Jaquaying the Quake Maps - Analysis of Quake 1 and 2 Maps (PART II)

(for those of you just joining us, see my Part I on Quake I)

Outer Hangar map

QUAKE 2 ANALYSIS

Quake 2 Preface

Before going on with the analysis, let's speak a few words. Quake 2 is generally considered an inferior FPS to Quake 1 in single player for multiple reasons. Some even consider its multiplayer weaker too. That being said, I'm taking a look only at the level design. My initial feeling will be that levels will be more complex because the engine offers more and they were more mature about it. Lastly, most of the levels are much shorter (I'm guessing because of processing/rendering concerns), so it is more difficult sometimes to grade them hence I was more generous. 

There are a lot of "hidden" (not hidden that much but still...) alternate paths, so keep in mind my Secret category also cover those. I'm saying this because the secrets in Quake 2 are much smaller in scope so people might be asking themselves why some maps have the Secret point if that is so. 

Also, I'd just like to point out 26 level design tips by Jaquay from an older interview. Not all of them are relevant to general dungeon design outside of video games, but it's still worth a read. 

1. Know what you want to do with a level before you start. Don't expect a map that you start as a single-player map to be easily changed into a multiplayer map. The reverse holds true for trying to make a Deathmatch map into a single-player challenge.
2. Sketch out a diagram of the map to use as an initial guide.
3. Don't start with grandiose projects. Try making something fun with a few rooms.
4. If possible, build your level with a "gimmick" in mind--some tricky gamism bit that players will remember. Popular gimmicks that have been used in the past include wind tunnels, numerous portals, lava maps, trap maps, water-filled maps, maps with large, slow-moving hazards, and low-gravity maps.
5. Try to be fresh and original with every new design. Do something that you haven't seen done before.
6. Test gimmicks of gameplay, tricks, and traps in test levels before building them into your game level.
7. Do architecture and texture studies ahead of time to establish an architectural style. Stick to that style.
8. Block out your level with large pieces of geometry. Think of the architecture you'll use, but concentrate more on how gameplay will flow through the level. At this stage, I try to keep my map grid at the largest possible setting (in Quake II or Quake III, that's the "64" grid). Avoid fussy details at this point and go for massiveness. At this stage of development, try to keep your frame-rate speeds well below the amount allowed by the game (for Quake II, we aimed to be below a maximum count of 500 triangles of architecture in any view). A good rule might be to try for no more than a third of your total possible polygon count in the worst views in and near your larger rooms.
9. Once the flow is established, you can start adding architectural detail and refining hall and room shapes.
10. Build in a modular manner. Make prefabricated pieces that be can fit together easily to make your level. Build tricky pieces of detailed architecture (such as door frames, complicated cornices, or furniture) once and set them outside the boundaries of your map. Clone them as needed for placement in the map.
11. When designing architectural elements, study the real world. Try to duplicate the look and feel of impressive works, but with less complicated geometry. Set yourself challenges in this regard.
12. Strike a balance between the use of real geometry and textures that imply three-dimensional depth when building architectural details. Textures that appear to be 3D should be used with caution. When viewed from a distance, they can fool the eye into believing that the architectural geometry is significantly more complex than it actually is. But the same texture viewed up close and at eye level completely destroys the illusion of depth.
13. Compile the map often. Don't wait until everything is placed to see what things look like (or if you have leaks in the map hull).
14. Complete your map geometry before adding monsters and items.
15. When building single-player game maps, don't put every game feature in the level. Having every monster possible in the game in a single game level is a glaring sign of amateur work. Generally speaking, the only place you're going to see all the monsters at once is in the AI programmer's test level.
16. The same goes for tricks, traps, items, weapons, and power-ups. Unless your map is as massive as the 64-player DM maps created for Quake II, restrict the number of different items you put in the map. Use a few things cleverly, rather than many poorly.
17. Small maps can be relatively similar throughout. Large maps should have distinctive, memorable locations that the player can use to orient himself in the map. "City64," a large DM map for Quake II, featured a huge canyon area, a massive alien temple, underwater caverns, a vast deep tank with water in the bottom, and numerous stretches of twisty corridors. The corridors were often similar, but they ended in distinctive large play areas.
18. For DM maps, give the players frequent opportunities to avoid pursuit and dodge for cover. Long hallways with no exits are bad. Avoid forcing players to make long trips to dead-end rooms--even to get good power-ups.
19. Place lights to achieve drama. If you have a choice between under-lighting an area and over-lighting it, err on the side of darkness. Just don't go overboard. Dark levels may look nifty, but stumbling around in the dark while playing gets old fast.
20. Light as you go--even if you're only placing temporary lights.
21. Don't forget the audio elements of a map. Sounds can provide important game clues.
22. If possible, allow multiple solutions for puzzles. You can still reserve the greatest rewards for players who solve them in what the designer has decided is the "best way."
23. Give the player a variety of game experiences and challenges in each map. All combat or all puzzles can get old quickly.
24. Be kind to your players; don't over-challenge them unnecessarily. Well-placed environmental hazards add to the tension of game play, but falling into lava or slime every third step or being crushed to death by falling weights every time you turn around quickly becomes frustrating.
25. Study maps you like and make an effort to duplicate or even improve situations and settings.
26. Finish what you begin.

