|
| Thread: A HoM&M Mini-Game, Creature Analysis and Creature Charts |  This thread is pages long: 1 2 3 · «PREV |
|
baronus
Legendary Hero
|
posted June 23, 2022 07:00 AM |
|
Edited by baronus at 07:03, 23 Jun 2022.
|
baronus
Yes! Using you tool we can fast find correct number of ballance when all unit is = other unit. So we can have Real unit power!
Its only way to find it because price is not Real value of power. We need test all vs all because if 1200 peasants Beata 50 nagasband 50 nagas Beata 1 azure dragon it doesnt mean that 1200 peasants beats 1 asure dragon. So we need only the same wins/defeats number. Using alghoritm its simple! And you can add variuos stats eg. SoD orc HoTa orc. SoD is a base...
Its never 1 vs 1 so I suposse 2 months growth eg 16 angels. Its real gameplay number in middle game.
|
|
Wub
Responsible
Famous Hero
|
|
posted June 24, 2022 11:20 PM |
|
Edited by Wub at 17:43, 25 Jun 2022.
|
Amazing, LouWeed! My life is so busy these days that staring at HoM&M3 charts (and writing the occasional post) has become perhaps my main pastime activity, so thanks for giving me some quality leisure time in idle moments  .
I think the chart gets by far the most fights right. As you said, you haven’t done kiting yet, which I believe you mean to be synonymous to maneuvering, so a few thoughts on that (which would probably need to be double-checked still):
• As for creatures with No-Retaliation, I think if they have Speed 9 or higher they can indefinitely hit-and-run all human-controlled Speed 4 creatures (as long as positive Morale is not triggered). Against Speed 3 creatures I think they need to have only Speed 6. Against human-controlled Speed 5 creatures indefinite hit-and-run is only possible if the No-Retaliation creature has positive Morale and the Speed 5 creature has not, because then evading until Morale is triggered becomes an option.
• As for Ranged creatures, I would need to look into it more, but probably the exact same holds, except that they need longer than every other turn to get a shot in. Also, Medusa Queens cannot outrun Speed 3 creatures because of their large creature size.
• In some cases, Ranged creatures can get a free shot in every other turn by moving away when blocked by a melee creature even if the Speed difference is smallish to medium. Think Titans versus Chaos Hydras. But in some cases it may occur even if the Speed difference is only 2.
• If I understand correctly, then especially in Horn of the Abyss Ranged creatures without Melee penalty should sometimes engage Ranged creatures with Melee penalty in melee combat. This allows Medusa Queens to take down Master Gremlins, for example (Round 1 Move, Round 2 Attack-Move).
• Regular Melee creatures that can outrun their opponent should Wait-Attack-Move rather than Wait-Attack-Attack. Think Unicorns versus Peasants. That way they maximize their first strikes.
Talking about Unicorns versus Peasants, it strikes me that there are many battles in the chart with as outcome 1.00. But when there is a chance of Morale bonuses and Specialties such as Blinding triggering, you would expect that at least sometimes the weaker stack still wins. Perhaps the number of simulations is too low to capture anomalies.
It would also be good to explicate the rules in the Explanation tab. I personally have no problems with draws due to stand-offs or infinite evading (could even be different for HotA than SoD, given the 100-turn rules) but if you want to force decisive outcomes it would be good to note down the exact decision rule. I do think that slow melee creatures do not deserve a win if they start by Defending, because then their faster opponent could do the same.
Baronus, as for your idea, two months may be quite a long time for a game, but sure, we could in theory take, say, 3 weeks’ worth of Sprites and test how large all other creature stacks would have to be for the Sprites (and all other creatures) to win only 50% of all match-ups. The Mini-Game would then center more around Growth instead of Cost.
In either case, as my first post explains, the Creature Strength charts are not meant to capture Speed and Creature Specialties, as their value is too situational. This allows for calculation of Stack Strength, which is meant merely as preliminary input for, say, the Mini-Game.
