What makes a good Max?

We haven't had many discussion threads so far in this forum, but I think this should prove a useful topic.

• How should a Max be planned?
  
• Are there any techniques and ideas that are unique to the Max category?
  
• How best to reduce the risk of missing monsters?
  
• Once you've got a first exit, what should you look for to try to improve it (apart from the obvious things like missed shots, poor running, etc.)?

These are questions I've been grappling with recently, and I'm interested to hear what other people think.

(Nice idea with this thread)

I'm not a player, just a watcher. I admit I really appreciate outrageously fast max demos. So, I guess, once a player has been able to trace a route for a max, he should look for shortcuts and record and record again until missed shots are a souvenir ;)

I take the opportunity to congrat all the demo players around here. You guys are making a tremendous job. Your work is really appreciated.

EDIT:

I'll be a little more precise for the shotcut idea. I see finding the fastest route kinda like optimizing a program. Sometimes, it's not about making every pieces faster but rather find another organisation of the pieces that makes the whole thing faster.
Another thing I have spotted in quite a lot of demos is how some players are stuck for a second or two (no biggy) because of an unexpected monster position. I have the feeling it's good to always have 2 or 3 alternative roads in mind when entering a room... in order to adapt to the configuration.
Maybe that doesn't make sense, but this is how I see it.

I think most players recording today understand or naturally do all of these things.
If it is an easy map, I'll just jump in and IDDQD to the exit for a first look. On many maps that means I will get 100% kills, but miss some secrets. Then it's to Deepsea and see what I missed(since the map will be familiar now, I'll know what to look for or where I didn't go).
If it is a harder map, then I look at the editor first, then IDFA and IDDQD. then editor, then just IDDQD or a real recording.
Once I have a real recording under my belt, then a second attempt will undoubtedly be faster.
DOOM is so dynamic that different weapons and monsters will result in a much different experience. (imagine any HR map without rockets, or without cells, it would be quite different)
It is very hard to optimize a max recording for a well-designed, well-balanced map. Look at the Compet-N records for E1M7, for example.
One of the finest demos ever recorded is ocelot's max record on there.
http://www.doom2.net/compet-n/database/cn.cgi?map:udoom:17+2003:12
Your average DSDA demo would be similar to a 4 or 5 minute run on here... bringing that time down to 2:24 is amazing.
Note how he groups monsters and even avoids firing to conserve ammo for a better opportunity later... his speed, aim, and movement are nearly flawless.
If you are on a map that monsters will stray if awaken, then usually you should kill them as you wake them up. see Marswar map10 for a good example... Cacodemons and shotgun guys ride lifts and seep into the cracks... they follow you geographically, often locking themselves away in some dark recess of the map to never be found.
But if you are playing Hr22, infighting; no one will get lost. :)
I'd say weapon choice makes the biggest difference. One of the things that sets apart the very best players is their ability to never pause... they are always driving.
Watch Vile drive through a map... on UV monsters are coded to take a moment before attacking. If you attack during that time, they have no offense. Other than planned infighting, if you drive INTO the monsters, you tend to control them. Ever play a map and everyone seems to slowly close in on you, then you get to the next rooms and they are all empty. :)
Besides perhaps bad sound_triggering wad design, you are also allowing the map to control you.
Watch Anders' TVR map16 demo or Arno Slagboom's 10sector Map02 demo for examples.
I agree with Julian about making the whole thing come together and be faster.
There are several excellent HR2beta maxdemos in the next DSDA update where they use weapon choice perfectly, ssg-ing here so they can RL the masses of monsters to the exit.
I was always impressed with Ryback's long nightmare demos... knowing when to use his one BFG shot or at what point he must needs to use the RL because of respawn troubles... that was fascinating and awe-inspiring to me.
sheesh, I sure can ramble... sorry about that.

i just watched that 224 e1m7 run .... twice ... it's bloody amazing. what's best about it is the sheer audacity of it, e.g. launcing two rockets into a space and running of without even checking for survivors... someone who knows every square inch, er, pixel of doom!

as for the thread, i'll think about it and get back to y'all!

cycloid said:

i just watched that 224 e1m7 run .... twice ... it's bloody amazing. what's best about it is the sheer audacity of it, e.g. launcing two rockets into a space and running of without even checking for survivors... someone who knows every square inch, er, pixel of doom!