Jaquay also speaks of other interesting things, such as: choosing the unconventional solution, blending flavors, making it real (understood as verscimilitude), etc. 

Overall

Serial	Episode Serial	Map Code	Map Name	Map Designer	DR	Loop	Secret	ES	Entry	Landmark	AT	EI	Intuit	SRA	PA	Total	Commentary
1	e1m1	Base 1	Outer Base	Tim Willits	1	1	1	0	1	0	1	1	1	0	4	11	This starts easy but high in opportunity. Right frm the get go, you can do three things: break glass and go through them, take normal way down which leads to an open gap in a wall or a door. The glass break can lead to potentially two ways to get different bonus equipment. You can check around to find ways to kill enemies from the sewer and so on. This map is small and takes but a few seconds to cross if you speedrun, but it offers a large choice of verticality and alternate paths. It is compact and fun. Grenades and explosive barrels can open passage? That's great. Very impressed with the first one and Willits doesn't dissapoint.
2	e1m2	Base 2	Installation	Tim Willits	1	1	0	1	0	1	0	0	1	1	4	10	Explosive bridge, some nice verticality in two different open room. Good use of contrats. I love how Quake 2 uses the "cramped corridor vs 3D open space with enemies above you". Again, Willits doesn't dissapoint. Also, since there is a second mission here after Comm Center, it's a very nice loop: you come back to the same level!
3	e1m3	Base 3	Comm Center	Tim Willits	1	1	1	1	0	1	0	1	1	1	3	11	Very nice verticality, some nice loop, secret area to open, use of a small "hub". Nothing exceptional but overall well done.
4	e1m4	Train	Lost Station	Jaquay	1	1	1	1	1	1	0	1	1	1	4	13	Secret area. Very compact but solid map: some good looping, small secret areas with loops and paths, a little train ride to change the orientation of the player, etc. The central room with the stone pillar is very nicely done where you circle around your objective and defend yourself. It is solid without being the most amazing map I've ever played. We'll see how the other level from Jaquay. Also for some reason my port shuts down the music when a new objective pops up. So when I picked the main feature of this level (the Super Shotgun), music stopped. It created a nice something as I went through the newly spawned enemies.
5	e2m1	Bunk 1	Ammo Depot	American McGee	1	1	1	1	1	1	1	1	1	1	5	15	If you've read my part I, you might know I am... skeptical at first, seeing the name of the designer. But oh wow. After a very early gotcha (of course, it's McGee), it rapidly evolved into the best map of the serie. It checks literally every single point: you have magnificient use of the 3D space with layers of bridges, a landmark room with said bridges, use of switch with direct impact (i.e. you see what you modified to the environment), use of water, modifying the landscape to access new area (exploding a door), plateforming, fast paths to alieviate previous difficulties (such as the small side-wall rampart of the water-flooded room), traps (that you can even use against your enemy such as lowering them into lava (of course it's lava, it's McGee) or dropping crates unto their heads for gibbling), and the level itself act as an hub between two other maps! And the level felt bigger than the other ones too. When I said that the 3D environment of the improved engine could lead to more mature mapping, that's a great example. The only downside is that I would have like the secrets to be more meaningful (they are mostly room secret with only power-ups and no change in paths except for the Chaingun one that offer a very small and kinda useless loop by bringing you back to the same place) and the level more "compact", less "sprawly".