____________
|
|
baronus
Legendary Hero
|
|
posted June 25, 2022 06:20 AM |
|
|
|
Since skills are extremely different, we have to accept additional conditions. 4 months of play is rather average. You won't finish the bigger map in 6 weeks. So 2 months is an average gain. It should also be assumed that the hero has 10 attack 10 defense. Units should also be in 3 groups.
|
|
LouWeed
Adventuring Hero
|
|
posted June 28, 2022 03:12 AM |
|
|
All good points wub, those all sound like good test cases once I get the two-dimensional battlefield & kiting working.
With the kiting logic, seems there's two sides to it. You've got your faster creature, and generally that part is simple, they just want to stay as far as possible away from the slower creature. But the slower creature (assuming they can't actually reach the enemy) does not necessarily want to get as close as possible to the faster creature - if the faster creature is in a corner for instance, then the smarter move is for the slower creature to move just far enough that they can reach them on their next turn, otherwise the faster enemy may be able to escape on their turn.
So my working logic at the moment for slower creatures - in the case where they cannot reach an enemy creature that is faster than them - is to move to the cell that maximises the number of cells that both creatures can reach on their next turn. Obviously this is the ideal strategy when the faster creature has no cell it can move to that the slower creature can't reach, and is probably close enough to optimal in other cases too.
baronus said:
Since skills are extremely different, we have to accept additional conditions. 4 months of play is rather average. You won't finish the bigger map in 6 weeks. So 2 months is an average gain. It should also be assumed that the hero has 10 attack 10 defense. Units should also be in 3 groups.
Script can already do attack & defense, and growth to x number of weeks with/without castle and hoard buildings - but splitting the stacks is probably not going to be doable
|
|
baronus
Legendary Hero
|
|
posted June 28, 2022 02:11 PM |
|
|
|
We need number of unit when each has the same wins/loses number! Not 4 weeks growth! Growth tells us nothing! I only told it not to have 2 angels in test. Yes stacks are to hard to simulate so for dragons hydras cerberi and other area units it must be corrected upper.
|
|
LouWeed
Adventuring Hero
|
|
posted June 29, 2022 05:57 AM |
|
|
|
So the idea will be to take, say 2 months' (or whatever) worth of sprites, and then take each other creature in turn, starting with 2 months worth, comparing them, and then increasing the number until a balance is found. The number of sprites used will be static. Probably some sort of binary search to speed things up. Shouldn't be too much trouble, depending how efficient it ends up being
|
|
baronus
Legendary Hero
|
|
posted June 29, 2022 08:52 AM |
|
Edited by baronus at 09:12, 29 Jun 2022.
|
baronus
2 months growth of angels 16 and how many each units is needed to have remis.
For area units we can make another. Eg hydras real test between 7 lvl units 3 stacks vs hydras. And average result as real hydras power. The same magic elementals but lvl 6.
Or another way to make it faster. We can make hydras with no area attack eg. giving hydras stats like chaos hydras in crtrairs and removing area attack. How many hydras vs chaos hydras we need to have remis. In this way we can rate hydras ability. 3 stacks vs 3 stacks.
|
|
Haloneya
Tavern Dweller
|
|
posted July 05, 2022 07:20 PM |
|
|
|
LouWeed
Adventuring Hero
|
|
posted July 06, 2022 05:28 AM |
|
Edited by LouWeed at 01:36, 07 Jul 2022.
|
Sorry yeah the kiting is not working 100% yet
I've added more info to the explanation of the spreadsheet
https://docs.google.com/spreadsheets/d/1dJx321tUnFox8FoC8PksRTuduGC0QaMAsQUwFDCrrzg/edit#gid=1477466469
Examples of kiting working:
Sharpshooter vs Dendroid Soldier:
https://pastebin.com/vNMHW93R
Titan vs 2 Chaos Hydras:
https://pastebin.com/jm86xM7d
3 Marksmen vs Dendroid Guard:
https://pastebin.com/0W7fERtc
Still to do:
-Creatures that occupy 2 hexes
-Melee creatures employing kiting tactics
|
|
baronus
Legendary Hero
|
|
posted July 06, 2022 06:25 AM |
|
|
|
Your work as I see is better and better! Very good! Ballance of units is more important question all strategii games. Do your work help us to find Real power of each creature.
