Author Topic: Dragon Warrior III (NES PCs)  (Read 5115 times)

Taishyr

  • Guest
Dragon Warrior III (NES PCs)
« on: August 28, 2008, 12:30:58 AM »
Weapons

Sword of Kings: Hero (120)
Thor's Sword: Solider (100)
Zombie Slasher: Pilgrim, Sage, Merchant (65)
Staff of Force: Wizard (55)
Golden Claw: Fighter (55)*
Sword of Illusion: Goof-Off (50)**
Iron Claw: Fighter (30)***
Iron Spear: Goof-Off (28)****

*This weapon basically forces you to get into a fight every step and is worthless in game. If you discount it in the DL for that reason, the Iron Claw is 30 power.

** One of a kind and requires the Goof-Off to be female. I'm vigin

*** Next best weapon for fighter.

**** Next best weapon for the Goof-Off.

Armor

Armor of Radiance: Hero (75)
Swordedge armor: Soliders (55)*
Angel Robe: Pilgrim, Sage, Wizard (35)
Flashy Clothes: Goof-offs (28)
Half Plate Armor: Merchant (25)
Cloak of Evasion: Fighter (20)

Helms

Iron Mask: Hero, Solider (25)
Iron Helmet: Sage, Pilgrim (16)
Turban: Goof-Off, Merchant (8)
Golden Crown: Wizard (6)*
None: Fighter (0)

NOTE: Everyone can use the Golden Crown, but it's Wizard's only helmet. Throwing it a bone here. Fighter can't use any helmets.

Shields:

Shield of Heroes: Hero (50)
Shield of Strength: Solider (40)
Silver Shield: Pilgrim, Sage (30)
Iron Shield: Merchants (12)
Leather Shield: Goof-Off (4)
None: Fighter, Wizard (0)


Hero:

STR: 132
AGI: 83
VIT: 166
INT: 75
LUC: 110

HP: 329
MP: 128

ATT: 252
DEF: 191

Physical: 100

Spells:

Lightning: 200 damage, 30 MP
Sleep: 75% Sleep, 3 MP.  Lasts an average of 1-3 turns.
Stopspell: 70% Silence, 3 MP.
Heallall: Full Healing, 7 MP

Notes: Armor of Radiance halves Fire and Ice.

Comments: Solo without Fendspell. Seems like a solid Heavy, though the MP issues don't help matters.

Soldier:

STR: 147
AGI: 35
VIT: 183'
INT: 26
LUC: 30

HP: 349
MP: 0

ATT: 247
DEF: 137

Physical: 99

Notes: Swordedge armor reflects half the damage back to the person attacking the Solider. Damage reflected is ITD/E.

Comments: Good HP and having the swordedge armor makes him something of a counter whore. Light/Middle. He can't handle healers or status users at all.

Merchant:

STR: 84
AGI: 68
VIT: 114
INT: 71
LUC: 53

HP: 236
MP: 0

ATT: 149
DEF: 79

Physical: 60

Comments: Half average damage, slow, no tricks, not durable. Light/Puny. Not a single redeeming DL trait here.

Goof-Off

STR: 35
AGI: 48
VIT: 79
INT: 45
LUC: 215

HP: 158
MP: 0

ATT: 88
DEF: 68

Physical: 35 (24 with the Iron Spear)

Notes: Wastes 25% of his turns doing nothing.

Comments: ..Kyar. Makes Merchant look like Orlandu with Attack up and math skill. Puny.

Fighter

STR: 178
AGI: 179
VIT: 132
INT: 37
LUC: 125

HP: 252
MP: 0

ATT: 228
DEF: 93

Physical: 91 (83 with the Iron Claw)

Comments: In spite of insane speed, he.. really has nothing. He lacks Alena's damage and durability, and doesn't seem to have the same crit rate. Middle.

Wizard

STR: 27
AGI: 129
VIT: 102
INT: 106
LUC: 146

HP: 197
MP: 206

ATT: 82
DEF: 105

Physical: 32

Spells:

Upper: Doubles defense, stacks to 999 defense, 3 MP
Slow: Lowers enemy speed to 0, 90% success rate, 3 MP
Blazemost: 180 damage, Fire elemental, 12 MP
Explodlet: 140 damage, Explosion elemental, 18 MP
Bikill: Double attack, doesn't stack, 6 MP
Chaos: 80% chance of confusion, 5 MP. See DW4 notes on what chaos does.

Notes: The Angel Robe immunes ID.

Comments: Upper's good at handling fighters. Chaos gives him the time to get it off most fights, and Blazemost is a solid 2HKO. Explodelet gives him a non elemental 2HKO if he needs it as well. Not as frail as you'd think, thanks to the scrubs in the cast.Heavy

Pilgrim

STR: 52
AGI: 73
VIT: 109
INT: 109
LUC: 125

HP: 216
MP: 205

ATT: 117
DEF: 117

Physical: 47

Spells:

Infermost: 100 damage, Wind elemental, 9 MP
Stopspell: 70% Silence, 3 MP
Sleep: 75% Sleep, 3 MP.  Lasts an average of 1-3 turns.
Heallall: Full Healing, 7 MP
Revive: 100% Revival, 20 MP
Speedup: Doubles speed up to 255 AGI, 3 MP
Sap: Cuts target's defense in half. Two uses lower defense to 0. 90% hit rate? Unsure, it's high but I know it can miss.
Surround: 72% chance of Blind, 4 MP
Beat: 50% chance of ID, 7 MP

Notes: The Angel robe immunes ID.

