HoMM3 | Dungeons & Devils 1: A Devilish Plan | Impossible difficulty+ | Slayer scroll bonus | Inferno demon farming
Unprecedented victory
Table of Contents
Intro
In my walkthrough of Long Live the Queen, I found the AI was losing too quickly to fully demonstrate my unorthodox strategies. Therefore I resolved to increase the difficulty beyond Impossible in the next campaign, Dungeons and Devils. By choosing the most useless starting bonus in the game, I achieved Infernal difficulty, and found it satisfyingly spicy!
Meridian's playthrough of Devilish Plan confirms my decision. Meridian blitzed the AI's main town in a week, skipping the map and ending any strategic threat from the AI. There would be no point in beating his time.
The 100 imp starting bonus is simply too strong, and the AI is not programmed to prioritize defending its main town. Presumably it does not ponder construction restrictions.
I did beat the more methodical MasterKD's time, by a meager 22%. More importantly, the devilish difficulty forced me to master demon farming, a counterintuitive mechanic for one accustomed to Castle and Rampart.
[PIC: Witch summoning demon blood.]
[PIC: Aztec human Sacrifice]
More on Inferno mechanics in a future post.
If you've already beaten this scenario on Impossible, you should play it again, but take the Slayer scroll bonus. That will force you to learn Inferno like nothing else. Later scenarios in a campaign obscure the early game mechanics with the overwhelming strength of leveled heroes. This is the map to learn Inferno.
The 6.5h gameplay video, with Inferno soundtrack and no commentary, is on YouTube. Below are my notes.
Starting bonus
100 imps help win adjacent fights, and are easily upgraded.
Armor of Wonder: +1x4 is less useful initially due to lack of mage guild and strong local creep.
Slayer scroll is garbage since dragon kills happen after AI is defeated.
The roadless prairie is the player's main defense against the AI, who might get a dragon from its two Refugee Camps any week. So the Slayer scroll is theoretically useful as insurance, sort of like the overpriced two-year warranty on electronics that cashiers always try to upsell you.
In practice, the AI won't do a level-7 creature rush because it has too many nearby targets. The player will have time to build up enough troops to handle the odd level-7 creature. Thus the 100 imps' faster start is better.
Map
Summary
Middle towns can't build, but have troops to hire.
Hill Fort is strategic, but its guard is strong.
AI has immediate access to the rich roadless prairie, whereas player must clear 2+ roadblocks.
Observatories and wells are key.
AI gets 2 refugee camps but can't built Dragon Cliffs.
Towns
The redwood observatories near each Rampart town are especially important due to the large, roadless map.
The main Rampart town is weaker than the Inferno. Rampart can't build Capitol or Dragon Cliffs.
The other towns are disabled except for a couple buildings each, such as tavern, mage guild, and level 1-2 creature dwellings.
Turtle power
The main Rampart town is weaker than the main Inferno town. Therefore if Tan survives into the late game and prevents Red from killstealing the Queen, Tan wins.
Red can't kill the Queen without visiting the keymaster's tent in Tan's southeast corner. This favors turtling.
Hill Fort
The Hill Fort is strategic. If the AI denies the player access, then player can't upgrade Rampart troops, and must pay for Inferno upgrades, which is crippling. AI can sit in middle Castle and use stables to raid far. Big problem.
However, the player only needs to visit it weekly, and can visit it safely in force.
Might vs Magic
Wells are key. Mages march along the edges following the wells. Mighties fight in the middle grassland.
Mages hold the flanks; Mighties raid the middle.
Dragon Queen
I turtled with cheats and fast-forwarded to see what the AI would do. On m2w4d1 Red defeated the Dragon Queen! Moreover, this counted as victory for Red!
This means that player can open the gate guard and bait AI into taking casualties fighting the dragons, opening the way for player to killsteal or counterattack.
The queen is just a lone Gold Dragon, much easier than the battles with her retinue. An Archdevil does 33-44 damage on a Gold Dragon, with a +3 Attack hero. With Slayer, it's 45-60. The Archdevil wins, either way.
Player's minimum strategic priorities are as follows:
Defend Inferno from AI.