|
|
LouWeed
Adventuring Hero
|
|
posted July 12, 2022 12:45 AM |
|
Edited by LouWeed at 00:45, 12 Jul 2022.
|
Getting there, slowly. I've got a few days off work coming up, maybe I'll finish this off and find the Ultimate Balance. Maybe I'll just play Hades lol who knows
I've finished the logic for shooters, and added the new results
Comparing the old results to the new results here
https://docs.google.com/spreadsheets/d/1dJx321tUnFox8FoC8PksRTuduGC0QaMAsQUwFDCrrzg/edit#gid=1350400864
Shooters up, dawdlers down, but I can also see some changes that don't make a lot of sense - scorpicores, thunderbirds dread knights, harpy hags have dropped a lot. Will need to sort this out.
Mighty Gorgons have gone up because I previously had the kill formula wrong. Sharpshooters haven't gone up that much, probably because I added logic so the other shooters will try moving closer.
Anyway still needs some work
|
|
Wub
Responsible
Famous Hero
|
posted July 15, 2022 11:17 PM |
|
Edited by Wub at 23:23, 15 Jul 2022.
|
Great progress, LouWeed, sorry for not being able to check in for a while – enduring kind of a double-shift-little-sleep routine these weeks…
I agree that to try and counteract kiting, slow creatures should want to approach fast no-retaliation creatures and fast shooters no closer than necessary. My logic when stating that Arch Devils cannot outrun human-controlled living creatures with Speed 5 was based on exactly that idea.
Anyway, I think you have empirically demonstrated my preliminary claim that a single-hex Speed 9 shooter can outrun a single-hex Speed 4 walker. Same for Speed 6 versus Speed 3. Impressive!
Nice work also on the (rather close!) Titan vs Chaos Hydras battle. In all these cases the maneuvering is rather precise, so not straightforward to script, I imagine. I noticed only one tiny improvement: in the penultimate round, the Titan should finish off the remaining Chaos Hydra with a melee attack, because by continuing its kiting, it risks losing the battle to a Morale bonus from the Chaos Hydras.
I looked also at the scripting explanation in the spreadsheet. Sophisticated stuff already, even without kiting. The interaction between this logic and the new kiting logic may be the most complicated yet.
Take said example of Zealots vs Demons. The script now says that Demons should “move to just outside the enemy’s range penalty distance” and Zealots should then “use ranged attack against enemy”. But instead the Zealots can walk to the corner, where they are now perfectly positioned for the same type of kiting as the Titan vs Chaos Hydra example, while the Demons still need 3 turns to reach them.
Creature Specialties interacting with the script are also quite challenging, I believe. Going back to the Vampire Lords example, If I understand battle-to-the-death logic correctly, then Vampire Lords should use a different logic against slower, living, non-ranged creatures they cannot outmaneuver: After Waiting and Attacking, they should not Attack again but Wait, so they can profit from their Life Drain specialty more.
Another example may be Basilisks’ Petrify ability, which in some cases may lead them to Wait (or even move away) until it has faded, because their attack would otherwise do only half damage. But very slick work on ranged creatures trying to counter Sharpshooters’ specialty by moving closer .
Good to see that Mighty Gorgons now beat Firebirds. I recall that in my tests they did better against other high-level or even mid-level creatures too. Going back to your very first post in this thread, perhaps I oversimplified the formula by just calculating a Death Stare rate equal to 10% the number of Mighty Gorgons (rounded down). In that case they should at least beat Behemoths, too, even when allowing them one first strike.
