ALL_1949 by Yonatan Donner, Matthias Belz and Nolan Pflug, 5 June 1997. ------------------- Maps: START (untimed), E1M1-M7, E2M1-M6, E3M1-M6, E4M1-M7 and END as a constructed continuous run Skill: 3 (Nightmare) ------------------- See later statistics section for kills and secrets details ------------------- Time: 19:49 ------------------- Demo running time: 27 minutes ------------------- The individual episode runs are also available separately: EP1_0324, EP2_0601, EP3_0432, EP4_0516. There is also a movie version of this run, QDQ_1949. You can get them from ftp.cdrom.com, or from the QdQ page at (0) Index (1) Basics (1.1) Statistics (1.2) Overview (1.3) Credits (2) More details for those who want to read a bit more: (2.1) Timing (2.2) How The Run Was Recorded (2.3) Biographies (2.4) Project History (2.5) Information About The Levels (1.1) Statistics All maps were run on Skill 3 (Nightmare) ------------------------------------------------------------------------ Level Map Time Secrets Kills Player ------------------------------------------------------------------------ The Slipgate Complex E1M1 0:30 0/6 16/42 Yonatan Castle Of The Damned E1M2 0:38 0/3 10/42 Yonatan The Necropolis E1M3 0:43 0/3 7/65 Yonatan The Grisly Grotto E1M4 0:28 0/3 6/59 Yonatan Gloom Keep E1M5 0:22 1/5 2/60 Yonatan The Door To Chthon E1M6 0:12 0/4 1/31 Yonatan The House Of Chthon E1M7 0:31 0/0 21/21 Yonatan ------------------------------------------------------------------------ Episode One: Dimension Of The Doomed 3:24 sub-total ------------------------------------------------------------------------ The Installation E2M1 0:10 0/7 2/46 Matthias The Ogre Citadel E2M2 1:00 1/3 18/48 Nolan The Crypt Of Decay E2M3 0:44 0/4 10/48 Yonatan The Ebon Fortress E2M4 1:31 2/4 14/78 Yonatan The Wizard's Manse E2M5 1:07 0/2 13/61 Yonatan The Dismal Oubliette E2M6 1:29 0/2 14/85 Yonatan ------------------------------------------------------------------------ Episode Two: The Realm Of Black Magic 6:01 sub-total ------------------------------------------------------------------------ Termination Central E3M1 0:48 1/5 24/67 Yonatan The Vaults Of Zin E3M2 0:57 0/3 8/47 Yonatan The Tomb Of Terror E3M3 0:34 0/2 7/47 Nolan Satan's Dark Delight E3M4 0:49 0/4 12/43 Yonatan The Wind Tunnels E3M5 0:51 0/4 3/50 Yonatan The Chambers Of Torment E3M6 0:33 0/2 1/73 Matthias ------------------------------------------------------------------------ Episode Three: The Netherworld 4:32 sub-total ------------------------------------------------------------------------ The Sewage System E4M1 0:48 0/4 10/56 Matthias The Tower Of Despair E4M2 0:49 0/5 9/43 Yonatan The Elder God Shrine E4M3 1:21 0/3 17/105 Yonatan The Palace Of Hate E4M4 0:33 1/5 2/78 Matthias Hell's Atrium E4M5 0:16 0/4 0/71 Nolan The Pain Maze E4M6 0:26 0/4 3/81 Matthias Azure Agony E4M7 1:03 0/9 16/80 Yonatan ------------------------------------------------------------------------ Episode Four: The Elder World 5:16 sub-total ------------------------------------------------------------------------ Shub-Niggurath's Pit END 0:36 0/1 15/25 Yonatan ------------------------------------------------------------------------ Complete run through Quake 19:49 total ------------------------------------------------------------------------ Average time for a level: 0:44.04 12 grenade-jumps and 20 rocket-jumps helped us to achieve this. Average time for the episode 1 levels: 0:29.14 Average time for the episode 2 levels: 1:00.17 (thank you, John) Average time for the episode 3 levels: 0:45.33 Average time for the episode 4 levels: 0:45.14 Average time for American McGee's levels: 0:36.00 Average time for Tim Willits' levels: 0:37.57 Average time for Sandy Petersen's levels: 0:44.67 Average time for John Romero's levels: 0:52.78 (1.2) Overview This is a complete run through Quake on Nightmare skill in less than twenty minutes. (The precise way we calculate times is explained in the "Timing" section.) We're very proud of it. It was constructed by pasting together recordings of the individual levels that form a continuous run. More details of how we made it can be found in the "How The Run Was Recorded" section. We hope you enjoy watching our demos and movies. Although we think this is a very good demo, we know it can be improved still further. We want to encourage other fine speed-runners to help us to record an even faster run, and we intend to run a large challenge with prizes as part of this effort. As a start, we are running an E1M3 challenge. This is almost, but not quite, exactly in the style of the Ion Storm speed-run challenges on E1M5 and E2M5: run the level as fast as possible. You will have to be pretty good to beat Yonatan's time of 0:43, but we encourage you to try. Read the exact details of the challenge and download the challenge patch from . If you are interested in single-player demos in general, including 100% kills and secrets demos as well as simple speed-runs, be sure to check out Nolan's Demo page at . (1.3) Credits The four individuals who are responsible for this recording are: Yonatan Donner Matthias Belz Nolan Pflug Anthony Bailey Yonatan, Matthias and Nolan recorded the level demos that the recording is composed of. To discover who recorded what, see the statistics section. Anthony Bailey provided technical support in the form of the utilities that we used in processing our demos and making the continuous run. (Mostly DemoRelise.) All four of us worked together in coming up with ideas for the strategic aspects of this run (decisions about what routes to use, what items to collect and what to avoid, ideas for special tricks and short-cuts.) We also made use of LMPC and the accompanying unofficial demo specifications by Uwe Girlich, and Film At 11 by Eric Stern. Our thanks to them for making these resources and tools. And of course we will never be able to thank id software enough for Quake itself. (2.1) Timing We should probably clarify exactly what we mean when we say that our run of Quake is performed in 19:49. There are several possible ambiguities to deal with when timing a demo, particularly one spread across the whole game. We have settled on what we think are the most sensible ways of resolving them. Our aims were to make the timings sensible, verifiable from the finished demo, and given those two criteria, as convenient as possible to calculate. We hope others will use the same methods to calculate their own times for individual levels, episode runs and game runs, so that it will be easy to compare different runs. (2.1.1) Timing a regular level Of course, we use the Quake's built-in timer. In fact, this makes the times longer than they really should be, since the timer starts at 1.39999998 seconds at the beginning of a demo, but there you go. Normally we just read off the time printed on the intermission screen. These are always a whole number of seconds. We ignore any residual fractions as it takes extra work to discover what these were, and as already noted Quake has already stolen 1.4 seconds from us, so it hardly seems unfair to take a fraction back again. (c: There is however a possible ambiguity when Quake times a played-back demo. On playback, when a level is completed, sometimes the timer may tick on to the next second before the engine reads the time following the start of the intermission. This causes Quake to display a time that is one second larger than the time shown when the demo was originally recorded. Whether or not it occurs depends on how close to the next whole number of seconds the demo time actually was, and it also depends on the speed of the computer you are playing the demo on. So a slow computer may display a different time than a fast computer. This needs resolving since we want times to be definitive, not dependent on the computer the demo is playing on. One the few occasions when it is necessary, we settle the matter by looking at the demo itself with an external tool (such as "Film At 11.") The time we use is the time marked on the frame of the demo before the start of the intermission. This always gives us the time that would have been shown when the demo was originally recorded. So, if your computer shows a time for a level that is one second bigger than the one we have stated, that is the reason why. An example of where this happens is E2M5, which we time as 1:07. The time in the demo is 1:07.97866058m. This is close enough to 1:08 that most computers will probably display a time of 1:08 when they play the demo back, but it really is a 1:07 demo. (2.1.2) Timing START We don't bother to time the sequences on the START level. These sequences don't really seem like important parts of the proper game, they are just a chance for everyone to catch their breath. (Quake doesn't show any time at the end of them, as we're sure you know.) So, we didn't make an effort to run them as fast as possible. It would be trivial but very boring to do them perfectly, but really, what's the point? (2.1.3) Timing an episode The time for an episode run is simply the sum of the times for the individual levels. Incidentally, we don't consider an episode to be completed unless the player exits with the rune. We mention this because there is a design flaw in E2M6 which allows the player to reach the exit without picking up the rune - you can rocket-jump over the half-barrier into the top of the exit. This means you can get to the exit slightly faster if you ignore the rune. However, we decided that this didn't count as completing the episode since the whole point of running an episode is to get the rune at the end of it. And it certainly wouldn't work for a whole game run! (2.1.4) Timing END The END level doesn't display an intermission screen with a time on it upon completion, because Quake instead plays an end of game sequence with Shub exploding. However this is obviously a level which is worth timing, so we inspect the demo to calculate a time for this level. The time we use is that on