Open the gate guard to dragons.
Kill dragon queen before AI.
If the war with the AI looks lost, player can still take his whole army and attempt regicide.
Level 7 creature recruitment
The main Rampart cannot build Dragon Cliffs, but Inferno can build a Forsaken Palace. Thus player has an advantage in the late game.
However, Rampart has two Refugee Camps that compensate for this. They spawn a random creatures every week, of any level.
Things I've seen:
lol sure enough, Theodorus shows up on m1w2d6 with a few Wyvern Monarchs and a Black Dragon. Great.
m1w3d4 devil spotted, great.
Resources
Resource income is more abundant than gold, assuming one uses Mages to clear creep. This encourages building high-level Mage guilds, and using the wells to capitalize.
Eventually the player reaches a Trading Post, allowing him to sell surplus resources for gold. However, it is on Red's side of the map, so leveling Inferno's mage guild first is strategic.
Inferno saves on magical resources by not upgrading creature dwellings, relying on the Hill Fort instead.
Strategy
Victory condition
The de-facto victory condition is securing both main towns.
This requires either defeating the AI's main heroes or having adequate defenses to deter them, such as a castled week's growth.
The AI does not understand the necessity of holding a main town, so it is easy to trick into garrisoning a crippled middle town until it is too late.
Bad starts
Bad luck in the early game slows tempo. Some examples:
Heroes are all mages, most with useless starting spells. Solution: Prioritize level 1 mage guild.
Southeast Rampart has a strong garrison.
Strong sawmill roadblocks bottleneck wood and delay prairie scouts.
Might vs Magic
In the early game, the problem with the mage guild is that Inferno heroes have such low Knowledge it's hardly worthwhile to give them level 2 spells, or even a collection of level 2 spells. They can just continue casting their level 1 spells occasionally, for minor effect. It is better to improve the army so that their casting can make a difference against the overpowered creep, which is too numerous for level 1-2 hero spell damage to make a difference. A team of Mages levels slowly while sharing an army, and Inferno usually gains Spellpower, not Wisdom.
Thus a main Might hero is much better than a Mage at the start. This is unfortunate, since Inferno's theme is burst Fire Magic damage. The map's wells look like a strategic resource, but their early relevance is greatly reduced by Mage weakness. Fortunately, Heretics start with 1 Attack and Defense. Skip the well creep until the Mage army is strong enough for demon farming.
Keymaster closed
Opening the Dragon Queen's lair requires visiting keymaster tents:
red keymaster and southeast Rampart
southwest Castle and green keymaster
This is a bad idea, because it gives the AI a chance to win by killing the Dragon Queen first. The keymaster tent is irrelevant to defeating the AI in a quick game. Therefore the Dragon Queen victory point should be kept sealed until the AI is defeated.
Homecoming comparison
No blitz like home
In Homecoming, the solution was to blitz. The AI's forces were divided by roadblocks, snow, and long tunnels. Given time, however, the AI's three captive Castle towns would become dangerously powerful.
In A Devilish Plan, blitzing is risky. AI controls most of the map including three secondary towns. AI's territory has no roadblocks or roads, prairie. Player's blitz thus relies on the AI's incompetence to succeed. If AI concentrates its forces, it can counterattack with overwhelming force, either at the invading hero or player's base.
The problem is worse without adequate scouting, but scouts tend to die to AI's minor heroes. So player's Mages should expand along the edges of the map instead, shielded by roadblocks.
The middle towns in Devilish Plan are even more crippled than in Homecoming, rendering a blitz less urgent. Meanwhile, the player has a nearby neutral town, a gold mine and Hill Fort. Player should focus on taking these strategic objectives first.
Once the player has upgraded his army at the Hill Fort, then his main Might hero blitzes, conquering the towns in sequence.
Sea unites, land divides
In Homecoming, the lesson was that sea unites. In Devilish Plan, the lesson is that land divides. That central prairie may look invitingly green, but it is a trackless abyss of wasted movement points. Crossing is best made at the middle stables, making the middle Castle a battleground.