You gotta memorize where all the monsters are when you're really shooting (pun intended) to squeeze seconds from the final time, but unfortunately the element of luck is also required in doing so. In that case, ocelot couldn't stand around to check if the rockets eliminated all the enemies, but had to move on after the right number of shots were spent.

E1M7 also has a reputation for possessing the "99% kill bug", something the demo crowd will refer to (jokingly) as a map which has a very high possibility for missed monsters when optimizing the time. It also happens to be one of the front runners in that category. What does that equate to? Ocelot's motivation, as well as technique, were unstoppable there. :)

These are just some strategies I use. Others might do differently.

Planning a max:
* play through the level a couple of times so that you're familiar with all the monsters and items. Note the position of weaponry and secrets, as these will be what you plan the route around.
* using IDDQD can be useful to see what infights are possible, and also how monsters will react to you trying various things.
* once you know the route to the exit well (keys, switches etc), use IDDQD and try skipping monsters, running through crowded rooms. Be as audacious as possible. See what gets stuff killed the fastest then try to do that in a regular run.
* the best routes are those that allow you to use the ammo most efficiently, and having you firing more or less continuously. Running the shortest route through the level is not particularly important.
* you should be aiming to collect monsters in groups. This is because you don't have to waste time aiming at them individually and they die faster. But mainly because it's more efficient to get all the monsters coming to you, rather than you going to them. That's because there's hundreds of them and only one of you - they can attack in parallel while you're operating in serial mode :)

Techniques and ideas unique to max:
* infighting. Time in a max is determined by how long it takes to kill everything, so every monster you don't have to kill is a saving. Wasting time in order to provoke infights, though, is generally not needed unless there is an ammo shortage or a severe map imbalance (eg map08).
* efficient ammo usage. You should have a plan for every rocket and every cell. Know how the BFG works, and the damage done by each weapon to each monster.
* always line up monsters when firing at them. If your shot misses, it might hit the monster behind. If it hits, some pellets will pass through the monster and you want to use these pellets too. SSG in particular, with its long load time and large number of pellets, demands that you try to line up as many enemies in front of it as possible. Watch some Anders maxdemos for instruction.
* claudio used an interesting strategy in a DANG submission once. He picks up an invis sphere, not because he needs it, but to time his progress. When it runs out he checks how far he's gone in the level and knows whether he's over or under schedule. You need a way of timing your run. Work out what areas are slowing you down and come up with ways to improve them.

Handling missing monsters:
* know where monsters tend to congregate (playing through on the route a couple of times will tell you). Plan to include these spots somehow in your route.
* know the location of every locked door and fast-open door. Use these as gates. Know where the stairs that are too steep to climb are. Cybies and Spidermommas can't climb even one unit stairs. (i think). Check if lifts can be operated by monsters. Look for monster_impassable lines. Remember floating monsters are a special case.
* know how the monsters move (straight-arrow to the player, plus some random wiggling), and how fast they are. This lets you set an upper bound on where they can get to while you're handling other stuff. Some monsters are faster than others - the Archvile is rarely safe to leave alone.

Improving:
* a second or third exit will generally be ten to fifteen percent faster, no route changes.
* play more aggressively. Use the berserk fist, it has a damage hit rate greater than the SSG. Shoot monsters from closer range. Remember, RL is the only effective long-range weapon.
* time how long it takes to kill monsters and factor this into your route. If a certain room takes 4 SSG blasts, you should be moving to the exit as you fire the fourth shot.
* start playing the probabilities. If you know a certain route optimization will save twenty seconds but has a 40% chance of missed monsters, take it.
* keep circling at all times when shooting at monsters, this keeps them from spreading out.

If you want to see a lot of max strategies displayed, watch the Compet-N demos for Doom2 map10. See how the routes and playing styles have evolved...

Very nice post by Ryback there. A lot of that stuff is hard for the players to describe because it's second nature by now, but that pretty much sums it all up.