Strategic and tactical improvements for HMM3

Hello VCMI community!

First of all, thank you for the great project. I’m suprised you haven’t publicized yourselves better.

I have a significant experience playing HMM3 both single- and multi-player modes. And I would like to share some of my thoughts on what you can do to drastically improve the game, something that can’t be implemented in WoG and thus, it may help you differentiate your project from WoG.

Disclaimer: I don’t intend to barge in on your project and tell you what to do. My intent is not that. I just wanted to share some of my negative experiences with the HMM3 classic that you may consider addressing. I am sure the problems I list below are also very familiar to you as well.

Problem: As game progresses, tactical battles gradually shift toward a bag of standard repeated tricks making the game boring.

Potential solution: Tactical variety can increase 10X by linking tactical rules to the land type where the battle takes place. For example,
(a) battle on grass - wait option becomes disabled, making combat follow HMM2 rules
(b) battle on lava - all unit speed is capped to 6 hexes (slow creatures have no penalty)
© battle in snow - killing the enemy’s commander completes the battle like in chess
(d) battle in swamp - decreases defense of all units to 1 rendering the defense stat useless
(e) etc.

The above will make players learn different tactical skills, allow players to strategically select places for battles, drastically improving strategic planning.

Problem: Abundant resources in the later stages of the game, barren map after all resources have been picked up and creatures defeated, low strategic value and variety.

Potential solution: Introduce building and town upkeep. For example,
(a) construct a buidling in a town -> no upkeep payment needed for that town;
(b) nothing is constructed in a town -> upkeep is subtracted from resources;
© choose not to pay the upkeep for a town -> a random building falls apart, if no buildings left the town goes indepedent.

Resource and creature management: the places where they were originally placed by the creator are marked.
Every week 1%-5% of marked resource places are filled with random resources.
Every week 1%-5% of marked creature places are filled with creatures of the same level.

The above changes will make resource protection a vital part of strategy due to upkeep outflows. An attack on the opponent can be done by cutting their supply lines to start destroying enemy’s buildings. It will also make it less prudent to occupy all small towns with no purpose to build them up due to upkeep cost.

In campaigns the above changes will allow you to return to the previous map to find it nearly virgin. Since no upkeep was paid on that map, the towns would return to their basic state and creatures/resources will re-populate the map. Speedy conquest may be a new campaign strategy if the goal is to return and retain as much of the past empire as possible before it decays.

The campaign concept can be drastically improved due to your changes to the map size. Progressing to another map in the campaign may simply be done by traveling through a special type of portal. The above changes to the resource/creature regeneration will take care of map re-population when heroes return back to the old map.

Problem: Artificial Idiot (AI) is obsessed with resource gathering.

Potential solution:
(a) When AI comes close to the human player, it needs to gather most of its armies into a single big army. If the AI’s army is bigger or about same size it needs to go and attack human player and his closest town, the resource strategy needs to take back seat. AI should avoid splitting its army when it is in the fighting mode.
(b) Presently AI recomputes its path after every step which leads to bugs like coming in and out of town many times. Instead the path needs to be selected based on the target (a resource, a town, etc) and followed through unless a human player comes into its sight.

Hopefully, the above ideas will help you come up with your own ideas too.


The other proposals are just another bunch of scripts that haave been done to WoG already, without any spectacular success. But this intrigues me:

Did you notice AI coming back to town over and over again? I was pretty sure this problem was solved. Entire goal locking mechanism and abstract goals were made just to works as you describe it.

Also, AI does everything but gather resources. Currently it strives only to pick these that block his path to exploration, unless win condition requires them.

I apologize for not making it clear. I was not talking about VCMI’s AI. I was describing the AI of the classic HMM3 as released by NWC/3DO.

So you see, VCAI is completely different and, I believe, quite unique for computer games.

Except AI request, I see nothing there which could improve the game, but only create significant unbalance.
You are aware that there are races which start on grass and lava, are you? So castle and rampart will not be able to use wait feature, while inferno will be stuck with low speed. You are aware that swamp heroes have a specificity, high defense, still you propose to ignore it, so how they play then?

As for commander chess like, in chess the king is not middle board and unprotected, as the commanders start. A commander can’t handle the hit of several high level monsters, so battle will be reduced to who acts first, commander dead, game over.

You propose upkeep for towns not build. So, basically our start resources got eaten before being used for purpose, then when we built everything and we don’t need anymore resources, nothing is payed. Capitalist…

The MoP mod already have town upkeep concept, different from this one.
I think the things Wyrm said are just raw not fully thought examples, and Wyrm want something SIMILAIR not the things ALREADY to be included. It is just concept of the concept.

The MoP ideas give more balance than original game and makes also game much harder. The town upkeep in MoP is based on how Town Inhabitiants like you. Both balanced and reasonable. If Kreegan Player (one which have many of Inferno Towns) conquer some Castle Town, why inhabitants would treat him the same as old owner? And any revolution have medium to high resistance, so simply “recently captured town” logic is even enough.

(by MoP I mean Master of Puppets which is both a modder and his flagship mod).

I personally disagree on MoP issue. He gives from start all spells and goodies to AI, which is nothing but even more unbalance and frustration. This is the easy way. The right way would be to script a better AI, as VCMI does.

And about towns upkeep realism, then it turns into another game. Good player = fast conquer the map, which in MoP isn’t doable, because high upkeep cost. So we have a different game, but then we are stuck in external considerations, like MoP “Heroes”, which is no more Heroes in anything, being in direct concurrence with modern games. People return to Heroes because it IS Heroes, they would probably not care about a weird mod about Heroes, based only on author’s super technical skills and not on community wishes, where we are penalized the better we play. Which explains the tiny audience MoP has.