Comments: Tank a turn, speedup, set. You basically need to kill Pilgrim in a turn or get status/chipped to death. Sort of similar to Angelo. High Heavy.

Sage:

STR: 69
AGI: 84
VIT: 112
INT: 87
LUC: 98

HP: 222
MP: 143

ATT: 134
DEF: 123

Physical: 53
Physical after two uses of Sap and one use of Bikill: 268

Spells:


Infermost: 100 damage, Wind elemental, 9 MP
Stopspell: 70% Silence, 3 MP
Sleep: 75% Sleep, 3 MP.  Lasts an average of 1-3 turns.
Heallall: Full Healing, 7 MP
Revive: 100% Revival, 20 MP
Bounce: Reflects magic
Speedup: Doubles speed up to 255 AGI, 3 MP
Sap: Cuts target's defense in half. Two uses lower defense to 0. 90% hit rate? Unsure, it's high but I know it can miss.
Surround: 72% chance of Blind, 4 MP
Beat: 50% chance of ID, 7 MP
Upper: Doubles defense, stacks to 999 defense, 3 MP
Slow: Lowers enemy speed to 0, 90% success rate, 3 MP
Blazemost: 180 damage, Fire elemental, 12 MP
Explodlet: 140 damage, Explosion elemental, 18 MP
Bikill: Double attack, doesn't stack, 6 MP
Chaos: 80% chance of confusion, 5 MP. See DW4 notes on what chaos does.

Notes: The Angel robe immunes ID.

Comments: ..Yeah. Skillset is mind bogglingly awesome. Half dozen decent statuses to use, good 2HKO damage.. basically combines the best of Pilgrim and Wizard into one package. Heavy/Godlike.

Average damage:

1. Lightning, 200 (Hero)
2t. Blazemost, 180 (Wizard)
2t. Blazemost, 180 (Sage)
4. Infermost, 100 (Pilgrim)
5. Physical, 99 (Soldier)
6. Physical, 91 (Fighter)
7. Physical, 60 (Merchant)
8. Physical, 35 (Goof-Off)

Average: 118

HP:

1. 349 (Soldier)
2. 329 (Hero)
3. 252 (Fighter)
4. 236 (Merchant)
5. 222 (Sage)
6. 216 (Pilgrim)
7. 197 (Wizard)
8. 158 (Goof-Off)

Average: 245

AGI:

1. 179 (Fighter)
2. 129 (Wizard)
3. 84 (Sage)
4. 83 (Hero)
5. 73 (Pilgrim)
6. 68 (Merchant)
7. 48 (Goof-Off)
8. 37 (Soldier)

Average: 87 (78 if you toss Fighter)

DEF:

1. 191 (Hero)
2. 137 (Soldier)
3. 123 (Sage)
4. 117 (Pilgrim)
5. 105 (Wizard)
6. 93 (Fighter)
7. 79 (Merchant)
8. 68 (Goof-Off)

Average: 113

I tested on 70 defense enemies, and went by the Iron Claw/Sword of Illusions for equipment. Took this at 580k EXP, which is a bit oveleveled.

hinode

  • Enough expository banter! Now we fight like men!
  • DL
  • Denizen
  • *
  • Posts: 1352
  • And ladies! And ladies who dress like men!
    • View Profile
Re: Dragon Warrior III (NES PCs)
« Reply #1 on: December 22, 2008, 02:01:00 AM »
Stuff I found while browsing old gamefaqs board topics.

http://www.gamefaqs.com/boards/genmessage.php?board=587249&topic=46263973

Seems like Luck is used for status defense in this game.

http://www.geocities.com/arcanelore2001/articles/dw3bestiary.txt
http://www.geocities.com/arcanelore2001/articles/dw3monspell.txt
http://www.geocities.com/arcanelore2001/articles/dw3monact.txt

Stats, resists, and AI patterns for monsters in this game. Curiously, it seems the 3 types of fire magic actually share a single resist stat on the monster side in this game, unlike subsequent DQs.

superaielman

  • "Mordero daghain pas duente cuebiyar/The fear of death holds not my heart!"
  • Denizen
  • *
  • Posts: 9630
    • View Profile
Re: Dragon Warrior III (NES PCs)
« Reply #2 on: December 22, 2008, 04:15:20 PM »
I've suspected that luck effects status resists but all of my tests have come up with a grand total of nothing conclusive on the subject.
"Reputation is what other people know about you. Honor is what you know about yourself"- Count Aral Vorkosigan, A Civil Campaign
-------------------
<Meeple> knownig Square-enix, they'll just give us a 2nd Kain
<Ciato> he would be so kawaii as a chibi...

