Jump to content
Search In
  • More options...
Find results that contain...
Find results in...
Linguica

Empirical analysis of onscreen monster counts

Recommended Posts

I decided to do an empirical analysis of Doom IWADs to see how many enemies were ever onscreen at once.

I downloaded UV-Max demos from DSDA and ran them through a hacked Chocolate Doom that output the number of live monsters visible on screen each tic. (Note I say "visible" - if you are facing a wall, but there are 100 monsters behind you, it outputs 0. I felt that the number of active monsters actually being drawn onscreen was the most straightforward measurement, though.) I felt that UV-Max was the category that best simulated "normal" play, since the player is actually fighting all the monsters instead of just running past them.

The X-axis is seconds, Y-axis is the max # of monsters visible during that second. You may be surprised at just how small most of the "hordes" in the IWADs actually are.

Knee-Deep in the Dead:



The Shores of Hell:



Inferno:



Thy Flesh Consumed:



Doom 2 Map01 - Map10:



Doom 2 Map11 - Map20:



Doom 2 Map21 - Map30:



So as you can see, it's quite rare to even have 20 monsters onscreen at once - the only time it happens in Doom 1 is in E3M6 (I think). Episode 2 is more sedate than I expected, with it barely ever hitting 10 monsters, and Episode 4 definitely has smaller hordes than I expected.

The hordes seem to be generally a little bigger in Doom 2, hitting 30+ monsters on several occasions. The one outlier at 45+ monsters is (I believe) Map18, where you can let a ton of monsters congregate in the central courtyard.

Share this post


Link to post

Haha awesome work. This will be a nice trump card to take out when the discussion inevitably turns to the 'hordes' in Doom versus other modern games. I'm honestly not surprised at all, but I was paying more attention. Next it would be interesting the see the average age of people that continually make these claims...

Share this post


Link to post
Linguica said:

(Note I say "visible" - if you are facing a wall, but there are 100 monsters behind you, it outputs 0. I felt that the number of active monsters actually being drawn onscreen was the most straightforward measurement, though.)

That seems fairly arbitrary and misleading since this figure belies the exact nature of the threat that the player faces at any one moment. You say that you've measured "how small most of the 'hordes' in the IWADs actually are," but your numbers objectively do not measure this. What you're measuring is exactly what you've stated: "number of active monsters actually being drawn onscreen".

Even if not onscreen, an active monster roaming the level is much different from an idle (or non-spawned and ergo non-existent) monster. An obvious example: MAP11 of Plutonia isn't memorable just because you fight one, maybe two Arch-viles at a time through the map. The entire point is that you have 12 active Arch-viles immediately hunting you, not to mention the several more you can wake up to boost that active count. Another example: MAP07 of Doom 2. You fight two waves of 6 Mancubi and 12 Arachnotrons, respectively. Naturally, you'd be hard pressed to see all at once, especially the Arachnotrons, but knowing that you've got about 10 more rapid fire plasma-spewers barreling down on you when while you're working on the 2 or 3 you can currently see on your screen is a pretty big part of the experience.

Similarly, only counting the onscreen enemies doesn't reflect the actual danger that the player faces. The starting room of E1M2 has over a dozen Zombiemen in it, but you'd have to intentionally corral them to see them all at once. Nonetheless, having 12 monsters on the player at the same time even if only a few are visible onscreen changes the dynamic of the gameplay. The red key trap alone in E1M6 opens up 17 simultaneous monsters, but I don't think your numbers capture that.

In fact, using a UV-Max demo for this purpose is arguably flawed procedure as well since speedruns of this caliber certainly aren't "normal" gameplay. They're highly planned out and choreographed routes that maximize monster deaths against time. Monsters being cleared out as efficiently as possible may mean less will be onscreen at once, and the numbers may even be skewed in the other direction as it's not uncommon for speedruns to intentionally instigate large scale infighting to save killing time.

If we wish to compare Doom to Doom 2016 (are we trying to tiptoe around mentioning it?), I'm afraid there isn't a particularly good or fair grounds for comparison. Even if you were to do something like count the number of visible monsters within a full 360 degree view range of the player, this would still miss those about to turn around corners or traveling from other rooms and such. Simply put, Doom and Doom 2016 are very different in their aims. Doom and Doom 2 feature large and open levels with permanent static enemy placements, while Doom 2016 features scripted wave-based combat in confined arenas. The former is quite fluid and less predictable in its encounters, while the latter is always fixed.

Anyway, as long as you're measuring the IWADs, is there any reason for excluding Evilution, Plutonia, and even NRFTL?

Share this post


Link to post

You seem to be reading all sorts of judgments into these graphs that aren't there. Feel free to do it better.

Share this post


Link to post

That's the point: It can't be done better as, at least in the sense of measuring "how small most of the 'hordes' in the IWADs actually are" for the purposes of comparing it to other games, namely Doom 2016, it can't realistically be done at all. This is a fine data set for what it is, but ultimately it's inviting comparisons of apples to oranges.

