Systemic inaccuracy of monster pages

[onderduiker]

This is just a heads-up for anyone currently working on the Diablo II monster pages, or thinking of doing so in the future.

A monster's defence, attack rating, life, damage and experience are calculated based on their level. According to the Arreat Summit and this Wiki, that level is determined by the entries in the Level columns in MonStats.txt. However, since v1.10 that's only true for monsters in Normal, and boss monsters (as defined by the boss column of MonStats.txt) in all difficulty levels: monster levels in Nightmare and Hell are determined by area level. This means that the defence, AR, life, damage and experience listed for monsters in Nightmare and Hell will only be accurate if their MonStats.txt level is the same as that of the area in which they spawn.

For example, the Fallen page lists Fallen (fallen1 in MonStats.txt) as level 67 with 771-2,261 life in Hell, but this is only true of Fallen that spawn in the Blood Moor (level 67), which have 771-1,696 life off the Ladder and 1,028-2,261 life in Ladder and single-player games: those that spawn in the Underground Passage Level 2 (level 83) are actually level 83 with 1,116-2,455 life off the Ladder and 1,488-3,273 life in Ladder and single-player games. Fallen also spawn in six other areas of Act 1, each with a different level.

The definition of chill effectiveness listed on every page is also wrong:

Chill Effectiveness tells what % of a character's cold length applies to the monster. 100 is all, 0 is no freeze/chill at all.

Chill effectiveness is the -attack rate and -velocity applied when monsters are chilled: for example, when chilled, a monster with 50 chill effectiveness will have -50 attack rate and -50 velocity applied to it, the equivalent of being cursed with Decrepify or affected by level 12 Holy Freeze. The only time that it affects chill length is when it's <=0: a monster with 0 chill effectiveness cannot be chilled or frozen, or slowed by Holy Freeze (oddly, the aura's radial cold damage also won't be applied, although the cold damage added to weapon damage is if the monster isn't Immune to Cold).

Unfortunately this definition isn't part of a template, so correcting it would involve correcting each monster page individually.

I've also become aware of a few errors where there's been confusion of which monsters spawn where:

Stygian Fury
While it's true that Vile Witches spawn in the Worldstone Keep Level 1 in all three difficulty levels, those that spawn there in Normal (succubuswitch2 in Levels.txt) are different from those that spawn there in Nightmare and Hell (succubuswitch8). The resistances listed are only for the Vile Witches in Normal, which would be Immune to Fire (Fire Resist 110%) if they spawned there in Hell, but the Vile Witches that actually spawn there in Hell are Immune to Cold (Cold Resist 155%).

Megademon
There's been confusion of the Venom Lords that spawn in Act 4 (megademon3 in Levels.txt) and those that spawn in the Throne of Destruction in Act 5, as Ventar the Unholy and his minions (venomlord in SuperUniques.txt):

VENOM LORD   RESIST % (HELL)
             Damage   Magic   Fire   Lightning   Cold   Poison
______________________________________________________________
Act 4           50        0    130          75     50       75
Act 5            0        0     95           0      0      120

Incidentally, listing Resist % in a different order to that in MonStats.txt may also have led to errors in transcription.

I've explained how life is calculated in the Coldworm HP thread in this forum, and posted Monster Resistances.xls in the Diablo Files forum, a spreadsheet of the life and resistances of monsters in Hell (excluding the Pandemonium Event and the Diablo clone): where monsters spawn in Nightmare is identical to where they spawn in Hell (both are determined by the nmon columns in Levels.txt).

[Leord]

Good spot. This needs to be updated of course. I am currently not doing monsters though.

So, do we have to update every single monster article then?

[Elly]

As flux is the stat guy I talked to him about it yesterday and he said that the monster spawning by area thing changed in v1.11, and he didn't see the point/possibility of changing it. As that post says, monster hit points are determined by area they spawn in, and most of them can spawn in multiple areas. We'd need several hps values for almost every single monster in the game, and that would just be eye blurring in the wiki format. And it's kind of irrelevant, since they just go up or down like 1% from the listed values.

Perhaps we can just list the variance thing?

[SnickerSnack]

Onderduiker: can you (or anyone) give a description of what would need to be done in order to correct this? I have a bit of down time coming in december, so maybe I can help. This sounds like a minor, but glaring error that really should be fixed.