the frame before the first finale text. This is the time at which the player enters the final teleport and at which Quake cuts to the intermission point. (2.1.5) Timing the whole game Of course, the time for the whole game is the sum of the times for the individual episodes plus the time for END. (2.2) How The Run Was Recorded This is *not* a recording of a single session of Quake. We'd sure like to meet the person who could play as well as this through a continuous twenty-minute run! It might be an interesting idea to try, but unfortunately there would be no way to be sure from a final demo whether or not it was recorded in one take, because the Quake *.dem protocol is inherently unsafe that way - and indeed every multiple-level demo recording must be hacked slightly in order to get it to play back correctly at all. Instead, this demo was created by recording a demo for each level individually and then using a utility to paste them all together. However, we were careful to ensure that the finished product really is a continuous run through the game. We start each new level with the correct statistics relative to the previous level. (Basically, these will be whatever we finished the previous level with, except that the Quake engine always gives you a minimum of 50 health and a maximum of 100, and a minimum of 25 shells, whenever you start a new level.) Continuity is very important to us; a demo made simply by pasting together unrelated recordings of each level in turn would not be a proper run of Quake. Continuity adds a lot of strategy to the game, since items that one finds in earlier levels can be used in later ones, but those that one avoids to save time cannot. Is it worth the deviation to get that grenade-launcher? Can one afford to sacrifice health at some point to save time? Many subtle interconnected decisions have to be taken. It's great fun! Since some degree of demo hackery is necessary to make a complete run that plays back, and since this run features a great deal of absolutely first-class play, people may think that what we have produced is some hacked-up cheat. It isn't. The demo for each individual level *was* recorded in a single take, without any cheating or hacking. It was very important to us to make a recording of spectacular human play, not of some bot-driven hackery. This is as close to a "proper" recorded run through Quake as we believe it is possible to get. If you find any continuity problems with our demo, *please* tell us. We don't think there are any, but we will want very much to correct any mistakes that are found. Thanks! To record levels starting with the correct statistics, we used a patched version of the Quake-C module. There are details of how to record multiple-level runs in this way at . Also, we will soon be releasing a patch that allows you to select arbitrary starting statistics from within a single Quake session using an menu of options. This will make life easier for anyone who wants to record similar multi-level runs. Sometimes we wanted to use a demo we had previously recorded, but to run it relative to different starting statistics, because we had improved our performance on an earlier level. In order to adapt a demo in this way, we built a special utility called DemoRelise. This takes as input two demos of consecutive levels, and relativises the second one with respect to the statistics at the end of the first one. This kind of relativisation will not always be possible; for example, suppose a player uses some ammo in the second demo that they would not have following on from the first. In such cases, DemoRelisation will just produce an error, not an invalid demo. The author was very careful to make sure that DemoRelise produces correct demos, and it gives warnings about anything that is concerned may cause a problem. DemoRelise will be made publically available, but be warned that it is a reasonably complicated program and so its author does not intend to offer any technical support for it! (2.3) Biographies Yonatan Donner When the first Quake Shareware came out, I was already a Doom addict, so I downloaded it. At first I didn't like it, so it was some time before I decided to get the full game after all - only because I trust id too much :) For a long time I wasn't a real player, just watched some demos occasionally, until after some conversations with Nolan I started playing myself and started to like the game. So I owe this extra waste of time for yet another computer game to Nolan :) Then I discovered it was possible after all to record Quake demos for more than one level, and recorded a simple run through of Episode 1 on Nightmare. I hoped to improve it one day. Then Anthony wrote me about it and some time later I asked Matthias and Nolan if they wanted to try it for the whole game. A short time later we started this project, which took about two months, and was lots of fun, and I think the end result is worth it. Thanks to Nolan and Matthias for doing this