SnowFire

  • DL
  • Denizen
  • *
  • Posts: 4934
    • View Profile
Re: Dragon Warrior III (NES PCs)
« Reply #3 on: August 07, 2009, 05:07:55 AM »
It came up in chat awhile back, and I was bored, so...  averages without Goof-Off.

Why?  Straight Goof-Off (that is, a Goof-Off who doesn't switch to Sage) is the rare case of what's obviously a "challenge" character that is not meant as a serious alternative.  The fact that he's cast worst at everything but luck (okay, and speed, but not last there just barely) is a good tip-off.  Obviously this is a dangerous policy to follow in general - we don't throw out the scrubs in Suikoden casts - but this case is unusually clear cut.  IIRC, the instruction manual outright says that Goof-Off is basically hard mode.  Kinda like the Radiata Stories trainwreck, where some characters are clearly "earlygame only" and don't really get better as the game progresses to give you an incentive to mix in new characters, but they're still technically legal to use.

Besides, there's still Merchant as cast scrub to make everyone else look good.  At least he's clearly meant as a serious option, even if "monsters magically carry around more treasure chests containing Medicinal Herbs!" turns out not to make up for sucking in battle.

Average damage:

1. Lightning, 200 (Hero)
2t. Blazemost, 180 (Wizard)
2t. Blazemost, 180 (Sage)
4. Infermost, 100 (Pilgrim)
5. Physical, 99 (Soldier)
6. Physical, 83 (Fighter) (Iron Claw)
7. Physical, 60 (Merchant)
( Physical, 26.1, (Goof-Off) ) (35 * .746 to include wasted turns)

Average: 129 (323 kill point) (116 w/Goof-Off, 290 kill point)

HP:

1. 349 (Soldier)
2. 329 (Hero)
3. 252 (Fighter)
4. 236 (Merchant)
5. 222 (Sage)
6. 216 (Pilgrim)
7. 197 (Wizard)
(158 (Goof-Off))

Average: 257 (245 w/Goof-Off)

AGI:

1. 179 (Fighter)
2. 129 (Wizard)
3. 84 (Sage)
4. 83 (Hero)
5. 73 (Pilgrim)
6. 68 (Merchant)
( 48 (Goof-Off))
7. 37 (Soldier)

Average: 93 (87 w/Goof-Off)
I dunno about throwing Fighter out completely, but I can see adjusting this down a bit in light of the fact that past a certain point Fighter's speed is useless overkill in a turn-based system.  Still good for Defense, mind, but not so much for initiative.

DEF:

1. 191 (Hero)
2. 137 (Soldier)
3. 123 (Sage)
4. 117 (Pilgrim)
5. 105 (Wizard)
6. 93 (Fighter)
7. 79 (Merchant)
(68 (Goof-Off))

Average: 121 (113 w/ Goof-Off)
« Last Edit: November 18, 2014, 05:14:20 AM by SnowFire »

Reiska

  • DL
  • Denizen
  • *
  • Posts: 251
    • View Profile
Re: Dragon Warrior III (NES PCs)
« Reply #4 on: November 15, 2014, 05:37:04 AM »
[Edit 2/1/2021: Adding some more mechanics obscura here which I had back in 2014 but forgot about until now.  As before, these are courtesy of slartifer and/or 1whoistornapart from GameFAQs]

[Edit 8/13/2021: Adding a few more bits of extreme obscura re: the Chance spell, credit to airkix72 from the DW3 speedrunning community on Twitch for digging this data up and to MisterHomes for the run that prompted even looking into it]

I've included everything from the old post as well, just so that everything is in one place.

Damage

DQ3 NES's standard damage formula is: DMG = (ATK-(DEF/2)) * [102..153]/256
This makes ATK/2-DEF/4 a reasonable approximation, with variance of roughly +/- 20%.

If the formula would return less than 1 damage, you have a 50% chance to miss and a 50% chance to hit for 1. 

The weapons that do bonus damage to specific enemy types add a flat bonus of 16-31 damage.  These are Dragon Killer (against any enemy flagged as a "dragon") and Zombie Slasher (checks resistance to Expel).

Bikill status doubles damage on non-critical hits.  Critical hits are not affected by Bikill.

An attacker using the Multi-Edge Sword will receive backfire damage equal to (Damage/4)+1.
A defender wearing the Swordedge Armor will cause attackers to receive backfire damage equal to (Damage/2)+1 if the attacker fails an Expel resistance check.

Poison Needle bypasses the standard attack checks, including enemy evasion (read: it's autohit).  It has a 1/8 chance to kill a non-boss target outright (regardless of Beat resistance).  Otherwise, it will do exactly 1 damage.  It cannot instant-kill bosses, and will always hit them for 1 damage.

Critical hits

Critical hits use their own damage formula: DMG = ATK * [54..64]/64

In other words, a critical hit does between 84.375% and 100% of the attacker's ATK in damage (average ~92%) and ignores the target's defense.

Most PCs have a 1/64 chance to critically hit on each attack.  A PC wielding the Demon Axe has a 1/8 chance to critically hit which *replaces* the standard critical hit rate.  A Fighter has a (Level/256) chance to critically hit *in addition to* the standard critical hit rate.