One final thought: the Mini-Game consists only of final battles, but in the actual game you want to preserve your army as much as possible for future battles. Akin to your ‘Score’ statistic, it could be possible to calculate how many creatures are left after each battle, and make a ranking based on that.
My hunch is that according to this perhaps more realistic sccoring system Sprites would no longer top the chart. Instead, Vampire Lords would, because if they win, they often win without losses. If so, then the results would support our intuition that Vampire Lords are quite possibly the most overpowered creature in HoM&M3 .
____________
|
|
LouWeed
Adventuring Hero
|
|
posted July 28, 2022 07:10 AM |
|
|
I've updated the spreadsheet with the final results:
https://docs.google.com/spreadsheets/d/1dJx321tUnFox8FoC8PksRTuduGC0QaMAsQUwFDCrrzg/edit#gid=1477466469
The explanation tab contains the full battle logic. Still room for improvement, but it's close enough
Apart from tweaking a lot of things, the major changes I made to the script are as follows:
-Morale is neutralised where possible. So if both creatures would normally have +1 morale, I make them fight with 0 morale. If one creature would have +2 morale (angels), and the other would have +1, then the angel has +1, and the other has 0. Versus creatures that naturally have no morale, regular creatures still have +1. I made this change to make it more efficient, but also to reduce the wild variations in battle outcomes.
-Comparison logic is better. When comparing two creatures, the first battle is a regular battle with randomised damage, luck, morale, spells etc. The second battle is optimised so that everything goes the way of the loser of that initial battle: their damage is maximised, the loser's damage is minimised etc. If the loser still loses, then the comparison is finished, and only the results from the first battle are used. Otherwise at least 10 more regular battles are
used to determine which creature wins. Previously this was at least two regular battles.
I added a column to the spreadsheet for the each creature's average remaining # of units. Obviously this 0 whenever the enemy loses, so I added an additional column giving the average # of units in the case of a win. Vampire Lords do score very highly.
I haven't found the "ultimate balance" yet, but I will. Just need to find an efficient algorithm for it.
|
|
Wub
Responsible
Famous Hero
|
posted July 30, 2022 11:19 PM |
|
Edited by Wub at 19:21, 31 Jul 2022.
|
Formidable stuff, LouWeed, you have truly outdone yourself! And the scripting improvements you outline make sense to me. Some highlights:
• Separate Creatures, Results and Analysis charts for HotA, SoD and RoE. Awesome! I am going to pay special attention to especially the SoD charts, of course.
• Data for how many creatures are left on average. Nice to see that when Vampire Lords win (which happens often), they do so with fewer losses than any other creature.
• A Mighty Gorgon Death Stare chart, very considerate . I remember now, each Mighty Gorgon has a separate 10% kill chance, with maximum 1 kill per 10 Mighty Gorgons. So my manual calculations were a reasonable proxy, but your script undoubtedly more exact.
• Kiting works well! Very satisfying to see those Stone Golems, which topped my Chart 4, where they belong – somewhere in the middle. And Arch Devils now beat Iron Golems due to positive Moral, very nice! Also, Peasants lose to (War) Unicorns, as they should.
• A plausible Top 20 – with no less than 6 No-Retaliation creatures. Zealots make it to Place 20, as the only shooter, and beat Demons due to good kiting. Nice to see Mighty Gorgons beat Behemoths, but lose to Cerberi – a battle I had not even simulated.
• Good simulation of RNG, allowing creatures with RNG-based Specialties to occasionally achieve upset victories, for example Scorpicores vs. Efreeti. And you seemed to have figured out the Enchanters’ spellcasting, after all!
Assuming that is okay to you, I have linked to your charts in my initial post, so they hopefully get more attention than they otherwise would, buried somewhere on page 3.