with me, it was a pleasure working with you guys! And of course big thanks to Anthony, without whom this demo would have been a simple run and not a real movie, and probably 5 minutes slower without all his suggestions and shortcuts. I hope you'll enjoy this demo as much as we enjoyed making it... Personal info: I'm 16.5, I live in Rehovot, Israel. My home page is at: Matthias Belz I am a fan of 3D shooters since I played Doom on my 386 back in 1994. I especially liked to watch those great speed demos that were made for Doom and Doom2. I eagerly awaited the release of Quake and bought it right when it came out. I waited for Quake speed demos to appear, but to my disapointment there were very few of them. Nolan seemed to be the only player who constantly released demos. So when I discovered a nice route for e1m5, I decided to try it by myself. Some more demos of levels where I had found an easy route followed. I thought this would maybe cause other (better) players to try it too, but only a few more players sent in speed demos since then. A few weeks later Yonatan released his episode 1 run, and asked me if I'd like to work together with him and Nolan to create a better one and maybe even runs for the other episodes. Of course I agreed and you can see the result here. I hope you'll like it. Personal info: I'm 23 and study Computer Science at the University of Karlsruhe, Germany. Nolan Pflug I started recording Nightmare Speed demos after I found ones at ftp.cdrom.com for e1m1, e1m2, e1m3, & e1m5. I was very impressed by them. Jeff Preshing, who had made some of those demos, had a chart on his web page showing the best times. I picked e2m1 and recorded it in 2:35 with all the kills and secrets. I emailed Jeff about it, but he didn't seem to care anymore. I continued to make NSDs, thinking there was only one person who was watching them: Yonatan. Soon, demos without all the kills and secrets started appearing, Matthias was making them. Then one day, Yonatan made ep1-0444, and soon asked Matthias and I to start a project to do all four episode runs. I decided that there should be a chart to show the best times, like what Jeff used to have. I asked Matt of Single Player Quake if he'd be interested. He said he was going to make one, but after a month of waiting, I got fed up and made my own. It's up at BTW, in case you haven't noticed, I'm a keyboarder! I've impressed many people with my keyboarding skill. (Not my words per minute, mind you.) I'm 17 and live north of Pittsburgh, Pennsylvania, in the US. Anthony Bailey I'm the token crap member of our group. I had great fun playing through DooM II and Quake, but I'm no gaming god, and I found it all pretty challenging. So it's been a real pleasure to work on this with such cool people and players as Nolan, Matthias and Yonatan, and to bask in their reflected glory. (c: I get away with it by programming some tools and utilities for them. When I saw Yonatan's EP1_0444.DEM, and realised how much room there was for strategy and inventiveness in making speed-runs across multiple levels, I immediately caught the bug. Of course I still have to run the levels by proxy because I'm too crap to do them well enough myself. Maybe one day I'll wake up and be able to grenade-jump properly. (c: I'm 26, and coming to the end of a maths PhD in Manchester, UK Home page: (2.4) History Of The QdQ Project (Here's a brief overview of how this demo/movie came to be, by Anthony:) Speed-runs and other clever demos have always been my cup of tea. And when one day I saw this recorded run all the way through Dimension Of The Doomed called EP1_0444 by some guy called Yonatan Donner, the first thought was "now, this is cool" and the second "I wonder just how fast it's possible to do this... or even the whole game?" Figuring I could maybe barter some programming skills and ideas for a piece of the action, I wrote Yonatan a couple of months back with a few ideas about how the time might be improved. "I reckon episode 1 could be done inside four minutes, just about, by someone with enough talent", I e-mailed him. As you can see from the final 3:24 time, this was wildly pessimistic! Rather more optimistically, I also guessed we might squeeze a run through the whole game under twenty minutes. At the time I had no idea just how hard that would be, but I guess that time became a sort of Holy Grail for us as time went on. I'm glad we managed it in the end. Ticking clocks had begun to haunt my sleep. (c: Yonatan contacted Nolan and Matthias, since they were also making speed-runs through individual levels and sending them to Walnut Creek, and so we got down to planning how to run through the whole game and seeing what routes and shortcuts we could come up with. Figuring the episodes were essentially independent, we started with the M1s and went from there. We worked out how to patch Quake to make the consecutive recordings, and I hacked up a utility that would let us adapt existing runs to use different