The prairie is player's main defense against AI, which might get a dragon from its two Refugee Camps any week. So the Slayer scroll is theoretically useful as insurance, sort of like the overpriced two-year warranty on electronics that cashiers always try to upsell you.
Operations
Circle the volcano.
Freebies first: learning stone roadblock.
Flag ore pit.
Try southeast Rampart.
Hero chain unified army.
Clear and harvest volcano roadblocks.
Beware strong creep.
Once routing, leave a slot for volunteers.
Prairie blitz
Main Might hero follows this sequence:
southeast Rampart
gold mine (to pay for upgrades)
Hill Fort
demon herd 1
middle Castle
southwest Castle
observatory, northeast Rampart
observatory, main Rampart
middle Castle
Conan - What is Best in Life | YouTube
Defend main Inferno with two strong Mage armies clearing flanks, weekly growth and fortification to Castle.
Try to visit Trading Post to sell surplus resources.
Demon farming well
Main Might army's core is upgraded centaurs, gogs, imps, and elves.
Therefore probably won't demon farm those units, except the trash imps.
Level 1-3 irregulars and Cerberi get farmed. Upgraded dwarves can go either way.
With pikemen and archers from southwest Castle, there are enough irregulars to keep two demon farms busy.
Mages farm near wells to maximize MP spent per fight. No hurry.
Date
DONE run 3
m1
w1
Heroes: Calh, Marius, Edric. Calh was lucky first hero.
Built kennels.
Cleared 4/5 volcano roadblocks.
Took Rampart.
2/3 artifacts useful: Glyph of Gallantry and Still Eye of the Dragon.
w2
New heroes: Yog, Zydar.
Built demon and pit fiend dwellings, and mage guild 1 and marketplace.
Took gold mine.
Visited Warrior's Tomb for Red Dragon Flame Tongue.
Upgraded at Hill Fort.
Demon farm: Recruited 3 pit fiends. Allocated 6 Cerberi, 27 Familiars and irregulars.
Calh keeps Magogs due to specialty.
w3
Took middle Castle.
Took southeast Castle.
Major battle 1.
Took northeast Rampart.
On 135, Calh defeated Red's main hero Gem for 52 Familiars, gaining 3 artifacts. Gem sought the bag of gold artifact near Inferno. Calh just barely reached her from southwest Castle via rally flag. Her morale was -3, and she only cast power magic missile, with 30 MP.
This is why it's important for the main prairie hero to stay near the middle until Inferno's defenses can deter Red's main hero.
This defeat was the beginning of the end for Red. By the end of the week, Tan held all three middle towns, leaving only main Rampart to Red.
Demon farming proved effective with three Pit Lords per Mage. Zydar sent 11 demons to the Hill Fort, and Voy farmed four.
w4
Calh took northeast Rampart then main Rampart on 142.
Demon farmer Zydar reinforced northeast Rampart and then fled north from major army to well.
Red's main army garrisoned at middle Castle, blocking Hill Fort.
Voy demon farmed at southwest Castle, sending back
Red willingly conceded the main Rampart for two crippled middle towns, middle Castle and northeast Rampart. This was a fatal strategic error.
m2
w1
Calh with primary Might army defeated Red's main hero in middle Castle, autocombating for minor losses.
On m2w1d5, Pyre with secondary Might army defeated Red's last town and major hero, autocombating for heavy losses.
Nothing left but mop up of two minor heroes. Game over for Red. No further challenge in the scenario.
Summary
Player capitalized on a lucky start and never looked back. Calh's main Might army followed the recommended path:
southeast Rampart
gold mine
Hill Fort
demon farm 1
took middle Castle
took southwest Castle
major battle 1
took northeast Rampart
took main Rampart
retook middle Castle
Calh visited the middle Castle thrice because of its strategic importance: the stables dominate the prairie, and the Hill Fort dominates the scenario.
Demon farming was highly successful. The two strong Mage armies defeated or deterred minor Red's heroes and harassed major ones. Voy farmed at least 20 demons., Zydar 17. These probably did not reach the main Might armies due to the roadless prairie, but 11 of Zydar's were upgraded and garrisoned to deter invasion of Inferno while Calh took the main Rampart.
