Developing UV-Max [and other] Routes

Doom Speedrunning has been sitting in the back of my head for awhile now ever since I went on hiatus because of reasons that don't need to be mentioned here, and I want to ask what methods do you use to learn how to break down maps and create a streamlined route?

How many hours did you spend on certain maps?

Curious too, I've never recorded a serious uv-max.

I don't think theres a specific order of priorities that applies to every map, but usually its make a dash for the "main weapon" first. then closest secret with valuable contents like armor or soulspheres, killing any high danger sniping monsters in the process, or those inhabiting a linear path. Let groups of mixed species monsters live to provoke infighting, target isolated groups of like species. Utilize chaingun guys, mancubi, revenants, arachnatrons, spider demons and cyberdemons to do your dirty work for you. Pain elementals and archviles need to die asap (though technically, archies can be abused to repeat resurrect zombies and imps for easy +kill%; not sure a pe's lost soul swarm does enough damage to help), and you gotta make the judgment call on the rest of the species if their faster killed by you or with infighting.

40oz said:

...though technically, archies can be abused to repeat resurrect zombies and imps for easy +kill%

FYI every thing (except lost souls) must be dead when you exit the map in uv-max. It's not the % counter that matters.

From DSDA: "100% Kills and Secrets, exiting with all monsters dead (except Lost Souls)."

You'll probably find this thread of interest (especially Ryback's post).

Clarification of Max definition here.

I'll use this thread to complain a little. I don't like how quite many maps nowadays have this as the most efficient uv-max strategy: run through the whole map activating as much triggers and monsters as possible and picking up the weapons and ammo, which takes like 1/10 of the time, and then the rest 9/10 you just clean up the whole place.

Darkwave32.wad is so entertaining to play "as intended", with several very tricky fights where you need to pay attention to the monster movements all the time. But in uv-max you just do two arch-vile jumps to awake the monsters in the whole map, and the rest is just a boring clean up, as I said.

Seems like quite a few maps in Speed of Doom and recent Resurgence are like that too: they just become a lot more boring in uv-max. I almost want to start recording demos with subpar routes because of this, but wait, I haven't recorded anything in forever anyway.

I do minutes rather than hours. Max I spent on demos is 30-1 hour, since I want the rest of the day to spend. I normally think of something whilst I playthrough the map of a route before I attempt a max in any other occasion I think about a map, which normally would be in a couple of hours/days after playing.

It's pretty tricky and even the most experienced players will usually come up with suboptimal routes if they're the first to record for a virgin map, especially when non-linearity creeps in. Just yesterday I watched vdgg's first attempt at routing btsx e2 map15 and I've enjoyed it tremendously, but knowing the map better than most at this moment, I can only say it's just the start of the quest for a good route, heh.

I love non-linear maps and I love coming up with intricate optimizing strategies, but I never let myself think that this is it, the perfect strat and now it's just a matter of grinding the perfect execution. Last time I thought that, jario destroyed my polished, smooth sc17 (a tiny, tiny map!) max with a sloppy 20min effort run using an entirely different strategy. Some of my best routing work is in my dbimpact maxes, but that's what I thought originally as well and then I returned a year later and smashed my own times by entire minutes with more proper planning and riskier strats. I can already see ways of saving minutes in my aeternum map02 max too and let's not even start with my rjspace9f run or I'll rant for days, heh.

I guess my point is it's hard, so don't expect to get it right the first time, sometimes you need a different person more removed from the problem to point out what you cannot see anymore, heh. Out of the box and whatnot.

It's probably different for conceptual slaughter maps where you're locked in massive arena battles and the question is "who to kill first and who to let infight", but that's not my cup of tea, so I'll let others talk about that.

Memfis said:

I'll use this thread to complain a little. I don't like how quite many maps nowadays have this as the most efficient uv-max strategy: run through the whole map activating as much triggers and monsters as possible and picking up the weapons and ammo, which takes like 1/10 of the time, and then the rest 9/10 you just clean up the whole place.

This is my pet peeve with slaughters and horde maps in general. All those fat, strong monsters on ledges work in the initial chaotic stage of the battle, but once that subsides, it becomes a menial janitorial task to kill everything off. You singled out Joshy and Darkwave, but I see it everywhere. ArmouredBlood, Phml, Alm, Gusta, skillsaw, Huy Pham... you name it, you've got it.

I've recently tried to come up with effective ways of limiting this problem, but it's not always easy. There are the cheesy telefrag/crusher chain secrets Huy Pham liked to use so much, but lately I'm inclined towards letting the snipers into the fray after a certain action. Like, the afrit valley in res29 or the BFG viles in Fireking, releasing them from the sniper nests actually makes for a good secondary fight while killing them off one by one/from the ground was the antithesis of fun. But if that's your thing, you can still knock yourself off before the trigger.