6	e2m2	Ware 1	Supply Station	American McGee	1	1	0	0	0	1	1	1	1	1	4	11	A very good level. Less so than the other one, especially with regards to secrets and elevation. It barely passes the Doom Rule (I only remember/saw one room that was on top of another, the rest of the elevation was Doom-like). The looping and layout are awesome. The whole level is a nested loop that brings you back to the beginnign (and therefore to the primary hub of this episode). The environment is well used with crates moving around (even if they don't make sense), a little tramway to crash, electric rails to disengage before moving, etc. I'm only giving the landmark point because of the double nested loop (lhe level itself brings you back to the entrance that brings you back to the central hub of Ammo Depot)
7	e2m3	Ware 2	Warehouse	American McGee	1	0	0	1	0	0	0	1	1	1	2	7	Unless I missed something, there was no real loop (there is one bridge but it's useless as you don't have to go back down), which is kind of rare. The level felt much more linear. Good use of interactivity, lifts and verticality, but overall weaker.
8	e3m1	Jail 1	Main Gate	Tim Willits	1	1	1	1	1	0	1	1	1	1	4	13	The level design integrated many elements and had a superb loop through a secrets that make you skipped a portion of the level. I don't believe until now in Quake we've had a level so dramatically presented: two absolutely different path offered from the get go. There was a nice detour underwater and various interaction with the environment through the (damn) laser trap and alarms. Even a big canon at the end controlled by an enemy was a nice touch. Loop work was too minor to include since it's a 15 pace thing with glasses and all. My only grip with the alternate path is that it doesn't exactly tell you it's alternate: in other words, if it's your first playthrough, you might end up backtracking, which beat the purpose. This could have been sold by a window or something that show you've made progress. But more on this later.
9	e3m2	Jail 2	Detention Center	Tim Willits	1	1	0	0	1	1	0	0	1	1	4	10	Minor alternate path at the begining, either up or down to get the blue keycard. A nice touch again. Overall good but simpler level with less use of elevation (compared to the Quake 2 standard that is, especially from Willits). Fun map nonetheless.It also serves as a kind of hub to get back to with a key to open a door, which is a very nice use of interconnectiveness.
10	e3m3	Jail 3	Security Complex	Tim Willits	1	1	0	1	1	1	1	0	1	1	4	12	Also a mini-hub to come back to with a distinctive landmark: the pyramid. This create very nice foreboding and you can anticipate a return here and something "out of the ordinary" taking place.
11	e3m4	Jail 4	Torture Chambers	Christian Antkow (Xian)	1	1	1	1	1	0	1	1	0	1	4	12	One of the new level designer. The map carries very well it's theme with mutiple torture chambers, prisons, guards, etc. In terms of level design, there are multiple paths and branching to reach different location (i.e. the keys of the level, the data CD). Lots of bridge and corridors at different vertical level. Destructible environment in certain areas. Overall well designed if a little mazy.
12	e3m5	Jail 5	Guard House	Tim Willits	1	1	0	1	1	0	0	0	1	0	3	8	Very short level mainly for a specific key. Not much to say: even if it's very short, there are still some minor branching and verticality to provide for different approach.
13	e3m6	Security	Grid Control	Tim Willits	Boss area												Basically a boss arena. Some traps prior to it. Nothing to say really.
14	e4m1	Mintro	Mine Entrance	Tim Willits	1	1	0	1	0	1	1	0	1	0	3	9	Act as a hub to return here later to get access to the Factory. Not the strongest level, but still contains some alternate branching, some elevation stuff, etc. But it's weaker in general.
15	e4m2	Mine 1	Upper Mines	Tim Willits	1	1	0	1	1	0	1	1	1	0	4	11	Loops, branching path, elevation, elevators, bridges, water, crouching, etc. This level makes full use of 3D environment in a very satisfying way. Even the layout partially changes throughout the level as you activate a propeller, open up a bridge, etc.