Most useless award goes to the Slayer scroll, which collected dust in Zydar's backpack.
DONE run 2
Summary
Tough start. Going well, but taking too long. Learned important demon farming lessons.
DONE w1
Summary
Unlucky start. No suitable Might hero, just two Heretics. Second tavern hero was Charna; I passed in favor of earlier mage guild.
Mage guild spells mostly useless now: Bloodlust (Ash's starting spell), Prot Water, Stoneskin, Dispel, Summon Boat.
Tough pikeman, Halberdier and Marksman roadblocks. Fortunately, the learning stone artifact was good: Quiet Eye of the Dragon.
Got a +1 Knowledge artifact instead of gold from a chest, but it can help the Heretics.
Most importantly, the southeast Rampart garrison was weak, providing gold income.
Bypassing the sawmill might bottleneck wood, but Tan ends the week with 9 wood.
Overall, this is the Infernal difficulty I hoped to show. No elite battle skills used; strategy will suffice.
I accepted Tan's starting weakness and dallied with the 1st hero until reinforcements arrived.
1
1st hero (Ash) took troops, flagged mines, visited learning stone.
2
Ash won tough fight against pikemen for 10 imps.
Collected wood. She's delaying so next hero can reinforce her.
3
Ash visited well despite having Mysticism and 5/10 MP.
Hired hero 2 (Calid) who recruited all troops and visited learning stone.
4
Calid reinforced Ash.
Ash flagged ore pit.
5
Ash took southeast Rampart and recruited all troops.
Ash found Unicorn's Helm (+1Kn) and gave to Calid.
6
Calid visited well.
Ash defeated Halberdier roadblock, taking some imp casualties.
Ash garrisoned at Inferno.
7
Built mage guild. No direct damage spells. One useful now: Stoneskin.
Sawmill guard is lots of Marksmen.
Nearby artifact is Glyph of Gallantry (+1 morale). So I could've hired Charna week 1.
DONE w2
Summary
Main hero now Crag Hack with ranged square and 6Atk 4Def.
Mages supporting and sharing army.
Cleared northwest starter zone to well.
Built Kennels, Demon Gate, Hell Hole.
1
Ash defeated roadblock to mercury lab.
Calid recruited Rampart troops.
2
Hired 3rd hero (Olema), who gave troops to Calid, and took Unicorn Helm.
Ash flagged mercury lab.
3
Ash routed roadblock to gem pond.
Ash flagged gem pond.
Olema visited learning stone.
4
Ash visited mystic garden.
Calid harvested gold.
Olema visited well.
Built kennels.
Hired 4th hero (Crag Hack) who recruited all troop and visited learning stone.
5
Ash defeated walking dead guarding well.
Built demon gate.
Olema garrisoned at Inferno and recruited demons.
6
Olema gave demons to Crag Hack.
Ash gave troops to Crag Hack.
Crag Hack defeated Iron Golems guarding Yawning Dead Shield.
Gaining momentum now! Might hero with 6A4D.
7
Calid routed nearby roadblock and gave troops to Crag Hack.
Olema garrisoned.
Built Hell Hole
w3
Summary
Conscripted dendroid Hill Fort guard, luckily.
Traded heroes at Hill Fort; destroyed medium army for negligible casualties.
Hill Fort upgrading army.
Crag Hack took gold mine, raiding prairie, defeated Clancy's medium army.
Southwest Castle taken but not defended.
Mages demon herding imps. Second herd formed; 2 Pit Lords each.
Built marketplace, mage guild 2.
Momentum is strong. Unless Red's main hero attacks next week, Tan will snowball to victory. I expect the AI will shy from confrontation, lick its wounds, and lose by attrition.
A Red medium hero encroaches from the north, but at most can disrupt northern demon farming.
Crag remains close to Inferno, within range of his supply lines. Support heroes are vulnerable when visiting the Hill Fort and reinforcing Crag.
Crag sweeps the prairie of Red's heroes, who flee or fall. This allows Crag's Familiars to be gradually herded into demons.
Tan's other five heroes are busy herding or upgrading troops, so Crag has no support hero or prairie scout. Crag has a spyglass that helps. He will continue patrolling Tan's central frontier, while Mages methodically clear the flanks.