Evasion

PCs have a base 1/64 chance to evade any physical attack.  Enemies can have different rates, but 1/64 is the most common evasion rate for enemies as well.  After the standard evasion check, the following additional checks are made as well:

A PC wielding the Demon Axe has a 1/8 chance to miss.
A PC afflicted by Surround status has a 5/8 chance to miss.
A PC wearing a Cloak of Evasion has a 13/64 chance to evade any physical attack.  (Note that this is a separate roll from the standard 1/64.)

Magic damage

Every spell in DQ3 has a set damage range that it deals, and these ranges differ for PCs and enemies.

The ranges are as follows for PCs:
Blaze 8-13, Blazemore 70-89, Blazemost 160-199
Firebal 16-23, Firebane 30-41, Firevolt 88-111
Bang 16-23, Boom 52-67, Explodet 120-159
IceBolt 25-34, Snowblast 42-57, Snowstorm 88-111, IceSpears 60-79
Infernos 8-23, Infermore 25-54, Infermost 60-119
Zap 70-89, Lightning 175-224

The ranges are as follows for enemies:
Blaze 7-12, Blazemore 52-62, Blazemost 92-128
Firebal 10-18, Firebane 22-34
Explodet 60-80
IceBolt 16-24, Snowblast 32-42, Snowstorm 55-67
Infernos 6-18, Infermore 14-34, Infermost 30-62

Additionally, these are the damage ranges on enemy breath attacks:
"Weak" fire breath (Flamapede, Heat Cloud, Lava Basher, Wyvern, Voodoo Warlock): 6-10
"Medium" fire breath (Evil Mage, Sky Dragon, Lava Basher, Wyvern, King Tortragon, Orochi, Hydra, Green Dragon, King Hydra): 30-40
"Strong" fire breath (Salamander, Baramos, Baramos Bomus): 80-100
"Weak" ice breath (Demonite, Glacier Basher, Snow Dragon, Frost Cloud, Minidemon, Scalgon, Archmage, Putregon): 9-21
"Medium" ice breath (Snow Dragon, Scalgon, Putregon, Zoma): 40-60
"Strong" ice breath (Zoma): 100-140

RobMagic normally steals 5-10 MP from its target if it hits; due to a bug, when used against a target under Bounce, the caster usually gains a very large random amount of MP due to the game reading garbage data.

The "Chance" spell

In random battles, Chance generates a random number between 0 and 16 and chooses an effect based on that number as follows:

0 - All surviving enemies are forced to flee the battle, ending it and awarding EXP/gold for any enemies killed previously
1 - Restores 100% HP to all living party members
2 - Revives all dead party members with 100% HP and restores 100% HP to all living party members
3 - Restores [41..48] HP to all living party members
4 - All party members and enemies become Confused [ignores status protection/resistance]
5 - All enemies are instantly killed [ignores status protection/resistance]
6 - All enemies are instantly killed [ignores status protection/resistance]
7 - All spells are nullified for the remainder of the battle
8 - All party members' physical attacks are guaranteed to critically hit for the remainder of the battle
9 - All party members and enemies are put to sleep [ignores status protection/resistance]
10 - All remaining actions in the current turn are skipped; for the next 3 turns, only the caster of Chance may act
11 - The marching order of the party is randomized
12 - The current battle immediately ends with no EXP or gold awarded, even for enemies previously killed
13 - All remaining enemy actions in the current turn are skipped; for the next turn, only the party acts
14 - Drains all MP from all enemies and gives it to the caster
15 - Cures Numb (paralysis) status for all party members
16 - Does nothing

In any scripted battle, Chance generates a random number between 0 and 15.  (In technical terms, I believe it generates a number between 0 and 255 and then simply ignores the four high bits, but this shouldn't actually matter.)  If the generated number is 12 or greater, then it restores 100% HP to all living party members; otherwise, it does nothing.

Initiative, speed, etc.

At the start of any non-boss battle, enemies have a 1/32 chance to ambush the party.
If they do not, the party then has a 1/32 chance to ambush the enemies.

Turn order is loosely determined by Agility, but the variance is extreme: the game picks a random number between 25% and 100% of each combatant's Agility stat, and then each combatant acts in order from highest result to lowest result.  In case of ties, priority goes through all PCs left to right, then all enemies left to right.

Enemy HP and regeneration

Enemies are generated with HP randomly between 75% and 100% of their maximum.

Some enemies regenerate HP each round (mostly bosses) as follows:
Weak regeneration (16-23 per turn): Scalgon
Medium regeneration (44-55 per turn): Stone Hulk, Troll King, Baramos Bomus, Baramos Gonus, Kandar 2
Strong regeneration (90-109 per turn): Orochi, Boss Troll, King Hydra, Baramos, Zoma

Enemy regeneration and/or other healing is capped at maximum HP, regardless of the HP value the enemy actually generated with.