Oops, a tangent happened. :P

You pick whatever you think is the best route and hope it works. That's about it.

dew said:

I love non-linear maps and I love coming up with intricate optimizing strategies, but I never let myself think that this is it, the perfect strat and now it's just a matter of grinding the perfect execution. Last time I thought that, jario destroyed my polished, smooth sc17 (a tiny, tiny map!) max with a sloppy 20min effort run using an entirely different strategy.

Yup, that can happen. Sometimes I spend way too much time on a single map (up to a couple of weeks on and off), or I spend whole evenings on getting exits using that specific route, and my biggest fear is that somebody will come along and say, "hey, I made a run for this map earlier, but I forgot to upload it, but you reminded me, here it is, btw it's 2 minutes faster" *cough* 4shock *cough*

On strategies: I think if you really want to, you can optimize on the finest detail level, but it comes down to what is practically possible. I generally don't do that (with exceptions: for example, it was fun to figure out which combination of weapons or routes was faster in some areas of my runs, for example by doing each combination five times and writing down the time it took, and then taking the average).

For example, I tend to play by ear when it comes to ammo and weapon use. This is probably nothing new, but I gradually learn if, for example, I will need a couple of rockets near the end to quickly kill an archvile, and use all of them except for about four or so (for example in my last Jenesis run). SSG is generally used more, so if I can use the chaingun without losing time, I do so (obviously ignoring the cases where the chaingun is actually better, such as picking off scattered zombies).

Again, probably nothing new: For routes, I usually learn the nomo run first, and then adapt where necessary as required by secrets, monters, ammo and (necessary) items. I try to maximize (1) staying as close as possible to the original route, (2) monster infighting (3) weapon damage, and I try to shoot as often as possible.

Of course the usual bag of tricks apply, such as pressing buttons or opening doors while reloading the SSG in the middle of a fight, etc.

Of course, sometimes there are spanners in the works. For example, in Unholy Realms Map 5, I think it was j4rio (again) discovered that one of the doors (blue I think?) could be opened from the inside by one of the monsters. I'm not an advanced enough player to pick up hacks like that (that's why I don't do (UV-)Speed)) but at the same time it's educational and amusing if it does happen. :)

ahaha looks like memfis is thinking the very same as me about dooming...
same keyboard setup, same gameplaying, same hatred for some kind of maps and the maxing levels going straight to the end and then backwards.

I find it not the way the mappers intended it. But yea its' efficient.

blob1024 said:

ahaha looks like memfis is thinking the very same as me about dooming...
same keyboard setup, same gameplaying, same hatred for some kind of maps and the maxing levels going straight to the end and then backwards.

I find it not the way the mappers intended it. But yea its' efficient.

Interesting.

I'm not much of a speedrunner as I'm too intimidated to compete with the records of most maps greater than the MAP01's of most megawads. My most recent attempt at UV-Maxing MAP01 of HR2Final yielded 2:02 as my fastest time, which is over a full minute longer than okuplok and kimoxvirus' records.

But I'm curious what the list of general complaints for speed and max runners, as well as regular demo watchers looks like. In general I'm usually far too impressed with the skills of the player to really judge the quality of the map in relation to it, but I'm sure a regular watcher might be looking at it with a different perspective and the runners could probably write essays about a single monster's placement :)

I've been studying the Doomed Speed Demo archives in brief increments of time trying to find relationships between the most popular wads to record demos for. I've been contemplating designing maps specifically catered to both the players and spectators of this forum. At the moment I've only really made a mental note of high difficulty traps, and perhaps secrets that are carefully and structurally designed to require thing wobbles, elastic collisions, or glides to get in normally unreachable areas, or at least without a considerable addition to your time.

That's an interesting thought, but note that for some of the speedrunners that like to devise map-breaking tricks, part of the joy is that you are indeed breaking the intended sequence. If a trick-laden route is actually the intended sequence, then it's a little different.

So if you're going that way, you might need to make the trick-route truly spectacular; simply leaving deliberate scope for 32-unit glides or strafe-50 jumps (etc.) might not make them appealing to speedrunners. (At least have them have to work to make them happen rather than spoon-feeding them.) And it would be an idea to make the maps sufficently open-form that there may be other ways to use the possible trick ideas. Then speedrunners can try many different routes, and have the possibility to find unintended ways to short-circuit the intended route.

I suppose what I'm saying is that being on a monorial is not intellectually challenging, however interesting a route in itself the monorial is taking.

I've only been attempting UV-Max since April of this year, so I'm far from an expert, but I have to say routing is the hardest part for me. Usually I do what j4rio said and just pick something and run it. Then, as I do attempts I'll organically discover certain things are faster. Like, "oh these cacos usually end up at the exit if I miss them, maybe I should just wait and kill them later" or "hey I can telefrag this mancubus if I skip these chaingunners" etc. etc. These kind of discoveries are really fun, but require the kind of map knowledge and gameplay knowledge you only get from playing the map numerous times.