It's week three and the AI still hasn't folded. This is the difficulty I wanted to see! I get nervous every time I leave a weak hero next to the shroud. Losing Ash with her artifacts made me wince, but that only made avenging and rehiring her more satisfying.
Foolish mortals. Have they learned nothing from Pompeii? Better to kneel for the axe than to choke on ash.
1
Crag Hack routed Marksmen guarding sawmill, and fought prairie roadblock for ~15% level 1 casualties.
Hired 5th hero (Axsis) who gave troops to Olema and visited learning stone.
Olema gave to Crag Hack.
Ash visited mystic garden.
2
Crag Hack conscripted dendroids guarding Hill Fort.
Crag Hack and Ash upgraded troops at Hill Fort.
19 Dendroid Soldiers and Stronghold irregulars hero chained to Axsis.
Red's major hero Synca routed Ash with 27 archers, 42 pikemen, 2 nagas, 8 battle dwarves, 27 gremlins.
Now what?
Push south, that's the safest direction.
Take the gold mine.
Hero chaining covers town defense.
3
Crag Hack routed Synca for 2 Familiars.
Crag Hack cleared Crypt for 1.5k gold.
Recruited two pit fiends and hero chained to Hill Fort, to begin demon herding.
Axsis recruited Rampart troops and visited observatory.
Basically won the scenario, I think, after winning the Hill Fort battle decisively.
Red still has a main hero lurking out there, but his Castle army was decisively defeated, so I have breathing room, which is all I need.
4
with correct demon herding. distribute properly. give morale artifact to mages to help them achieve sufficient army size. actually, if I resume recording from here, I'm too spoiled. Must start a new run, sigh.
old
Crag Hack routed sulfur pit guard and roadblock, gave 18 Familiars to Axsis, visited learning stone, defeated gold mine guard.
Hired 6th hero (Jabarkas), who recruited 2 pit fiends and 2 demons and gave troops to Olema.
Olema gave 2 Pit Lords to Calid.
Demon farming plan:
Stronghold irregulars form south mage army with 2 Pit Lords. Castle troops will replace finite Stronghold troops.
Dendroid Soldiers clear from northeast well towards northeast Rampart, herding lvl 1-3 Inferno or Rampart. Familiars first.
5
old
Calid ferried Stronghold and unupgraded troops from Axsis north.
Olema took unupgraded towards Hill Fort.
Axsis won two battles and herded two demons from 18 Familiars; recruited 14 centaurs.
Crag Hack flagged gold mine.
Jabarkas harvesting near prairie.
Built marketplace.
6
old
Crag Hack defeated Clancy garrisoned at southwest Castle for minor casualties.
Rehired Ash who recruited 2 Pit Fiends and two Demons, took Axsis' unupgraded troops and gave to Calid, and kept upgraded Fortress troops.
Calid gave troops to Olema.
Olema upgraded troops at Hill Fort.
Jabarkas flagged sulfur pit.
7
old
Ash took Pit Lords from Olema.
Crag Hack visited Rally Flag and gave Olema 18 Familiars.
Olema gave Crag Hack upgraded centaurs, elves and demons.
Built mage guild level 2.
Axsis recruited 15 imps from Inferno.
Jabarkas visited windmill and learning stone.
w4
1
Crag Hack took center Castle, defeating two weak heroes.
Hired 7th hero (Calh) and recruited imps and gogs.
Axsis received rest of his 18 imps.
Calid took surplus Hobgoblins from Ash.
Jabarkas visited windmill.
Red:
Main hero invading southeast Castle.
Medium hero raiding north of Inferno.
2
Built mage guild 3.
Calid garrisoned Inferno and recruited demons and hellhounds.
Crag Hack visited stables, fought artifact guard and visited windmill.
Axsis herded guard and visited star axis, recruited free centaurs.
Olema visited learning stone.
Jabarkas visited rally flag.
Ash retreated towards Inferno.
Calh neared Hill Fort.
Plan:
Better to fight Red than creep.
Bring Rampart Mage army to fight Septienna.