Encounter rates and the Golden Claw
DQ3 uses a simple probability check for random encounters happening, made every step, and varying by terrain.  For every step, there is an X/256 chance for a battle to start, where X is equal to:

On ocean: 4 in the daytime, 5 at night.
On plains: 10 in the daytime, 13 at night.
On bushes or desert: 15 in the daytime, 19 at night.
On forest or swamp: 18 in the daytime, 22 at night.
On hills: 25 in the daytime, 31 at night.
On any tile in a dungeon: 10 regardless of time.
When stepping through a screen transition in a dungeon such as a doorway: 84 regardless of time.

If you're carrying the Golden Claw, X is always 100 regardless of terrain and time.  Therefore, it multiplies the encounter rate by roughly 6.

Goof-Off Antics

Goof-offs learn new "antics" in the same way spellcasters learn spells.  All goof-off antics use the 50/50 method of spell learning, meaning that once the minimum eligible level is reached, the antic has a 50% chance to be learned every level. 

A Goof-Off has a 65/256 chance to perform an antic each turn instead of the command you selected.  Here is the list of antics with the minimum level each can be learned; if no effect is stated in parentheses, the antic simply wastes the Goof-Off's turn.

L2: Assess the situation
L3: Sing a song
L5: Take a break
L7: Smile
L10: Get lost in thought
L11: Kind glance
L12: Cheerlead
L14: Attempt to flee (Prints the normal message for attempting to run, but always fails)
L15: Trip over self
L16: Spruce up
L18: Parry (actually executes the PARRY command)
L19: Beckon the heavens
L20: Hone body
L21: Smug remark
L24: Recoil
L26: Point finger
L27: Pinch
L28: Jazzercise
L29: Curious dance (Prints the same message as the enemy action, but always fails)
L30: Call for reinforcements (Prints the same message as the enemy action, but always fails)
L31: Fake Explodet
L32: Pokes dead monster
L33: Fake arms merchant
L34: Chew gum
L35: Numb self (Inflicts Numb status on the Goof-Off!)
L37: Feign death
L39: Badmouth
L40: Satori
L45: Bluster
L50: Soft soap
L60: Expose tattoo (Only female goof-offs will use this antic; male goof-offs will still "learn" it, but if it's selected in the process described below, they'll perform the command you selected instead of doing it)
L70: Blow kiss
L85: Awful pun
L95: Sock whiff

The antics are organized into magic "pages" as if they were actually magic, and the game selects which antic the Goof-Off uses the same way it would select the spell used by a confused character: it chooses a page of the magic list at random (with equal probability), then chooses a learned spell from that page (again with equal probability).  These are the "pages" for Goof-Off antics:

Page 1 - beckon, recoil, take a break, lost in thought, spruce up, satori, hone body, jazzercise
Page 2 - sock whiff, assess situation, poke dead monster, trip over self, sing, dance, chew gum, smile
Page 3 - smug remark, point finger, badmouth, feign death, call reinforcements, attempt to flee, bluster, fake Explodet
Page 4 - pinch, cheerlead, expose tattoo, fake arms merchant, awful pun, blow kiss, soft soap, kind glance
Page 5 - numb self, parry

As mentioned above, if a male Goof-Off attempts to use the "expose tattoo" antic, it will be ignored, and the Goof-Off will instead use the intended command.

Running from battle

Under normal circumstances, the chance to successfully flee from battle in DQ3 is 50% on the first and second attempts, 75% on the third attempt, and 100% on the fourth attempt.

If the Hero's level exceeds the level of all enemies in the battle by 10 or more, fleeing will always succeed (unless the battle is a boss battle).  The highest level of any normal enemy in the game is 45, so a level 55 Hero's party will always succeed at running away from almost any battle.  (Exception: Red Slimes, for some reason, are flagged as level 63, which is the level assigned to boss enemies.)

Fleeing will always fail if the battle is not a random encounter.

Resistance values

Enemies in DQ3 have a resistance value between 0 and 3 for each of the following categories of attacks:
Blaze/Blazemore/Blazemost/Firebal/Firebane/Firevolt/Bang/Boom/Explodet
IceBolt/Snowblast/Snowstorm/IceSpears
Infernos/Infermore/Infermost
Zap/Lightning
Beat/Defeat
Sacrifice
Sleep
StopSpell
Sap/Defence
Surround
RobMagic
Chaos
Limbo/Slow
Expel

Resistance values against a magic category cause spells in that category to fail at the following rates:
0 resistance = 0%
1 resistance = 77/256 (approximately 30%)
2 resistance = 179/256 (approximately 70%)
3 resistance = 100%

Enemy targeting

Enemies' basic physical attacks will obey one of four targeting flags set on a monster-by-monster basis.  The four possible targeting flags are as follows:

Normal: Target is chosen weighted by positioning at the start of the round.  (The weighting will be explained below.)
Unweighted: Target is chosen randomly with an equal chance of hitting any party member at the start of the round.
Smart: Target is chosen weighted by positioning at the monster's turn (meaning it'll adjust for party members dying before its action).
Concentrated: Target is chosen randomly with an equal chance of hitting any party member at the start of the battle, and for the duration of the battle, all concentrated-targeting enemies in the battle will hit *only* this party member.