As you said, it’s close enough, and I agree, these would be fabulous final results already. Should you still be interested in improvement, here are a few thoughts:
• Not only Arch Devils, but also Devils are fast enough to beat Iron Golems. I think this refers to line 104 in your script. When cornered by Iron Golems, Devils should not move to a corner, but to a hexagon that is (a) out of reach of the Iron Golems and (b) no more than 11 hexes away from the diagonal corner, so they can reach this corner in case of a Moral bonus.
• Faerie Dragons should not end last on the list. They can kite successfully against creatures with Speed 3 and 4 (but not 5). EDIT: On second thought, turns out that even after casting all five of their spells, Faerie Dragons still lose almost any battle, so it seems your chart is correct after all . EDIT 2: Actually, I think Faerie Dragons don't even beat Enchanters 100% of the time, as the chart says. They really need good spells for that.
• Pixies should not lose to Behemoths. In fact, if Behemoths do not defend, Pixies win easily. Not sure what goes wrong here, as Pixies, too, are allowed to defend (line 142). Perhaps the Behemoth’s temporary Defence bonus is not factored into the preliminary Battle-to-the-Death simulation.
• It would still be nice if the rules regarding stalemates (or for that matter infinite evading and battlefield obstacles) were clarified in the Explanations tab. I believe you stated you force creatures to break the stalemate, but not sure if it is alternated which creatures have to do this, and how this is done (e.g., stop defending, move forward) in how many simulations.
• I like the Upgrade statistics in the Analysis charts. I would say the column showing increases in number of Wins is actually a better indicator for Value than the percentage increase in number of Wins. Going from 0 to 1 win is an infinite increase percentage-wise, but less impressive than going from 110 to 130 wins. Increases in (a) rank order in the list and (b) “Any % Remaining’ could also be considered as a proxy for Upgrade Value.
I might come up with other ideas, but I first have a lot of entertaining chart staring to do .
____________
|
|
LouWeed
Adventuring Hero
|
|
posted August 03, 2022 06:50 AM |
|
|
Faerie Dragons & Enchanters: I never bothered getting the simulation right for Faerie Dragons or, more importantly, Enchanters. When I say simulation I mean the AI part of the script that relates to decision making - both the Enchanters/Faerie Dragons and the creatures they're fighting should make decisions based on what the spellcaster is most likely to cast - but they just randomly pick a spell and base their decision based on that one spell - so the results for those two creatures should be taken with a grain of salt.
Faerie Dragons, I just didn't care since they have zero chance of doing very well even if it was done properly. It would be interesting to see how enchanters do if the only spell they ever cast is "Slow"...
I've found out why Faerie Dragons beat enchanters - it's because the Faerie Dragon's spell power never gets recalculated - so all 5 spells are cast at spell power 83*5 = 415. With this bug fixed, Faerie Dragons lose to everyone, 100% of the time, with the exception of Rust Dragons - Faeries win 60% - and Crystal Dragons -Crystals win 90%. They may not be as bad as this suggests, as the logic for Faerie Dragons is to cast a spell if at all possible, if the enemy is not immune. They might do better to attack sometimes, or at least move away from their attacker.
Pixies Vs Behemoths: it's quite close. I ran 1000 battles, pixies win 544 (0 draws). Looks like it has to do with damage variation in the pixie's first strike - here's the logs from two battles, first one pixies win, second one they lose:
https://pastebin.com/4vnevUjB
https://pastebin.com/baTiVP5j
So first round both creatures move forward so that there's 7 cells between them - close enough for the pixies, but not the Behemoths. Second round, pixies wait, Behemoths estimate that the pixies will win a fight to the death (based on average damage) if they move forward, so they defend. The pixies then estimate that they would win a fight to the death if they attack, even with the Behemoths defending (again based on average damage).
But the damage range is huge - anywhere from 137 to 275 behemoths will be slain in that first hit. So in the first battle (the one that they win), the pixies do more damage than expected, and the 29,514 pixies remaining at the start of round 3 can easily beat the remaining 446 behemoths. But in the other battle, the 27,942 pixies are at a disadvantage against 473 behemoths.