16	e4m3	Mine 2	Borehole	Tim Willits	1	1	1	1	0	0	1	1	1	0	4	11	Very nice layout level with good contrast of small and big spaces and almost two different level (one above, one below) that loops and branches up. It also act as a small hub after you activate the machinery.The secret in this level even give you access to an alternate entrance into another level.
17	e4m4	Mine 3	Drilling Area	Tim Willits	1	1	0	0	1	1	0	0	1	0	2	7	Short level with some traps and minimal interaction. It doesn't have much to offer except that it's kind of two level in one. It's not a bad level, it's just very short and made to transition to lower mines.
18	e4m5	Mine 4	Lower Mines	Tim Willits	1	1	0	0	0	0	1	1	1	0	2	7	Lots of environmental stuff going on (lava, platform, traps, etc.). Layout looks intricate because of the 3D aspect but it's really simple in reality. It's rather short with few things going on, very "corridor-linear". It has some traps that can be used to advantage and some (one room) elevation going on, but the rest is very dry. A transitional area (which is sad because it's like 2 transitional area in a row).
19	e5m1	Fact 1	Receiving Center	American McGee	1	1	1	1	1	0	1	1	1	0	4	12	Act as a small hub. Of course it has lava. As is now McGee routine, it's awesome: mutliple branching paths, multiple objectives, envirnment interaction, elevation (a little bit less than is usual), etc. It has a lot of very awesome point (and even a stupid secret area).
20	e5m2	Fact 2	Sudden Death	American McGee	Special secret level												Said secret level. Not a real level. You have 30 seconds to grab as many items as possible including a railgun (which is already in Receiving Center in a secret anyway).
21	e5m3	Fact 3	Processing Plant	American McGee	1	1	0	1	1	0	1	1	1	0	4	11	Again a level with loops and multiple objective. This one feels more dynamic than the other because of a more "complex" objective, but it's less complex in layout and more linear.
22	e6m1	Power 1	Power Plant	Tim Willits	1	1	1	1	1	1	1	1	1	1	4	14	Very tight level with multiple branching, a secret area that's worthwhile (act as a branching path AND gives the BFG). Elevation, bridges, rampart, ladder, complex layout. It checks all boxes in a very tight way. It's an amazing level that feels dynamic even if compact and short. It has all the ingredient. Not giving it a full 5/5 for personal appreciation because I feel the shortness of it gives less weight to the different positive pointsa and "check in the box" since, for example, you can loop only so far.
23	e6m2	Power 2	The Reactor	Tim Willits	1	1	0	1	1	0	1	0	1	0	4	10	Solid level with a heavy dose of elevation shift, multiple entry into rooms and good overall layout. It even have some nice details, like showing you the maps of the episode. Serves as a hub for Cooling Facility and Toxic Waste Dump.
24	e6m3	Cool 1	Cooling Facility	Tim Willits	1	1	0	0	1	0	1	1	1	0	3	9	The level is the "sewer level" of the game. You follow corridors of waters. It has some nice thing going on in specific portions, but it wasn't the best of Willits. Elevation and contrasts in space was good.
25	e6m4	Waste 1	Toxic Waste Dump	Tim Willits	1	1	1	1	0	0	1	1	1	0	4	11	Contrary to the name it's not a sewer level. Lots of elevation, nice shortcut opening, loops, brahcing paths and stuff to do. Overall a great level.
26	e6m5	Waste 2	Pumping Station 1	Tim Willits	1	1	1	1	0	0	0	1	1	0	3	9	Both Pumping Station are simple but interesting level with some loops. They are kind of short so I'm putting them together.
27	e6m6	Waste 3	Pumping Station 2	Tim Willits	1	1	0	1	0	0	0	1	1	0	3	8
28	e7m1	biggun	Big Gun	Nil	Boss area												It's a boss area. At least it has a loop and a countdown?
29	e8m1	Hangar 1	Outer Hangar	Tim Willits	1	1	1	1	1	0	1	1	1	0	4	12	Good level with a secret level hidden. Some looping, branching path, elevation and different interactions. A diverse and compact level.