starting stats in case we needed to do that. About a month ago, we had a run for Episode 4 that we reckoned was about as good as it was going to get and we noticed that the QuakeLab MultiMedia contest had opened its doors to simpler demos as well as to more elaborate movie productions, so figured we might as well enter it. We wondered if maybe we could alter the camera position in our demos to make them more movie-like to watch. It was just a passing fancy but after I'd written a simple program to do that sort of thing, we all got fairly into the process of filming demos. It's great fun, and if you want to film your own demos, give the utility that these initial experiments grew into a try (it still takes some nous and patience): . All the way through this project I have been outrageously demanding about what these guys should try to achieve considering that I can't play for toffee. And they have constantly surprised me by coming up with the goods and then some. Now we finally have a complete run under twenty minutes. We hope everybody else reckons that it's as cool as we do and enjoys watching it. But the challenge doesn't stop there, and I'm looking forward to being even more demanding for even better times as we extend the project to get the rest of the Quake community involved. Stop by our page at if you want to offer us your ideas or your playing skills, or if you want to win some stuff. (2.5) Information About The Levels Below each of us talks a little about the levels that he did. E1M1 : The Slipgate Complex (Yonatan, 0:30) A rather small level, the runthrough is simple. However it is hard to survive with all the humans firing at you, and not delaying is very hard, if someone is on your path, he must be dead before you get there. I think 0:30 is the optimal time, I wasn't able to do it faster even on skill 0. E1M2 : Castle Of The Damned (Yonatan, 0:38) A very nice level, running through is simple here too. There is only one critical area, it's the switch that raises the bridge to the silver key, you need some luck to have a clear path to the switch and back, the knights can get in the way. The level is pretty easy. E1M3 : The Necropolis (Yonatan, 0:43) This was by far the hardest level for me in the whole game. At first I tried the conventional route, taking the key and then going all the way back to the door, then I thought I could perhaps get back up right away. I managed to do it with a rocket-jump, so I knew a grenade-jump could be done too. But nothing prepared me to just how hard it is. First of all, throwing the grenade is tricky: it can easily bounce off that platform. You can throw it up, but then it's harder to jump in the right moment. The jump has to be extremely accurate, or else you won't get up. You have to be running up the slope at the moment you do the jump - this gives you extra upward velocity. I managed to do it once every one hour of trying, or so. Afterwards there is another possible trick which Matthias found, doing a grenade-jump over the opening to the sewers, which saves some seconds. However I never had enough health to do it. The last trick is not getting locked in the lift in the end. It takes it a very long time to rise, and I didn't want to waste that time. Anthony suggested a long time before I tried it that I try and find a way to get out of there (then go up using the teleporter), maybe using an explosive. I found a better solution: use the fiend to stop one of the bars. It's hard, because the fiend has to cooperate with you and get right into its position to block the bar. In this try I was lucky and he cooperated with me and got very fast into his position, usually it takes several seconds of luring. To sum it up, the level itself is easy except for the tricks which are damn hard to do. E1M4 - The Grisly Grotto (Yonatan, 0:28) The grenade-jump to the silver key was suggested by Anthony. It is easy to do, the optimal jump is to land right on the key. There is one ogre in the end who blocks the way, so I had to kill him. The disadvantage of this route is not getting the super nailgun, but it's not very much needed anyway. The level is quite easy. [Stop Press... we have a new idea now of how to do this level even faster. - Anthony] E1M5 - Gloom Keep (Yonatan, 0:22) This is a different route than the standard "from-scratch" route. It was suggested by Nolan and Anthony. It is quite hard, the first grenade-jump is always successful but the second one is hard to do without getting hit, and then you mustn't get hit until you get the rocket launcher and do the rocket-jump. Even so, if you get more than 94 points of damage during the two jumps, you have to take the healthpack. It is possible to do it, I've managed to do it on skill 0, but never with everything that went around me. It might just be too hard. If it works, the time can be maybe 0:20, which is probably faster than what can be accomplished with the standard route (Shadow's) on skill 3. E1M6 - The Door To Chthon (Yonatan, 0:12) The rocket-jump to the key was originally done by Matthias. I knew I had to leave with a good health so that I'll be able to do many rocket jumps in E1M7. It takes some luck to get both 0:12 and a good health, but it was quite easy, took very few tries. The key is to have the shambler fight an ogre and to get both ogres after the gold key into pain, when they don't attack you for a brief moment. E1M7 - The House Of Chthon (Yonatan, 0:31) This isn't very much the standard route. It was first done by Nolan. I first jump to the rune and to the right, so that I won't waste time with a normal jump. Then I rocket-jump up and do 3 more jumps later, from one switch to another. This saves a lot of time usually wasted running around from one switch to another. It is quite hard to get all rocket-jumps go well without getting hit from Chthon. [We had hoped we could use even more rocket-jumps, to go from the third switch back to the first again. But if you are that quick, the pillars seem to get in the way of your jump across the lava. - Anthony] E2M1 - The Installation (Matthias, 0:10) This is probably the easiest level to run fast. One jump of medium difficulty is all you have to do. Nolan first made a 0:11 demo for this level, I improved this to 0:10. I think this the optimal time, so this is one of the few levels that cannot be improved any more IMHO. (The other levels of this kind are E1M1, E1M6 and E4M5.) E2M2 - The Ogre Citadel (Nolan, 1:00) There were several possible routes for this level, but I had to get the grenade launcher so we could do Matthias's grenade-jump in E2M3. I imagine under a minute is possible with this route, but I got help from an ogre as it was... :) [There's another route that gets the grenade-launcher and is probably a bit faster, but it requires two tricky grenade-jumps and more importantly, misses the super shotgun. And that weapon is always very useful, so we decided we had to get it, really - Anthony.] E2M3 - The Crypt Of Decay (Yonatan, 0:44) I used to think this level is extremely hard, mainly because of the deadly end. But now I find it rather easy. The grenade-jump shortcut was discovered by Matthias. It is quite a hard grenade-jumps, almost maximum height. In the end I had to kill the first hell knights to open the exit and then wait for the ogre and shambler to come out, because they block the passage. I got quite lucky in this try and they got out quite fast and I managed to sneak to the exit with quite good health and armor. E2M4 - The Ebon Fortress (Yonatan, 1:31) This is the hardest level to survive in the whole game, considering I wanted to leave with lots of red armor for E2M5, because of the two rocket-jumps and the underwater ride. I took the super nailgun, because it's very useful and very needed. Then I took the yellow armor, because the rest is extremely hard to survive without it. There are several possible routes in this level, this one was done first by Nolan in a skill 0 speed demo, with some small changes by me. The silver key area is very hard, getting the silver key and jumping into the slime alive is the first hard part, then getting out of the slime, you will be bombed with tons of grenades and shambler lightnings. I decided to get out of there with a rocket-jump, because the damage taken from the rocket is much smaller than the damage caused by the environment when doing it the normal way (regular jumps). Then I took a health pack because I had to let the Quad end just when I'm entering the exit area, for the final rocket-jump, which was discovered in a mistake by Nolan, who accidentally pressed his rocket-jump key in an early demo and jumped up there. I thought it'll be easier that way, since you get a clear shot at the switches, without monsters disturbing you. E2M5 - The Wizard's Manse (Yonatan, 1:07) I didn't do the rocket-jump in the start because I didn't think I'd have enough health. Then I did two jumps to the cage instead of one because it's very hard to survive that one, there is a shambler right behind you there. The cage ride is very boring and you must have more than 40% health, because you lose that while drowning. Romero, why did you do those extremely boring sequences?? [We love you really, John. (c: - Anthony] The last room can get violent, but those monsters don't stand a chance against me and my Quadded super nailgun :) The level is medium in difficulty. The route here is also Nolan's. E2M6 - The Dismal Oubliette (Yonatan, 1:29) Another very boring sequence - the descending lift. The rocket-jump in the start is very simple to think of, I guess I was the first to do it, plus the later return to the yellow armor and the second jump. The level is quite easy, but very boring to do. I let one shambler telefrag the other and be telefragged by me. The rocket-jump to the rune is Anthony's idea. [One could even try rocket-jumping out again as well - Anthony.] We know we could finish the level without taking the rune, but we don't think we