Stronghold Mage army can clear "lots of Evil Eyes" creep, but will need reinforcement. Take some Cerberi from Crag Hack.
Crag Hack will visit observatory and take northeast Rampart, keeping Red busy.
3
Crag Hack conscripted 25 dwarf guard.
Olema recruited elves and centaurs and gave to Axsis, taking dendroid herd and visiting star axis.
Jabarkas visited stables and recruited pikemen.
Ash garrisoned Inferno.
Calh gave Crag Hack Magogs and Familiars and took 8 Cerberi.
Calid took unupgraded from Axsis and Inferno towards Hill Fort.
Red:
Northern heroes fled east from Crag Hack.
Other heroes moved towards middle Castle.
4
Crag Hack visited observatory and took northeast Rampart.
Hired 8th hero (Raksha) at northeast Rampart with no troops but Scholarship.
Raksha bought spellbook, learned Magic Arrow from mage guild, took dwarves from Crag Hack, recruited dwarves from dwelling.
Axsis garrisoned Inferno.
Jabarkas upgraded pikemen, gave to Calh, and visited Swan Pond.
Ash took Stronghold herd with Cerberi towards Septienna; army too small though.
Calh took unupgraded towards Hill Fort.
Olema herded artifact guard.
Red:
Took center Castle with strongish hero.
5
Crag Hack defeated pikemen dwelling guard.
Calh upgraded troops, keeping Cerberi and dwarves.
Jabarkas ferried reinforcements to Crag Hack and recruited pikemen.
Ash headed towards Rashka.
Summary
Realized this is taking too long because I didn't allocate enough troops to demon farmers, underestimating the creep. Need three Pit Lords per farm, not two.
DONE run 1
played with map revealed, on Impossible
m1
w1
Summary
Heroes gained learning stone free level.
Used hero chaining to blitz south creep for town, gold, ore and troops.
Recruiting demons.
1
1st hero (Zydar) takes troops, flags sulfur pit, crystal mine.
2
Zydar visits learning stone, clears roadblock.
3
Zydar flags ore pit.
2nd hero (Calh) recruits all imps and gogs and visits learning stone.
4
Calh takes troops from Zydar.
Zydar advances to southeast Rampart.
3rd hero (Calid) visits learning stone.
5
Calid gives Zydar troops.
Calh defeats roadblock to Inferno and gives Zydar army.
Zydar takes southeast Rampart, recruits all elves and some centaurs, visits observatory.
6
Zydar defeats roadblock to grassland.
Calh wins twice to recruit free centaurs.
Inferno builds demon dwelling.
7
Zydar hires rest of centaurs and gives to Calh.
Calh gives troops to Calid.
Calid hires demons and heads towards sawmill.
w2
Summary
Cleared neighboring creep roadblocks.
Built Inferno creature dwellings through pit fiends by now, and mage guild level 2.
No dwellings upgraded. Cleared path to Hill Fort.
One hero harvesting south grasslands.
1
Calid flags sawmill and gives troops to Calh.
Zydar visits learning stone in grassland.
4th hero (Raksha) garrisons troops and visits learning stone.
2
Calh flags mercury lab.
Zydar visits windmill.
3
Calh flags gem pond.
Zydar harvests grassland.
Calid harvests near mercury lab.
4
Rashka recruits free centaurs.
Built pit lord dwelling and mage guild by now.
5
Rashka recruits centaurs and elves, and gives troops to Calh.
Calh returns to Inferno.
6
Built mage guild level 2.
Calh recruits all Inferno troops and gives to Rashka, who defeats west roadblock to grasslands.
Zydar visits warrior's tomb.
Calid visited mystical garden by now.
7
Built kennels.
Raksha defeats artifact guard, hands troops to Calh.
Calh defeats Hill Fort guard.
w3
Summary
Hill Fort upgrades army.
Two Mighties deny Red the center with a strong low-level archery square. Defeated two heroes and conscripted 18 Basilisks.
Trading Post converted surplus resources to gold.
Mages demon herd locally.
Artifact buildup from creep blitz beginning to tell.
Built City Hall and Citadel.
1
Upgraded all troops at Hill Fort.