I have some familiarity with SAGE (and PARI [computing software]), so I may be able to write code to semi-automate the process.

[Leord]

What Elly said is also important. Might be we need a gazillion new tables and stats over relatively small improvement in info quality on the site...

[Saesa]

So for level 68 fallen spawning in Dark Wood it’s: 792-1743 and 1056-2324 for ladder games
For level 69 fallen: 814-1791 and 1085-2388
For level 80 fallen: 1051-2312 and 1401-3083
For level 81 fallen: 1073-2360 and 1430-3147
For level 85 fallen: 1159-2550 and 1545-3400

[onderduiker]

Good spot. This needs to be updated of course. I am currently not doing monsters though.

So, do we have to update every single monster article then?

Yes, at the very least to edit or remove the definition of chill effectiveness: anything else will depend on what you make of what follows.

As that post says, monster hit points are determined by area they spawn in, and most of them can spawn in multiple areas.

It's not just monster life that's level-adjusted, it's defence, attack rating, damage and experience as well.

And it's kind of irrelevant, since they just go up or down like 1% from the listed values.

While this is probably true most of the time, I've already given an example of when it's not: the maximum base life of Fallen in Hell is currently listed as 2,261, but the maximum base life of any Fallen (those in the Underground Passage Level 2, a level 83 area) is actually 3,273, which is almost 45% higher.

Perhaps we can just list the variance thing?

As far as I can tell the listed values already span the minimum values off-Ladder to the maximum values in Ladder and single-player games, so this is probably the best idea for the monster pages, to keep things clear.

If anybody were motivated to do so, they could then put tables of the level-adjusted stats for Nightmare and Hell on the area pages: at the very least this may be worth doing for level 85 areas, and any other areas that are frequently visited.

Onderduiker: can you (or anyone) give a description of what would need to be done in order to correct this? I have a bit of down time coming in december, so maybe I can help. This sounds like a minor, but glaring error that really should be fixed.

I have some familiarity with SAGE (and PARI [computing software]), so I may be able to write code to semi-automate the process.

It might be worth waiting until after v1.13 is released before doing anything, although unlike v1.10 I doubt the delay has anything to do with making extensive changes and is most likely due to it being low priority (although you never know...).

If you (or anybody else) wants to go ahead anyway, then you need to extract Levels.txt, MonLvl.txt and MonStats.txt from patch_d2.mpq: you could do so by downloading a utility like WinMPQ and the v1.11 data file (which still works for v1.12) from the Phrozen Keep, or I could extract the files and upload them to this site's Diablo Files forum (although it may be worth paying the Phrozen Keep a visit anyway, to consult the relevant file guides).

Levels.txt's nmon columns tell you the ids of the monster types randomly generated in each area in both Nightmare and Hell, and the MonLev2Ex and MonLev3Ex columns contain the expansion's area levels for Nightmare and Hell respectively. MonLvl.txt's columns contain the base stats for particular levels (column headings prefixed L- are for Ladder and single-player stats) and MonStats.txt columns contain the percentage adjustments for the ids of particular monster types.

That may sound bewildering until you've actually looked at the files (like Saesa), but I've explained how base life is calculated in the Coldworm HP thread, and the other base stats follow a similar principle for normal monsters. If I upload the extracted files, I'll post more detailed step-by-step instructions.

I've already posted Monster Resistances.xls, which contains levels, life and resistances for all monsters in Ladder and single-player Hell. You can use the filters to find the level and life ranges of particular monsters: for example, after filtering the Class column for Normal and the Monster column for Fallen, the drop-down box for Level tells you that Fallen are level 67-83, and those for Min and Max HP tell you they have 1,028-3,273 base life (just to be clear, I'm referring to the Fallen sub-type, not the whole species).

Although this won't help you with other Hell stats (aside from the level to use) or Nightmare stats, it will tell you where particular randomly generated monsters spawn in Nightmare since this is identical to where they spawn in Hell. It will also tell you about monsters that always spawn in set locations (like Fallen encampments) rather than being randomly generated.