Stalemates: I know that some battles result in stalemates because both creatures determine that defending is the only action that will not result in defeat. I don't why this didn't happen once in 1000 pixie v behemoth battles (see above), I think it's more likely in 1v1 battles for some reason.
I did put logic in explicitly to deal with melee creatures that have the same speed - one of them has to move into attack range of the other and lose the advantage.
Anyway, after 3 rounds of neither creature attacking, a draw is declared. The one exception to this is if one of the creatures is kiting the other - in this case 20 rounds are allowed to pass before a draw is declared.
I'll add the stalemate logic to the Explanations tab
|
|
NimoStar
Responsible
Legendary Hero
Modding the Unmoddable
|
|
posted August 03, 2022 02:42 PM |
|
|
LouWeed said:
Getting there, slowly. I've got a few days off work coming up, maybe I'll finish this off and find the Ultimate Balance. Maybe I'll just play Hades lol who knows
I've finished the logic for shooters, and added the new results
Comparing the old results to the new results here
https://docs.google.com/spreadsheets/d/1dJx321tUnFox8FoC8PksRTuduGC0QaMAsQUwFDCrrzg/edit#gid=1350400864
Shooters up, dawdlers down, but I can also see some changes that don't make a lot of sense - scorpicores, thunderbirds dread knights, harpy hags have dropped a lot. Will need to sort this out.
Mighty Gorgons have gone up because I previously had the kill formula wrong. Sharpshooters haven't gone up that much, probably because I added logic so the other shooters will try moving closer.
Anyway still needs some work
Will you take into account the Dendroid Guard and Dendroid Soldier ability in kiting/melee kiting?
Once the other creature is embrambled, it can no longer kite.
____________
Never changing = never improving
|
|
LouWeed
Adventuring Hero
|
|
posted August 03, 2022 11:28 PM |
|
|
Dendroid's binding is accounted for. You can kind of tell, looking at the results:
https://docs.google.com/spreadsheets/d/1dJx321tUnFox8FoC8PksRTuduGC0QaMAsQUwFDCrrzg/edit#gid=1224542768
It's almost exclusively shooters and no-retaliation creatures that can beat dendroids
|
|
Wub
Responsible
Famous Hero
|
|
posted August 05, 2022 10:13 PM |
|
Edited by Wub at 21:40, 18 Sep 2022.
|
I agree, LouWeed, focusing more on Faerie Dragons will not affect who will top the chart, but based on your ideas I looked into it further, and it is actually hilarious how hopeless they are.
Mixing in melee attacks doesn’t work for Faerie Dragons: even Zombies do over 5 times more damage. In fact, Faerie Dragons even lose a melee showdown against Marksmen and Grand Elves. Just blocking them to follow up with casting spells also doesn’t work: Marksmen just move away until they get Morale.
Still, I found there are four match-ups that Faerie Dragons win 100% of the time. Usually, their problem is that even if all of them can cast all 5 of their spells, they still just don’t whittle the opponent down enough to survive a subsequent melee showdown. That is, unless their opponent has vulnerabilities. Indeed, with endless kiting, they can beat Earth Elementals, Magma Elementals, Fire Elementals and Air Elementals.
Magma Elementals lose because, if I remember correctly, they cannot save themselves by casting Protection from Earth on themselves .
The battle against Air Elementals takes around 500 turns, though, because Faerie Dragons have to wait for positive Morale to coincide with having a Lightning Bolt ready, and this for 5 times  .
That also means your now added stalemate logic (thanks!) will usually work well, but the above battle will be an exception.
In line with what you wrote earlier, I also believe that when a creature keeps endlessly evading a stronger opponent, you count this as a loss, rather than a draw. That is an option, of course. It may not always be clear which creature is trying to avoid the loss, though. Vampire Lords may be endlessly evading Iron Golems, but Iron Golems may also be frantically hanging on to avoid a no-retaliation attack
Anyway, nice to see that the script simulates the Pixie vs Behemoth battle well, and thus that it is just a matter of calculating enough runs to arrive at a reliable victory percentage. By the way, it is never a draw because once Pixies and Behemoths are toe-to-toe, unlike in your earlier one-on-one battles, a first strike reduces the enemy damage more than defending does.