The targeting weighting uses four sequential rolls, presuming four party members:
First, the attack has a 113/256 chance to hit the lead party member.
If it does not, then it has a 180/256 chance to hit the second party member.
If it does not, then it has a 235/256 chance to hit the third party member.
If it does not, then it will hit the fourth party member.
This overall results in targeting chances by slot of approximately 44.14%/39.28%/15.22%/1.36%.

All attacks *other* than basic physical attacks always use unweighted targeting.

Other enemy AI bits

Some DQ3 enemies are flagged as "MP aware".  Enemies with this flag will not attempt to cast spells they lack sufficient MP to cast and, if they are afflicted with StopSpell, will not waste turns trying to cast spells.
« Last Edit: August 13, 2021, 11:22:27 AM by Reiska »

SnowFire

  • DL
  • Denizen
  • *
  • Posts: 4934
    • View Profile
Re: Dragon Warrior III (NES PCs)
« Reply #5 on: November 18, 2014, 05:13:02 AM »
Hey, let's talk about an old not particularly good game some more with dated mechanics!

So...  not that it matters too much for me...  but on reflection, Jester's 25.4% chance of turn skip should really be reflected in the damage average, it's kind of like a particular brand of horrible accuracy, except worse (since it might mean failing a crucial item heal in-game).  Of course I'd only use that damage average for Jester herself, because Jester = Jogurt, but still.  Also I didn't notice but Fighter's damage was listed with the cursed Golden Claw rather than the Iron Claw, and yeah no to that.  Edited the above damage numbers that I'd posted.

For Defense, thanks to the formula from the stat topic...  it turns out that an enemy with Attack of 267 will deal 104.5 damage to the cast, which is 40% of the no-Goof Off HP average.  (I have no idea if any such enemies EXIST, but hey, DL assumptions & all.)  (EDIT: From Reiska's list, Zoma is better than this, final dungeon randoms are worse.)  This comes up with "effective defense" multipliers of the following:

1. Hero - 0.83
2. Soldier - 0.96
3. Sage - 0.99
4. Pilgrim - 1.01
5. Wizard -1.04
6. Fighter -1.07
7. Merchant - 1.10
8. Goof-off - 1.13

Yeah, it's a pretty narrow curve.  Usual disclaimer that subtractiveish defense is more potent against weak attacks and less useful vs. very strong attacks (but very few super-strong physicals exist).

For speed, since the speed formula is now known...  this gives a more attractive way to compensate for the fact Fighter has super-overkill agility than throwing him out, while still reflecting that Agility past a certain point has diminishing returns.  I propose that finding the agility score in which the cast, on average, goes before this agility half the time and after this agility half the time is a fair "true" agility average.  Giving Fighter an extra few points of Agility only knocks him from 91% chance to 94% chance, while those same few points to a median dueler help more.  (Also note that this devalues agility to a degree - there are other games where 101 speed always goes before 100 speed despite only statistically being a tiny bit better.)  As it happens, the enemy speed whom party members (no goof off) go before/after it 50% of the time on average is 83.  (With goof-off, 78).

Here's the results:

By % chance of beating out an 83 Spd enemy:
1. Fighter - 91%
2. Wizard - 78%
3. Sage - 51%
4. Hero - 50%
5. Pilgrim - 40%
6. Merchant - 35%
(7. Goof-Off - 16%)
8. Soldier - 7%

So yeah, even slugs like Soldier occasionally go first, and while Sage likes being mildly above-average now, he can't exactly rely on outspeeding consistently.

No, I don't have any idea what actual speed values of enemies are in-game, all of the tables seem to be reprints of the chart that came in the NES box that was quite extensive for the day, but listed only HP / treasure drops.  I suspect it'd be something more like 60, which would make the likes of Fighter look worse (geez you already win) and everyone else better.

EDIT: So sure, I threw together the results based on a sample of enemies from Reiska's data (including 2 bosses) - so the average of when you go first vs. an enemy with AGI 99, 80, 72, 67, 59, 43 are the following:

1. Fighter - 94%
2. Wizard - 85%
3. Sage - 64%
4. Hero - 63%
5. Pilgrim - 54%
6. Merchant - 50%
(7. Goof-Off - 28%)
8. Soldier - 16%

Average is 61% chance of outspeeding that packet of enemies (57% w/ Goof-Off).
« Last Edit: November 21, 2014, 12:26:51 AM by SnowFire »

Reiska

  • DL
  • Denizen
  • *
  • Posts: 251
    • View Profile
Re: Dragon Warrior III (NES PCs)
« Reply #6 on: November 20, 2014, 08:34:57 AM »
I went ahead and crunched the following numbers.  These are enemy averages for HP, ATK, DEF and AGI on final dungeon randoms, excluding Metal Babbles for reasons which ought to be obvious.

HP: 184.85
ATK: 134.62
DEF: 76.23
AGI: 65.46

Here's the individual monster data.