really finish the episode if we don't take the rune. E3M1 - Termination Central (Yonatan, 0:48) The explosives jump is the idea of Anthony and Matthias, but is quite hard to do. I had to take the red armor, because otherwise I didn't manage to survive. But this way I got to the yellow key fast. You need to blow one of the explosive boxes before blowing the one using which you jump, because otherwise for a certain reason you don't take any damage. [It seems to be a bug in Quake. - Anthony] The level is hard. E3M2 - The Vaults Of Zin (Yonatan, 0:57) This level is quite easy, even without taking the ring of shadows. The whole route was done first by Matthias. I modified it a bit to have a faster time (different weapon usage, killing less, etc.). There are no specific tricky parts, you just hope everything goes fine. E3M3 - The Tomb Of Terror (Nolan, 0:34) This was about the only grenade-jump I could do, since you don't have to go up very high. I only have a 486-133 Mhz, so I only get about 10 FPS in Quake. And, in case you didn't know, the higher your FPS, the higher you can reliably go in a rocket/grenade-jump. [Because there can be less time between the jump and the explosion, so you're closer to the blast, so you get more of a boost - Anthony.] I got VERY lucky in this level, near the end, the monsters just stepped aside and let me pass... E3M4 - Satan's Dark Delight (Yonatan, 0:49) One of my favorite levels. The route was suggested by Anthony. It's also quite easy. I killed the ogre who guards the switch above the lake so he won't disturb me later, it's not a waste of time since I did that while swimming. The rocket-jumps are rather easy, the second one is optional but it's faster that way. E3M5 - The Wind Tunnels (Yonatan, 0:51) The brilliant grenade+rocket-jump was discovered by Nolan. The trigger was placed too low, obviously we can reach very high :) This is the only tricky part in the level, the rest is easy. I like watching both shamblers hitting me from the camera view. E3M6 - The Chambers Of Torment (Matthias, 0:33) The Ring of Shadows makes this level quite easy. Yonatan first played it, Anthony suggested the jump in the beginning. There is another possible route here where you don't have to get the gold key, but I think it's a few seconds slower than the one we used. E4M1 - The Sewage System (Matthias, 0:48) Another level where you just have to run through, no shortcuts here. Surviving is not easy and you need some luck too to dodge all those grunts and enforcers. Well, theoretically there is a shortcut: Anthony pointed out that one could use the explosive box to reach the gold keycard. But since there is no armor or megahealth in this level, there's no chance to survive it. E4M2 - The Tower Of Despair (Yonatan, 0:49) Quite an easy level. The route was done first by Nolan, who also was the one to suggest skipping the invulnerability and the first Quad. No tricky parts here, maybe jumping on the wood beams without falling. [Stop Press: we've now done this level a couple of seconds faster, just too late to make it into this demo. These guys just keep on improving! - Anthony] E4M3 - The Elder God Shrine (Yonatan, 1:21) The basic route was my idea, but Anthony and Nolan helped make it much faster. I was quite lucky in this demo, no monsters blocked my way back from the silver key area. Surviving is easy, but many things can go wrong. I killed the scrag in order to open a shortcut on the way back. I got rid of a pursuing spawn by throwing a grenade forward, which exploded right on time to blow it to bits. E4M4 - The Palace Of Hate (Matthias, 0:33) Again it was Anthony who discovered the route for this level. I used a triple grenade-jump sequence to finish the map quite fast. The jumps itself are not too difficult, but it's difficult to do them all in a row. Also you have to take care that you don't open the room with the fiends while getting the super nailgun. [Matthias has run the map a second faster from scratch by missing the super shotgun, but we wanted it for the rest of this episode - Anthony] E4M5 - Hell's Atrium (Nolan, 0:16) Not much to say about this level. Just run, turn some, jump some, and exit. E4M6 - The Pain Maze (Matthias, 0:26) Anthony and I discovered this route at about the same time. I realized that a normal grenade-jump is sufficient to reach the opening that leads to the exit. Yonatan made a 0:27 demo real soon, and it took me some time to improve it by a second. E4M7 - Azure Agony (Yonatan, 1:03) It is easy to get to the exit alive using this (Nolan's) route, but getting a good time is harder. It's hard to aim at the vores when you get hit from rockets all the time and can't see where you're firing. END - Shub Niggurath's Pit (Yonatan, 0:36) A bit of a boring end level, and pretty simple. I had to wait for the teleport destination (the spike thing in case you don't know) to enter Shub, so it can't be done much faster than that. Yonatan Donner Matthias Belz Nolan Pflug Anthony Bailey