Perhaps this would seem less inclined to a particular viewpoint were the "small hordes" bit was omitted as this statement leads the viewer in regards to an aspect that the data does not actually correlate to. Of course, the definition of "horde" is well up for debate as well, but I imagine most people would consider this to be something along the lines of "the number of monsters wot are trying to kill me" and not "how many can I see on my screen at once?". One might not want to look back at that gaggle of Cyberdemons on your tail, but rest assured, they are indeed part of the horde you're facing (not literally facing, but you know what I mean).

Share this post


Link to post

Very interesting, thanks for putting this together. I'm curious to see data for Final Doom as well.

Share this post


Link to post

Well, even with a conservative 10-12 monster "horde" average, Doom would still regularly beat most FPSs released after it by an order of magnitude. If anything, this analysis reveals that you don't actually need 100s or 1000s of visible monsters (2 and 3 orders of magnitude) to convey that "horde" feeling.

Besides, with more than 20-30 visible monsters, vanilla slowed to a crawl back in the day.

Share this post


Link to post

This will be a nice trump card to take out when the discussion inevitably turns to the 'hordes' in Doom versus other modern games.


How? Like Maes says, even a dozen monsters on screen is far above your average corridor shooter pitting you against three or four at a time. Most discussions about hordes in Doom have assumed a number between 10 and 16 as reference, this changes nothing...

Besides, visible monsters =! active monsters. Even in games more ambitious with their visible monster count, it's quite rare this is reflected in their active monster count. An unseen monster lurking towards you plays a role in the overall game flow, in regards to your movement options as well as your own strategy; this is a dimension lost in games that spawn monsters dynamically in the very room you're in.

Share this post


Link to post

I'm actually pretty surprised the numbers are so high. Though it's not unusual for a UV-MAX demo to skim past all the monsters and farm them all into a large area only to come back and blast them all at once.

The examples cited by Revenant100 are not too difficult to "simulate" in a modern game I don't think. You can still "feel" like you're facing hordes without actually fighting them and using the processing power of your device to make them actively seek the player out on their own. Doom doesn't draw the textures on walls you're not looking at, and there is nothing lost there.

I could be pulling this out of my ass but I think I read something about Dead Rising's massive hordes of zombies that can be displayed on screen at once and part of the reason that was possible is that many of the zombies are just hollow shells until you get close enough for them to become active. I'd imagine something similar could be done in video games. You gotta fake it till ya make it.

Share this post


Link to post

A big missed point here: Doom shows roughly a 90 degree view. So if there's 12 monsters in front of you (visible), there may be 12 to the right, 12 to the left, 12 behind, 12 behind a wall, 12 in a side corridor, etc.

Besides, 12 *is* a horde. Imagine walking home, and 12 guys jump out to rob you. I think most people would be giving up their wallet. And later, they might say: "A horde of guys jumped out and threatened me.", right?

Linguica:
Some interesting changes to your charts:
1. Instead of monster count, add up there health values, and display that. That will show the visible amount of threat. Divide by 100 to show the number in "human" units.

2. Add together all monsters in currently visible sectors. That would get the monsters behind you, as well as some in side corridors. Not perfect, but decent.

Good stuff, as usual, man!

Share this post


Link to post

Here's E1M6 with a further modified exe that also counts how many enemies can see the player each tic. By "see" I mean there is a direct LOS between them, whether or not the monster is actually activated or is targeting the player.

Share this post


Link to post
Phml said:

How? Like Maes says, even a dozen monsters on screen is far above your average corridor shooter pitting you against three or four at a time. Most discussions about hordes in Doom have assumed a number between 10 and 16 as reference, this changes nothing..


Uh.. no kidding dude, obviously Doom has more enemies on screen than FEAR. No one is arguing that. But Destiny? Serious Sam? Painkiller? My statement was more geared towards the flippant 'Fuck modern games I miss fighting hundreds of enemies like in Doom' claim that pops up every couple of years.

Phml said:

Besides, visible monsters =! active monsters. Even in games more ambitious with their visible monster count, it's quite rare this is reflected in their active monster count...


Yes a good point, but when I see or hear the term 'horde' my mind doesn't immediately turn to a couple imps dicking around after you in a hallway. To me a horde is opening a door and bam, face full of monsters. Teleporting into total chaos like in map21, shit like that. I think this graph is a fair representation of that term.

Share this post


Link to post

I can't really be bothered to do a lot of messing with graphs so here's the raw data:

https://www.doomworld.com/linguica/hax/episode1.txt
https://www.doomworld.com/linguica/hax/episode2.txt
https://www.doomworld.com/linguica/hax/episode3.txt
https://www.doomworld.com/linguica/hax/episode4.txt

https://www.doomworld.com/linguica/hax/d2ep1.txt
https://www.doomworld.com/linguica/hax/d2ep2.txt
https://www.doomworld.com/linguica/hax/d2ep3.txt

the format goes tic, # of enemies on player's screen, # of enemies with direct LOS to player.

Sometimes the second value is even larger than the third, because the second one counts an enemy as visible to the player if even a single pixel of their sprite is drawn on screen, whereas the third one is casting a line between their centerpoints.

Share this post


Link to post

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×