[Flux]

It sounds like a project to embark upon after v1.13, assuming/hoping there will be some other changes to the monsters made in that patch. I saw the changes (in v1.11, IIRC) that made monster hit points vary so much by location, but didn't think it was worth the hours to add in those on every monster page, when the listed base hps were fairly close in most cases. And after all, it's not like anyone has ever read the wiki page hps, and then gone into the game planning only to do X damage to a monster and been surprised that it wasn't dead. Monsters are all treated with a "hit until dismembered" strategy anyway.

Wiki pages are clear and quick to read, but the drawback is that you can only present so much info on them. A javascript with hover options would work better for something like monster hps, which change dynamically based on so many options.

As for the chill effectiveness notes, I wrote those, and the ones on defense, foot speed, blocking, etc, based on the best definitions of those features I could find in forum discussions, back in early 2008 when I made those pages. I assume the game's function hasn't changed, so it's the player understanding of how chill works that's evolved/self-corrected?

[Leord]

We have discussed allowing certain javascripts in there, for the sake of tabbed navboxes, for instance. Besides, when Diablo 3 is released, it might be cool to have javascripts in there for item hovers etc. we should look in to this.

[onderduiker]

I saw the changes (in v1.11, IIRC) that made monster hit points vary so much by location, but didn't think it was worth the hours to add in those on every monster page, when the listed base hps were fairly close in most cases.

Again, it's not just life that's affected, it's defence, attack rating, damage and experience as well. Admittedly, complete accuracy isn't that important, and any significant differences between the levels from MonStats.txt and area levels in Nightmare and Hell are most likely restricted to early Hell (and possibly even just Act 1).

However, looking at Fallen (probably one of the worst cases), it's clear why stats shouldn't be listed for every possible level on the same page (particularly if the stats for Normal, Nightmare and Hell are listed in the same table):

Fallen     MonStats ID   # Areas   #Levels (Hell)   MonStats Level (Hell)
_________________________________________________________________________
Fallen     fallen1        8        7 (67-83)        67
Carver     fallen2        9        6 (68-83)        68
Devilkin   fallen3       10        8 (69-85)        85
Dark One   fallen4        8        3 (71-73)        71

Carver     fallen6        2        2 (81-82)        83
Devilkin   fallen7        1        1 (83)           83
Dark One   fallen8        1        1 (81)           83

Warped Ones (fallen5) never spawn as normal monsters, they only spawn as Colenzo the Annihilator's minions.

Implementing javascripts could be interesting: you could list the range of values for a monster in the table, then have a box with the more detailed stats pop up when you hover the cursor over it. I don't know how easy that would be to implement or use, though: it could also be confusing and annoying if there are too many boxes.

As for the chill effectiveness notes, I wrote those, and the ones on defense, foot speed, blocking, etc, based on the best definitions of those features I could find in forum discussions, back in early 2008 when I made those pages. I assume the game's function hasn't changed, so it's the player understanding of how chill works that's evolved/self-corrected?

As far as I know it's the understanding rather than the mechanic which has changed. I think the previous understanding of chill effectiveness was a misconception based on the following:

• Drain effectiveness is known to be a percentage modifying life and mana returned by % Life and Mana Stolen per hit, so it was reasonable to assume that chill effectiveness would be something similar.
• It would have been observed that monsters with 0 chill effectiveness couldn't be chilled or frozen: that's 0% chill or freeze length... hmmm...
• Although (I hope) it's common knowledge that chill and freeze length are affected by difficulty penalties (reduced to 1/2 in Nightmare and 1/4 in Hell), I'm not sure how many people realise that Cold Resist % doesn't just apply to cold damage, it applies to chill and freeze length as well: so if someone observed reduced chill or freeze length and didn't know that was caused by CR %, they'd conclude that it was caused by chill effectiveness.

I've done tests in both the unmodified game and -direct -txt mode after editing text files (to eliminate variables) to investigate how chill effectiveness actually works (and discovered that if Frozen Horrors weren't always Immune to Cold in Hell, -33 chill effectiveness (actually 33 coldeffect in MonStats.txt) would indeed result in them attacking and moving faster when chilled :shocked.

I noticed that on at least one monster page (unfortunately I can't remember which one), the Chill Effectiveness column heading was a link to a Chill Effectiveness page (although no text has been added to it yet). Instead of having definitions on every page, you could put a link to this page in the column heading so that if people didn't know what it was they could click on the link to find out. You could do something similar with the other definitions as well.