The more demos I make, the more I'm realizing that blob1024 and memfis are right. Usually the fastest route involves beelining for the power weapons, progressing to the exit, and then killing everything. I'm still fresh on the scene so I haven't quite decided that this type of route is uninteresting. However, most UV-max routes don't end up making use of cool UV-Speed tricks like keybumps or rocket jumps. Granted, this isn't always the case, but it has been in my experience.

Getting back to the original question, the routing and executing the route kind of develop in tandem. So while there is a "routing" phase, routing the map still continues as I iterate on one specific route. So by the time I get a satisfying exit, I might be using an entirely different route than when I started. As far as how long it takes, it really depends on how low you want your time to be. There are a ton of unoptimized UV-Max demos on DSDA. If just want to beat a run, then it might take more or less time depending on how well optimized the demo was in the first place. Sometimes your first exit will beat the DSDA time. What then? Why bother optimizing even more? I dunno, that's a personal decision.

UV-Max is a very interesting category to route for. You have to worry more about ammo usage and infighting than you might with UV-Speed. This lends UV-Max to having much more competition. Outside of highly optimized runs there are usually a few things that can be improved. I.e. tatsurd-cacocaco improving some Compet-N times for Requiem maxes. Those maps have been played to death, and yet are still improvable. This is what intrigues me most about UV-Max.

This post probably doesn't make any sense as it is (in my opinion) rambling and self-indulgent. Hopefully it is intelligible enough to be of worth to someone.

^ Far from it, I enjoyed reading it quite a bit!

Although I find it more discouraging, perhaps because I tend to make my UV-Max attempts blind by comparing the times of my personal bests to the records on DSDA than to have watched a record setting demo and developing a more efficient way to emulate what happened.

I guess that's an important part of the question. Does studying existing demos thoroughly have a lot to do with it? Perhaps looking at the demo with IDDT IDDT on and searching for monsters that are better killed earlier than later that the demo-recorder might not have noticed in gameplay is a vital part in planning a better route.

It's likely a personal preference rather than the definitive Right Way, but I very much like having a demo to base my route on. Usually it jumpstarts the organic routing thing I mentioned. Jumping into a virgin map that no one has routed for Max before is very intimidating to me. Some runners may enjoy that approach, but I usually prefer to have something to base my initial attempts off of, and develop the route from there.

Try not to get discouraged though. UV Max is a very hotly contested category for demo recording. Spending a week routing something and getting a satisfying exit only to have it crushed by a stellar doomer comes with the territory. Get inspired, not discouraged.

Maxes are interesting in that you can keep improving them just about eternally. Even if you end up getting limited by route improvements, there is almost always room for performance and overall luck improvements. The shortest maxes may be the least probable to improve if they are done well enough (~30-40 seconds per demo), but anything over 2 minutes is always improvable unless there are timed triggers present. Check times of maxes here for example, note that most of those demos have little to none routing changes and times usually boiled down to petty details.

Indeed, nowadays I too favor going for "blind" attempts since I feel it's much more satisfying to come up with your own routes and strategies. Not that I strategize my runs much beforehand; it's a few playthroughs to familiarize myself with the map and then I start recording. I let the route evolve during play as I pick up things I could do differently. Only after a few decent exits or so I check out the pre-existing demo to see what he or she has done different. If there are things done considerably better, then I adjust my own strategy accordingly and continue recording, if I feel like it, to produce a higher quality demo.

To me it feels that if I watch an existing demo before recording myself, it kind of skews my views and I may miss potential better strategies because of having this basis route predefined. "This other guy used this route so it must be good, right?" Minor adjustments are easy to do sure, but changes that deviate from the original route considerably can be harder to see when you haven't formed an independent opinion of a good route yourself beforehand. Hard to explain properly like this but maybe you catch my drift. Much like when recording for a "virgin" map it's impossible to instantly come up with the perfect route and even the final product has flaws that you can't see but someone else may pick up. Two independently recorded demos by two different players give you an opportunity to have this "dialogue" with the other demo where the other runner shares his "opinion" in demo form and you have yours. Then combine the best of both worlds for a better quality demo.

All of this depends on the map in question obviously. If it's a linear pipe shooter then there's probably only minor adjustments to be made and basing your demo on a pre-existing route doesn't hurt much. For stuff like Can't Run From Evil you need to have even more than two players providing demos for a truly perfected route and strategy. It's exactly like j4 said though, maxes can never be optimal and the longer the map the more there is room for dependency on player performance.

Veinen said:

Only after a few decent exits or so I check out the pre-existing demo to see what he or she has done different. If there are things done considerably better, then I adjust my own strategy accordingly and continue recording, if I feel like it, to produce a higher quality demo.

To me it feels that if I watch an existing demo before recording myself, it kind of skews my views and I may miss potential better strategies because of having this basis route predefined.

Also, for me it's easier to learn the map if I figure out my own routes first. One of my first demos was Doom 2 Reloaded map 20, a pretty big map with some choice in the first half on where to go next. I learned the map from the existing demo but I felt that it took way too long. It's like using your GPS to drive everywhere - it just takes longer to familiarize yourself with the surroundings because you're looking at that damn map all the time.