Calh has tactics, archery, gog specialty. Takes centaurs, gogs, elves, imps.
Rashka takes Pit Lords and Horned Demons
Zydar is routed and rehired.
2
Rashka gives troops to Calid.
Calh takes middle Castle and visits stables.
Zydar recruits free centaurs.
Build marketplace.
3
Rashka dodges Red heroes in grassland, running towards Calh.
Calh conquers mercury lab, conscripts 18 Basilisks.
Zydar takes 19 Centaurs and Pit Lords and demons to demon farm.
Calid gets remaining unupgraded centaurs and elves.
Built Blacksmith.
4
Calid gets demon farming army with 9 Centaurs.
Raksha visits stables.
Calh already visited windmill, conquers gem pond, heads for Trading Post.
Zydar defeats Star Axis guard, gets unupgraded troops including Demons.
Built Citadel.
5
Calid visits well, clears roadblock, heads for Star Axis.
Rashka visits windmill near Hill Fort.
Calh trades surplus resources for gold and conquers sawmill.
Zydar returns to base.
5th hero (Fiona) hires imps and hellhounds to reinforce Calid.
Built City Hall.
6
Calh gives troops to Raksha, who visits Rally Flag and defeats two distant heroes. Raksha has Advanced Logistics.
Calh taking Basilisks to Hill Fort.
Fiona visits Star Axis.
Zydar recruits all he can afford, low level last, and heads to Hill Fort.
7
Calid defeats southmost roadblock to grasslands.
Fiona (with the Necklace of Dragonteeth Zydar looted from tomb) visits well and gets 50 MP.
Zydar upgrades troops at Hill Fort.
Calh gives basilisks to Zydar, and gives Zydar's reinforcement imps, centaurs, elves, gogs to Rashka.
Rashka ends turn at stables.
6th hero (Arlach) hired for his ballista, which goes to Rashka, who also bears Necklace of Swiftness, Ring of Life and Golden Bow.
w4
Summary
Main and northeast Ramparts secured.
Red's invasion of Inferno defeated.
Central Castle lost. Red's main hero dominates center.
Built growth buildings in Inferno to maximize demon farming. Built mage guild 5.
Mages use wells to clear creep at southwest Castle and northeast of Inferno.
Demon herding shortage due to Calh taking imps and centaurs northwest.
1
Calh took main army, conquered southwest gem pond.
Fiona recruited free centaurs.
Zydar took Horned Demons and basilisks towards southwest Castle.
Arlach took Pit Lords and misc.
Calid recruited elves and centaurs.
2
Raksha visited pikeman dwelling and windmill.
Calh conquered sulfur pit.
Zydar visited windmill.
Arlach took Halberdiers from Raksha.
Calid returned to Inferno.
Built mage guild level 3.
3
Calh visited windmill.
Zydar took southwest Castle.
Arlach gave troops to Raksha.
Raksha defeated Fafner with medium army at Inferno's gem pond.
Built mage guild level 4.
4
Calh flagged gem pond, visited Marletto tower.
Built mage guild level 5.
Raksha returns to Inferno.
Zydar recruited archers from dwelling.
Calid heads north to farm demons from Cerberi; inefficient.
5
Calh defeats guard, visits northwest learning stone.
Zydar garrisons Horned Demons to fix morale and heads for monks.
Arlach recruits pikemen from center Castle and runs southwest away from Red.
Calid defeats artifact guard by gem pond.
Rashka recruits almost all, takes troops towards Hill Fort.
7th hero (Lord Haart) heads towards Calid.
6
Calh takes northwest Rampart from token garrison, flags sawmill.
Zydar loses archers defeating guards for well and ore pit.
Arlach visits windmill.
Calid defeats well guard.
Rashka heads towards northeast Rampart and Calh.
Lord Haart gives starting troops for Calid to herd, and takes demons.
8th hero (Olema) hired at northwest Rampart and garrisons imps.
Red's main hero Neela took center Castle.
7
Calh garrisoned in main Rampart to defend against possible hero chain.
Arlach gives Zydar pikemen.
Lord Haart visits mystic garden.
