far the most obvious of the lot - is a speedrun of Doom2 MAP01 to MAP10 Uwe did. There's lots of stuff that's very obviously cheated in it, but the most blatant is that in MAP07, every single arachnotron is killed with just two rockets. Normally it takes three rockets to kill an arachnotron. Occasionally (because of Doom's partially random weapon damage) you can kill one with two rockets - someone's calculated the exact probability somewhere, but I reckon it's something like one in ten or one in fifteen. If we take one in ten, there's twelve trons on MAP07, so the chances of them all dying with two rockets in an uncheated run is 1 in 10 to the 12th power, which is an EXTREMELY large number :). (In all its glory, it's: 1,000,000,000,000 - one trillion). So it's basically proven that Uwe must have used some kind of cheat, most likely some form of TAS executable which allowed him to stitch runs together like tasdoom and mbf do nowadays, to make that run. There's lots of other stuff he does in the run which would be very hard to explain in any way other than "he was cheating", but that's the most painfully obvious.

so I downloaded all that 18MB for nothin'? typical. bloody zdoom. /me taps foot waiting for the Linux port of zdoom 2.0...