I am thinking to post in this thread a picture of your chart with the final rankings. Basically like my Chart 4, but then proper  . Unless you object (also understandable), I could manually change a few hard-to-simulate outcomes as follows:
• Iron Golems lose to Devils (requires kiting and waiting for positive Morale)
• Faerie Dragons beat Fire, Air, Earth and Magma Elementals
• Pixies beat Behemoths
My chart staring did not yield other discrepancies. This is partly because I eyeballed most outcomes rather than actually simulating them, but when I did simulate surprising outcomes, you chart was right all the time.
The only more fundamental improvement could be said idea to let fast shooters without melee penalty engage slower shooters with melee penalty in melee combat (e.g., Medusa Queens vs Master Gremlins). But that probably only works for HotA, which I haven’t played yet, so all good .
____________
|
|
Wub
Responsible
Famous Hero
|
|
posted September 23, 2022 09:14 PM |
|
Edited by Wub at 21:15, 09 Oct 2022.
|
Running double shifts made it again take longer than hoped, but I have now also created a chart of LouWeed’s amazing scripting. I made a few small adjustments:
• Pixies beat Behemoths, Devils beat Iron Golems, Faerie Dragons beat Fire/Air/Earth/Magma Elementals (see above).
• Faerie Dragons lose to Enchanters and Crystal Dragons and beat Rust Dragons only 60% of the time (see above).
• Faerie Dragons beat Wraiths. This one actually takes a few thousand (!) turns: kite until Meteor Shower and positive morale coincide (happens about once every 480 turns). After 4 Meteor Showers, cast Chain Lightning (another 480 turns). The Faerie Dragons are then just strong enough to beat the remaining Wraiths in melee if they kite until positive morale allows for hit-and-running.
• Arch Devils beat Black Knights and Psychic Elementals. This also requires kiting until positive morale allows for a hit-and-run.
So, without further ado (also posted now in opening post):
Chart 5: Mini-Game Rankings
Interestingly, because with these final adjustments Faerie Dragons win some more battles and Rust Dragons no longer always lose against Faerie Dragons, we now have a new tailender: Mummies! I think we can all have peace with that. Here are a few more musings:
Air Elementals best RoE Elemental
Also interesting to see is how Water Elementals, Fire Elementals and Earth Elementals all clump together in the tail. Of course Fire Elementals are still useful for Armageddon shenanigans, but despite being both faster and stronger than Water Elementals, they are not more cost-efficient. Earth Elementals are quite possibly the worst creature in the game: a hopelessly overpriced Level 5 creature with stats worse than any Level 4 creature.
Thus, Air Elementals are really the only of the traditional four Elementals with a somewhat acceptable cost-efficiency (see also Chart 3).
Storm Elementals are half-decent, too, though are ranked a bit too high in Chart 5, because unlike what the manual says they actually have a melee penalty, so should end up somewhere in the third column, below Cyclops Kings. For the same reason, Ice Elementals should be ranked even lower, too.
Unupgraded Level 2 creatures suck
It’s true, all unupgraded Level 2 creatures are at the bottom half of Chart 5. Probably only Rogues are somewhat decent in view of their Specialty. As Chart 3 also shows, Level 1 units (but creatures from other levels too) are substantially more cost-efficient than unupgraded Level 2’s (perhaps with the exception of Imps, Skeletons and Troglodytes, but they have, respectively, Demon Farming, Necromancy and high numbers from Tavern heroes going for them)..
To illustrate, Centaurs are more cost-efficient than Wolf Riders, Boars, Rogues, Harpies and Stone Gargoyles. Pikemen, Gremlins and Gnolls are more cost-efficient (and faster!) than Dwarves and Walking Dead. And Master Gremlins are more cost-efficient (and also faster) than Archers, Gogs and Lizardmen.