Rashka takes northeast Rampart.
Fiona visits windmill.
Olema visits learning stone.
Built Birthing Pools, Cages and Castle by now.
Built mage guild level 2 in Rampart.
m2
w1
Summary
Built Capitol and more mage guilds.
Hunting Red's minor heroes, and securing the map.
Calh received reinforcements from both towns.
Red's main hero Neela sallied and was routed. Game over.
1
Week of the wight.
Zydar recruits archers.
Calid fights wights.
Fiona visits learning stone.
Calh hires upgraded elves and centaurs, defeats minor Red hero.
Arlach ferries pikemen and demons to Inferno.
Olema garrisons main Rampart against minor Red hero.
2
Zydar and Calid fight wights.
Raksha and Calh recruit beholders from refugee camp and split army. Rashka gets beholders, demons, imps.
Fiona recruits free centaurs.
Lord Haart supports Calid.
Upgraded pegasus dwelling. Olema recruits pegasi and defeats minor Red hero?
3
Red's main hero Neela garrisons at center Castle, blockading Hill Fort.
Raksha trades resources for gold, gives army to Calh, takes beholders back to Rampart.
Minor hero Cuthbert approaches Inferno. Lord Haart defends.
Calh reflags mines and approaches center Castle.
Arlach gives troops to Fiona, who recruits centaurs and elves.
Olema flags sawmill and pursues Red heroes east.
-Built Inferno Capitol and Rampart mage guild 3.
4
Calh visits stables.
Arlach visits learning stone.
Fiona garrisons with reinforcements.
Lord Haart recruits troops and pursues Cuthbert.
Built Order of Fire.
5
Calid pursues minor Red hero.
Olema defeats minor Red hero.
Rashka recruits centaurs and ferries imps.
Lord Haart upgrades troops at Hill Fort, gives to Calh.
Fiona pursues invader Cuthbert.
Red's main hero Neela emerges from center Castle to defeat Lord Haart's minor army.
6
Calh routs Neela's army, which included a Chaos Hydra, with acceptable losses, mainly 80+ Familiars.
Familiars probably should've been farmed into demons.
Red has no more significant army anywhere. Game over.
Competitors
It's the same two competitors again, playing on Impossible. I found one other obscure guy, but he wasn't relevant.
Meridian
Heroes of Might and Magic 3 RoE [075] A Devilish Plan | YouTube
w1
Same starting hero: Calh! What are the odds.
100 imps makes early battles trivial.
Took middle Castle
Recruited two refugee camps, got death knights and pegasi. Lucky.
Took main Rampart from 14 centaurs 8 dwarves and citadel, no hero, on day 7. Autocombatted for minor losses.
Summary
Game is basically over by 117. Meridian blitzed so hard he skipped the map.
There is no point in watching the rest. Obviously the AI can't do anything with the crippled middle towns.
Looks like Meridian took the last town by 131, but it's not like he needed to hurry. Irrelevant metric. Let's call this a week 1 win.
Even though this was a lucky run, the 100 imp starting bonus is strong enough that the blitz probably works reliably for elite players.
Given Meridian's w1 blitz, the victory metric will have to be securely holding both main towns. The AI is retarded and fails to prioritize its main town, so this is pretty easy to accomplish.
Comparison
I took the main Rampart on 142. 7 days vs 23, so I was 16 days slower, or 329%.
Presumably I could've been more aggressive, but I don't like strategies that depend on the assumption that the AI will cooperate, even if it usually does.
MasterKD
HOMM3 - Restoration of Erathia - Impossible - Dungeons and Devils 1 | YouTube
100 imps bonus.
Methodical clear, using few heroes, taking each town.
Took main Rampart on 147.
I was five days (22%) faster at 142, despite useless bonus.
Conclusion
There are no relevant comparisons, since no one was masochistic enough to upload an Impossible victory using the Slayer scroll.
If you disagree with my unorthodox Mighty/Mage hero specialization, prove me wrong by beating my time on Infernal difficulty.
The obvious way to do that is to blitz the main Rampart earlier, since the AI overvalues crippled towns. However, having the same starting bonus will make comparison of early game performance possible.