30	e8m2	Space	Comm Satellite	Christian Antkow (Xian)	1	1	0	0	0	0	0	1	1	0	1	5	Vertigo part 2. However it's stricly worse. In Quake 1, Vertigo gave you the ability to move around at low gravity, resulting in interesting 3D exploration as you could almost "swim" through the space. In Quake 2 it's 90% corridors, and it's inundated with the worse enemies of the game.
31	e8m3	Lab	Research Lab	Brandon James	1	1	0	1	0	0	1	1	0	0	3	8	Not the worst level but definitively not the best. It has some nice theme and looping, but it has some weird enemy traps/placement and a confusing layout.
32	e8m4	Hangar 2	Inner Hangar	Tim Willits	1	1	0	1	1	1	1	1	1	1	5	14	There are a lot of loopings, branchings, opening shortcuts, environmental effect (platforming, elevators, etc.) and other things. It's a very strong level from Willits that takes most of his best skills as a designer and cram into one map. The only downside as you can see is that there are no secret alternate path or worthwhile secrets.
33	e8m5	Command	Launch Command	Christian Antkow (Xian)	1	1	0	1	0	1	0	0	1	0	3	8	Simpler level with a hub-like structure and multiple objectives at different points (i.e. branching). It has some loops and elevation shifts, but nothing really that stands out. It would probably have been a better earlier level since it has very few serious interaction and layout. Even if the score is low, it's not a bad level per se.
34	e8m6	Strike	Outlands	Christian Antkow (Xian)	1	0	0	0	0	0	0	0	1	0	2	4	This level felt weird and... unprofessional ? It's strange to say. The mission is short, easy and not very exciting. You have just one thing going on and it's a branching path. It doesn't even have the Doom Rule (which is the first for Quake 2 and one of the only for Quake in general). The only level I had some collision bug with the level itself. It felt unpolished and just tossed together. Even the ennemies were all early level. There was even a confusion: there were cracks on the ground that all over the rest of the game were destroyable to reveal secrets: well now in this level! It should have been an early level, and more interesting. Easily the worst level of the serie in terms of "blandness".
35	e9m1	City 1	Outer Courts	Jaquay	1	1	1	1	1	1	1	1	1	1	4	14	Here we get in the Jaquay string of levels, the endgame. Visually I have to say that it's legit they are the best levels. I don't know if it's because of the personal talent/attention to details of Jaquay or because they were made later with more experience by the team, but they are so much richer in details and way ahead in terms of aesthetics by use of real architecture, tall buildings, symbols, and even carpets! Level design wise, this map has nice looping, landmark, traps (can't say they are nice, but they are there), interactivity, cramped area contrasted with open areas, etc. The map got it all. The weakest aspect is the secret since it has a looping but not much more. The main door at the beginning act as a landmark as you have to find the way to get it through a switch. I love that technique: forebode where to go and put the opening mechanism elsewhere while still seeing what it actually opens. Romero used that in Quake 1 too. The biggest downside of this level (I'm not rating it per se) is the damn infinite spawning flying enemies.
36	e9m2	City 2	Lower Palace	Jaquay	1	1	0	1	1	1	1	1	0	1	4	12	Very solid level again, although unintuitive. It has less elevation shifts and contrasts, but it still has some thing going on on that front. The "throne room" with the computers is shoved with details and very nice.
37	e9m3	City 3	Upper Palace	Jaquay	1	1	0	1	1	1	1	1	1	1	4	13	Very good level with branching paths, loops, bridges, elevations, traps, etc. The whole package (mostly) again.
38	e10m1	Boss 1	Inner Chamber	Brandon James	Boss area												Boss area with a very tiny (and boring) mini level prior to it.
39	e10m2	Boss 2	Final Showdown	Tim Willits