Only when upgraded, some Level 2 creatures have something going for them. Not Battle Dwarves and Zombies, though: Halberdiers are still far more cost-efficient.
Horde buildings are not really looking good, either
Of course, the idea to spend a turn and at least 1000 Gold to construct a building that only the week after provides, in most cases, just a few extra Level 1 creatures is often not particularly attractive to begin with. But my point is that many Horde buildings seem designed not to provide a bonus, but to offset a penalty.
To elaborate, Pikeman and Centaurs – two of the strongest Level 1 creatures – have no Horde building available, but already have a high weekly base growth of 14. Skeletons and Gnolls are both not only much weaker, but also have a base growth of only 12. Troglodytes, Imps and Goblins are weaker still, so their growth of 14 or 15 is still measly. Thus, while Horde buildings increase the weekly stack strength of Skeletons, Gnolls, Troglodytes, Imps and Goblins, this is more of an equalizer (if even that) than a bonus.
Same for Hell Hounds, which have the lowest growth of all Level 3 units. Dendroid Guards are stat-wise more of a Level 4 than Level 5 unit (just compare them with Gorgons, to see what a proper Level 5 tank unit looks like), so also need the Dendroid Saplings to make up for this.
Stone Gargoyles and Dwarves may be an exception, but as noted they are not very cost-efficient, so getting a few additional, expensive Level 2 units is still not a great deal. In that sense, the Griffin Horde building seems the least unattractive as it provides a growth bonus (rather than offsets a growth penalty) for a not overly cost-inefficient unit. Perhaps this is why it is the only building in the game that has a higher level building as a prerequisite.
The above Horde building logic is no longer valid in expansions. Sprites are super cost efficient and already have good growth without Horde building. The same is true for Ayssids.
Logic of speed of flying creatures
Talking about expansions changing Restoration-of-Erathia logic: Have you noticed that upgrading non-level 7 flyers increases their Speed according to a set pattern?:
• Speed 5 Flyers (Wights) become Speed 7 Flyers (Wraiths)
• Speed 6 Flyers (Stone Gargoyles, Griffins, Harpies, Vampires) become Speed 9 flyers (Obsidian Gargoyles, Royal Griffins, Harpy Hags, Vampire Lords)
• Speed 7 Flyers (Wyverns, Manticores, Genies, Rocs) become Speed 11 Flyers (Wyvern Monarchs, Scorpicores, Master Genies, Thunderbirds)
• Speed 8 Flyers (Pegasi) become Speed 12 Flyers (Silver Pegasi)
• Speed 9 Flyers (Serpent Flies, Efreeti) become Speed 13 (Dragon Flies, Efreet Sultans).
Expansion creatures are again an exception: Pixies have Speed 7, but Sprites have Speed 9, not Speed 11. Similarly, Stormbirds have Speed 9, but Ayssids have Speed 11, not Speed 13.
That’s it for now. I am still musing a bit about Creature Speed. If there would be some community consensus about how useful Speed is (using perhaps Chart 3 as support), then it would be possible to arrive at approximate values for Speed-Adjusted Creature Strength.
For example, how cost-efficient should a Speed 5 walker without Creature Specialties be to be neither underpriced nor overpriced? Probably more cost-efficient than Imps, but less cost-efficient than Demons, so something like a Swordsman maybe (so strength equivalent of 3311 Demons for 1 million Gold)? Should a Speed 4 walker be about as cost-efficient as an Ogre (strength equivalent of 3779 Demons for 1 million Gold)? And a Speed 3 walker a bit stronger than Peasants (strength of 4368 Demons for 1 million Gold)? It would then be possible to compare the cost-efficiency of Creatures with different Speeds, including unupgraded units vs their upgrades. Anyway, just something to ponder .
____________
|
| |
|
|