Enemy           HP      MP      ATK     DEF     AGI     Other notes
Salamander      200     0       150     59      70      Dragon-typed; either attacks physically (with unweighted targeting) or breathes strong fire (iirc ~60ish? fire damage MT)
Goopi           70      0       90      75      59      Uses concentrated targeting; often calls adds (either more of itself or Granite Titans)
Voodoo Warlock  150     255     80      60      72      Casts Healall/Increase/Blazemore; sometimes does minor MP busting or calling (trivial) Hork adds; occasionally breathes weak fire (~15 fire damage MT)
Metal Babble    6       255     55      1023    150     flees 5/8 of the time, attacks 1/8, casts Firebal 2/8
Gold Basher     210     0       165     50      60      non-weighted targeting; flees 1/4 of the time; rare drop Golden Claw (1/256)
Granite Titan   350     0       200     90      35      double-attacks 50% of the time; 1/4 crit chance
Archmage        130     100     80      150     75      Casts Explodet/Revive; occasionally breathes weak blizzard breath; does not waste turns if MP depleted or silenced
Hydra           150     0       137     120     76      double-turns; same AI as Salamanders except that it's medium fire (~40 MT damage)
Troll King      250     27      135     78      67      may cast Limbo, which is really fucking annoying; regens ~50 HP/round
Green Dragon    120     0       140     75      62      often breathes medium fire; sometimes calls Elysium Bird adds
Barog           93      21      100     57      55      casts Defeat (MT ID) 50% of the time
Putregon        350     0       170     45      43      Dragon-typed, vulnerable to Zombie Slasher; often breathes medium blizzard breath, rarely weak blizzard breath
Lionroar        160     50      145     60      78      casts Bounce and Infermost
Swordoid        170     31      158     72      99      vulnerable to Zombie Slasher; double-turns, may cast Healmore, may crit

King Hydra      550     0       240     50      50      boss, but a simple one; regens ~100 HP/round; double-turns, often breathes medium fire breath
Baramos Bomus   450     255     210     200     72      boss; uses strong fire breath/Explodet, often double-turns and can triple-turn; regens ~50 HP/round
Baramos Gonus   900     0       360     0       0       boss; non-weighted targeting, does nothing but attack; regens ~50 HP/round
Zoma (1)        1023    255     550     300     255     final boss; this is his form before you use the Ball of Light; regens ~100 HP/round
                                                        double-turns, follows strict pattern:
                                                        strong blizzard breath (SBB) -> Snowstorm -> SBB -> Attack -> freezing wave (MT Dispel) -> SBB -> Bounce -> SBB -> repeat
Zoma (2)        1023    255     320     200     80      final boss; this is his form after you use the Ball of Light; regens ~100 HP/round
                                                        double-turns; may use medium or strong blizzard breath, Snowstorm, freezing wave or physical attacks, is not strictly patterned


EDIT: I should note that the average given doesn't adjust for e.g. Swordoids double-turning or anything like that; it could be argued they should have extra weight in the average because of that.  YMMV.
« Last Edit: November 20, 2014, 08:59:59 PM by Reiska »

SnowFire

  • DL
  • Denizen
  • *
  • Posts: 4934
    • View Profile
Re: Dragon Warrior III (NES PCs)
« Reply #7 on: February 28, 2015, 09:06:44 PM »
More DW3 mechanics obscura courtesey of Reiska in chat: Swordedge Armor only reflects basic attacks, and its % chance of proc is the chance of an Expel proc.  As a reminder, Expel roughly equals crappy instant death where you don't get XP or gold, so even worse than ID.

So, at bare minimum, Soldier won't get Swordedge hype vs. bosses & PCs that immune ID, and PCs that resist ID will cause it to have a lower chance of working.  Might also apply to similar "forced run" resistance.

Reiska

  • DL
  • Denizen
  • *
  • Posts: 251
    • View Profile
Re: Dragon Warrior III (NES PCs)
« Reply #8 on: February 01, 2021, 10:27:15 PM »
I don't have the headspace/expertise to figure out how these should work for averages/etc., but here's stats for DQ3 NES bosses. 

Kandar 1
HP: 113-150 (131 average)
ATK: 65
DEF: 42
AGI: 19
Evasion chance: 1/32
Targeting: Unweighted
Resistances: Immune to Beat, Sacrifice, StopSpell, RobMagic, Chaos, Limbo/Slow, Expel; 30% resistance to fire, ice, Sleep, Sap/Defence, and Surround

Actions: Attack (6/8 chance), Critical Hit (1/8 chance), Parry (1/8 chance)

He comes with three Kandar Henchmen for support, detailed below.

Kandar Henchman
HP: 45-60 (53 average)
ATK: 55
DEF: 31
AGI: 20
Evasion chance: 1/64
Targeting: Unweighted
Resistances: Immune to Beat, Sacrifice, StopSpell, RobMagic, Chaos, Limbo/Slow, Expel; 30% resistance to fire, ice, Sleep, Sap/Defence, and Surround

Actions: They don't do anything but attack.