As you can see, secrets are much tamer in terms of what they open (less secret level, less alternate path). It's generally a room with ammo/weapon, and it's generally lame since you get those all the time (and newer weapons aren't new for a long time). For example, the Super Shotgun level is very nice, but it takes the same time to do the level to get the Super Shotgun than to just continue the campaign and get it. My guess is that this design decision was made by someone else than the level designers themselves since it seems to have been really widespread, just like in Quake 1. This is logical from the perspective of a 3D engine: it takes more time to create secret areas in 3D with all assets included than in Doom, for example. 

Per Designer

Let me first say that I wish Romero would have done maps for Quake 2 as his map in Quake 1 were solid and I know he would have made great use of the engine. 

Antkow and James did the worst level of the bunch. Their level are unmemorable, generally boring, and sometimes broken. I've got nothing really to add or say except play them yourselves to get your own opinion. But even outside of opinions, as you can see their level integrate very few interesting features.

Willits again provided a solid stream of works. I dare say that some of his work in here are more wide across the spectrum: some are better than his best in Quake 1, but some are also worst than his worst in Quake 1 (although the Wind Tunnel is difficult to beat). Yet my favorable comments regarding Quake 1 also applies here: he made great use of the new capability of the 3D environment and the engine and it's a great result.

McGee is the biggest surprise as I've said in my comments and his level are among the best of the whole bunch. Top tier with Jaquay. I can't say much more since he only did a specific portion of the game levels.

You can see the difference in Jaquay level design, it's really interesting to see it shine so much: there are more traps, more mazes, more more application of the different level design technique (branching, loops, hub to come back to, etc.) than the other designers. There are also more attention to visual details and the level are richer for that. I'll be honest I really thought that the level would be good, but that the difference wouldn't be "that" big (especially after McGee): I was pleasantly surprised. Although to be fair, I don't necessarily consider it world apart from the others: Romero level in Quake 1, McGee level in Quake 2 and the overall quality of Willits are very much in the same league (although visually less intricate, but like I said, I can't say for sure if it's just Jaquay or if its the 3D texture maturity of the endgame having been developped). The last three level of Jaquay really felt like an interconnected whole, a "dungeon" so to speak. It is almost more akin to Doom than it is to Quake: normal Quake levels uses 3D to provide a more dynamic and fast pace action. Here Jaquay uses the 3D space to create more complex and intricate environment that feels more like dungeoncrawling than pure unfiltered action. Although none of them were 15, you can see that it's mostly due to my personal appreciation and some issues with the pacing. But the maps in themselves, emptied of everything, are an absolute beauty and are by far the best in the series so far.

[EDIT: I added the following block later. It came to me that this should have been there initially as it brings everything more together with examples.}

DETAILED ANALYSIS

I will now discuss in detail my two favorite maps, once for each game: e2m1 (The Installation) for Quake 1 and e2m1 (Ammo Depot) for Quake 2.

Quake 1 - e2m1



e2m1 map blueprint. Rights to the owner and graciously found on the quakewiki from http://www.benryves.com/

• Generalities: The map is not complex, and could even be said to be "simple". But it is simple in the good sense. It is short, compact and offer a very dynamic approach (i.e. you can rapidly navigate it and take head on the challenge without any space/time loss). I feel the map represent the best of a "compact" approach to Quake mapping. 
• Layout: As you can see, the layout is interesting in the sense that it's made into a square. In reality, the layout is more of a inverse U shape since there are no connections between the left/right section in the south. It's a very good level that's intuitive even with path closing behind you, loopings/branching, keycards, etc. 
• Looping, branching and alternate paths: Right from the get go, you are offered two different paths to take: a corridor left that's enclosed, or an open room with a bridge above water on the right. What's even better is that said water actually brings you to yet another possible path for a secret. Then right after that, you get to a passage that separates again left and right. And that same room has a looping secrets passage with a teleporter back to the beginning. And that's like 30 seconds in. This gives you an idea with regards to branching and looping in this level. I suggest looking at a playthrough 100% on Youtube (or better yet, play it yourself) to get a better idea. It really feels like a full fledged great Doom 3D level with all the compact and complexity added. But yet it's always intuitive. 
• Secrets and exploration: There are a lot of secrets. Some of them leads to teleportations, alternate rooms, and so on. It really encourage exploration, especially since most of them are not obtuse: going through the water is just a matter of exploring, not of hitting a random pixel with your gun. 
• Use of 3D space: Water, elevators, above/below enemies, etc. This Quake 1 level really is interesting. It is kind of simple in its 3D, with generally only very clear elevation shifts and so on. But still, it's present and well done. 
• Use of contrasts in time, space and visual: Like always, Romero does well the "corridor leads into an open room that leads into an enclosed space" loop. There are also more light variations. Also, it's interesting to consider the use of closing wall to create that sense of contrasts where the only way to go is forward, trapping you even more into the enclosed space (even if they actually reopen a bit later). 
• Environment: Apart from the walls, use of 1-2 traps and water, there isn't much. Some enemies are hidden behind walls, as always, but it's fairly light on this, even by Quake 1 standard. 

Quake 2 - Ammo Depot

I couldn't find any blueprint or map drawing, sorry! Here is my own general reconstruction of the map in a pointcrawl format. I am sorry for the low (paint) quality. Just to be clear, most of the big rooms with an adjacent elevated/underground room means they are all stacked on each other in one giant room with multi-level. You could probably just consider it one giant room, but since they sometimes serve different purpose and are on different height I separated them. 

• Generalities: A nice level that makes great use of the 3D space (multiheight layered) and looping/branching that, even though the layout is actually "simple", it feels like it's complex with clever use of change of height, switches and environment interaction. 
• Layout: The drawing I made didn't really give justice to the layout. In reality, there are a lot of overlapping bridges and passages, sometimes with up to four level of height in the same room. The layout is much more sprawling and distant, but because of the branching and looping, it feels all interconnected with very few backtracking. It also doesn't show justice to the alternance between elevators, slopes and passages to switch between rooms and heights. Finally, it also doesn't show well where you get shot from above by enemies, and then you finally reach that spot later on, creating a very nice forebodding sense "oh, I CAN get there!". 
• Looping, branching and alternate paths: As yiou can see, there are some alternate paths, some rooms with closed door that require looking for solutions, some one way passages, some looping, etc. The only downside is that in such an FPS, looping isn't "that" useful because you have no real reason to come back. That being said, since we take a look at it from a dungeon design perspective, it's a very nice touch. Players going through this particular level would have been more than happy to find the "hidden" ladder that brings them back to the 4th room at the bottom. 
• Secrets and exploration: There are quite a few secrets and hidden stuff, some of them leading to looping and such. Some require a leap of fatih (as per the one way passages), which is interesting. Nothing here is out of the ordinary with regards to that. 
• Use of 3D space: Very well used: elevators, slopes, ammo crates, enemies with the higher or lower ground, nested rooms, etc. For level design, this serves multiple purpose. It gives a sense of dynamic exploration: you can travel three different axis. It also gives weight to (perceived) complexity: if you go up and down, it means you are now above other rooms you have to explore and vice versa. Finally, its used for encounter design with enemies below/above. 
• Use of contrasts in time, space and visual: Lots of small passages open up to huge room and vice-versa. You have also contrats between the high and low: the lower you go, generally the more the space is confined (crates in big rooms, or small corridors). Whereas the higher you go, the more open it is (and even dangerously so since the high bridges/passages have no protection from falling, and enemies have kick-back attack!). 
• Environment: Lava traps, enemy traps, traps against enemies, switches, elevators, etc. This level has a lot of them, even if they are very simple in design and scope. It's almost a "showcase" of sort. It's much more refined than the showcase we shaw in McGee level from Quake 1. 

Another good coverage that could be made would be to cover all Jaquay end game levels from Quake 2 as a single unified dungeon and take a look at it.

[EDIT: End of the added block]

CONCLUSION

We come to the conclusion of this small series... or did we? I've had much pleasure in doing this serie and I will be pursuing it in two other direction now:

• First, I want to assess a mod/wad of Quake 1 that's been said is an absolute marvel in terms of level design, "Arcane Dimension". I will try to play it when I get the time, take note and write an article on it if it's worth it (i.e. it if brings something different or interesting enough to comment). I will take time to also review my criterias. For example, I feel branching should be taken into consideration outside of the other categories. Encounter emplacement and the type of objective should be counted also (because I'm looking at Hexen with the switch hunting and I need to be able to take points off of that). Also, I should be clearer (and simpler) in the definition of some criteria.
• Secondly (although it will probably be published first), I will make a recap of everything I've learned about level design in this little serie and condense it into a digestible article that can be used by everyone. I will post different technique, example and application within the context of D&D/RPG maps. 

It would be possible to expand this serie into an actual "game level design for D&D" that include classic like Dark Souls and such. I might do this... one day. 

Let the Jaquaying continue!