Kandar 2
HP: 300-400 (350 average), regenerates 44-55 per turn
ATK: 89
DEF: 40
AGI: 35
Evasion chance: None
Targeting: Normal
Resistances: Immune to Beat, Sacrifice, Sleep, StopSpell, Surround, RobMagic, Chaos, Limbo/Slow, Expel; 30% resistance to fire, ice, and Sap/Defence

Actions: Attack (63/64 chance), Critical Hit (1/64 chance)

Orochi
HP: 225-300 (263 average), regenerates 90-109 per turn
ATK: 120
DEF: 68
AGI: 50
Evasion chance: None
Targeting: Normal
Resistances: Immune to fire, Beat, Sacrifice, StopSpell, RobMagic, Chaos, Limbo/Slow, Expel; 30% resistance to Sleep, Sap/Defence, and Surround

Actions (2 per turn): Medium Fire Breath (1/2 chance), Attack (1/2 chance)

Note: I've never seen Orochi breathe fire twice in one turn on NES, so I think it may be rigged to not do this.  She can attack for both turns though.  Also, on NES, both Orochi fights are identical.

Boss Troll
HP: 240-320 (280 average), regenerates 90-109 per turn
ATK: 180
DEF: 60
AGI: 80
Evasion chance: None
Targeting: Normal
Resistances: Immune to Beat, Limbo/Slow, Expel; 30% resistance to StopSpell; 70% resistance to fire, Sleep, Sap/Defence, Surround, RobMagic, Chaos

Actions: Follows a strict 8-turn cycle.  He attacks every turn, and always criticals on turns 3 and 6 of the cycle.

Baramos
HP: 675-900 (788 average), regenerates 90-109 per turn
MP: 255
ATK: 220
DEF: 100
AGI: 85
Evasion chance: None
Targeting: Smart
Resistances: Immune to Beat, Sacrifice, RobMagic, Chaos, Limbo/Slow, Expel; 30% resistance to ice, wind, Sap/Defense, Surround; 70% resistance to fire, Sleep, StopSpell

Actions (2 per turn): Follows a strict 8-action cycle of Explodet, Attack, Strong Fire Breath, Blazemost, Limbo, Attack, Explodet, Chaos, then repeats.

Note: Baramos is MP-aware, so if StopSpelled, his pattern becomes Attack, Strong Fire Breath, Attack, then repeat.

King Hydra
HP: 413-550 (482 average), regenerates 90-109 per turn
ATK: 240
DEF: 50
AGI: 50
Evasion chance: None
Targeting: Unweighted
Resistances: Immune to fire, Beat, Sacrifice, StopSpell, Surround, RobMagic, Chaos, Limbo/Slow, Expel; 30% resistance to ice, Sap/Defense; 70% resistance to Sleep

Actions (2 per turn): Attack (5/8 chance), Medium Fire Breath (3/8 chance)

Baramos Bomus
HP: 338-450 (394 average), regenerates 44-55 per turn
MP: 255
ATK: 210
DEF: 200
AGI: 72
Evasion chance: 1/64
Targeting: Smart
Resistances: Immune to Beat, Sacrifice, Sleep, RobMagic, Chaos, Limbo/Slow, Expel; 30% resistance to fire, ice, wind, lightning, Sap/Defence, Surround; 70% resistance to StopSpell

Actions (1-3 per turn): Attack (1/4 chance), Explodet (1/4 chance), Strong Fire Breath (2/4 chance).  He has a 2/8 chance of attacking once, a 3/8 chance of attacking twice, and a 3/8 chance of attacking thrice.

Note: He's MP-aware.

Baramos Gonus
HP: 675-900 (788 average), regenerates 44-55 per turn
ATK: 360
DEF: 0
AGI: 0
Evasion chance: 1/64
Targeting: Unweighted
Resistances: Immune to Beat, Sacrifice, Sleep, StopSpell, Sap/Defence, RobMagic, Chaos, Limbo/Slow, Expel; 70% resistance to Surround

Actions: He does nothing but attack.

Zoma (without Sphere of Light)
HP: 767-1023 (895 average), regenerates 90-109 per turn
MP: 255
ATK: 550
DEF: 300
AGI: 255
Evasion chance: None
Targeting: Smart
Resistances: Immune to ice, Beat, Sacrifice, Sleep, StopSpell, RobMagic, Chaos, Limbo/Slow, Expel; 30% resistance to lightning; 70% resistance to fire/wind

Actions (2 per turn): Follows a strict 8-action cycle of Strong Ice Breath, Snowstorm, Strong Ice Breath, Attack, Disruptive Wave, Strong Ice Breath, Bounce, Strong Ice Breath.

Notes: As in other DQ games, Disruptive Wave removes all buffs from the party.

Zoma (after Sphere of Light)
HP: 767-1023 (895 average), regenerates 90-109 per turn
MP: 255
ATK: 320
DEF: 200
AGI: 80
Evasion chance: None
Targeting: Normal
Resistances: Immune to fire, ice, wind, Beat, Sacrifice, Sleep, StopSpell, Surround, RobMagic, Chaos, Limbo/Slow, Expel

Actions (2 per turn): Attack (3/8 chance), Snowstorm (1/8 chance), Medium Ice Breath (1/8 chance), Strong Ice Breath (1/8 chance), Disruptive Wave (2/8 chance)