I believe that any changes in AI behavior must go together with a revised work on random templates as well. Take for example Phoenix or MoP, on most of templates we are easy prey, what are the chances to survive to a nearby AI which starts with all dwellings and spells?

I meant balance features not cheating AI.

MoP’s town upkeep balances the game - it makes longer play and nerfes the bonus of more cities. Of course here strategy of rapid capture the town and flee is not successful, but once we defend the city long enough it turns out that inhabitatnts prefewr you than oponent.

Of course it’s a mod and not all players have to like it. I don’t like upkeeps, be it Phoenix or MoP, but if there would be a prize to balance the loss, I would give a try.

MoP used cheating AI just because WoG AI is too poor to play seriously well, partially because AI has almost no idea about new objects.

I agree, that this changes would make different playstyle of homam and it would be hard to ballance it. I disagree with the idea of different combat features depending on terrain, but it would be interesting, if there was more adventages or disadvantages on various terrain type. For example inferno on snow would have -2 speed, but on sand has +1, or something like that.

But what would be much more interesting for me is more complicated battleground.
-For example village would have palisade (weaker then a fort and creature can attack it, or it can be special building if you don’t have stones to build a fort)

  • or special combat terrain for garrison (gate). For example two towers for shooters and some walls, which would give some advantage to defender. Walls can have holes (like open two hex gate) or something like that.
    -Also there could be a building to improve a castle. Inner wall for shooters and commanders. Or wide walls where you can go with shooters for a shooting bonus, but would be vulnerable by catapult. Or outside palisade.

One more thing about cities, it would be good, if you’re be able to go back behind walls, when you’re defending the city.

Also there could be a special terrain features. For example in swamp, there could be swamp, where you would have -x to speed (and defence). Or in inferno terrain a fire lake, similar to firewall. Or sand terrain would have a quick sand. Or snow would have ice lake which would increase speed but decrease attack and defence or a snowdrift (similar to swamp) or something like that.
Or unit special skills could be used in some terrain. E.g. dendroind couldn’t use their skill in lava terrain, cause their roots would burn or units have a x% bonus to chance to make a natural magic on their home terrain and so on.

I’m h3 player, although I played WoG for some time, I don’t know many mods, so maybe somebody already did this, but it would be good to implement it to VCMI.

And one more thing to combats, I like tactical games and sometimes some creatures has special skill backstabing. When you’re attacking creature and on the next hex is a friendly unit, you got double damage. It’s tactical advantege similar to dragon 2hex attack. So at least rogues should have it (and rogues should also have no enemy retaliation, like in h2)

I can’t find right topics, so I type some of my ideas here.

  • fireball balance
    I think that fireball in h3 is really weak. It should be similar to meteor shower or cost less spellpoints.

  • combat terrain editor
    I know you’re not doing new map and campaign editor (yet :smiley: ). I hope that hota team does and you will implement it to VCMI. I agree with Wyrm, that map of any size is very useful in campaign. But what would be perfect feature for map editor would be special editor of battleground. You would place it on particular place in EM, like event and you would be able to draw your own battleground. It would totaly changed some maps.

  • plannable path for heroes
    it would be really hard to write this script and it’s more like mod. But when you have hero, which is traveling from one dwelling to another and visiting water wheels and windmills it would be perfect if you can plan the path for hole week and hero would repeat this way every week, untill you cancel it. It could be third option for a camp button.

I also recomand to use some features from wog heroes 3 hd (for example give all artefacts and stacks to another hero, or when you’re trying to place last stack to a town one creature remains and so on). It’s really userfriendly.
But be careful about some features. For example I don’t like concept of commanders. It’s good for systems like Age of wonders or Battle for Wesnoth, but in heroes the comander is the hero. And if there is a comander he should be placed in stack or be a hero (like a chess king, but it would completely change a playstyle of battles).

thanks to vcmi team for working on this great game. Great work so far, keep working. (it’s nonsence, but I’m really happy that you changed the color of hex grid, I always hated that green color, and when I found MMArchive, it was first thing I changed)

I’m looking forward to VCMI 1.0, cause current version is not stable and hardly playable, at least in win7 x64. But maybe I just have got a wrong version of wog. I hope you’ll get rid of the wog graphics in last version.

And some questions at last. Do you cooperate with Hota team? Or the WoG core team? Don’t you know what happened to the Bastion project (egypt town)?

And one more thing. I’d appreciate new random map generator. I think that I read somewhere that it’s also one of your plan. I know nothing about templates and I don’t know how works generator from original series. But it would be nice if you could set more things. At least forbid teleporters. Also number of towns and heroes, passability of terrain (open map vs. closed, complicared map), different win and loss conditiion and so on. More things you can customsize the better.

Yep a random map generator with more features would own hard, especially the feature to set the aggressiveness level of monsters, it has always bugged me how the native generator always produced hostile mob.

I believe that, in order to be effective, AI needs serious boost if human/AI game. Of course, I am not talking about the “give 1000 archangels and 99999 resources on day one”, but slight bonuses and specific tweaks which help him to remain competitive.

So far, AI problems (in SOD) are not logistically based, but management related. With several various towns, he is not coded to take choices, take best monsters from and protect main hero. He knows rather well to battle, if given right spells and he knows to expand fast, although he does not take advantage of this later.

I spent about 2 months to analyze its behavior on random maps and tried to find solutions and create a more challenging opponent. I know VCMI will code it from scratch, but still I will share with you what I found and coded for, maybe this can give ideas.

My mod (for ERA 2.4) Better AI simplifies AI expansion and management by tweaking/removing all that it is useless for him.

A bit of what’s inside:

  1. AI experience: based on difficulty selected every AI hero will be granted extra experience every day. On 200% for example, a starting AI hero will become cca level 15 on day 8.

  2. A random primary stat is granted every day.

  3. Every AI hero is granted with all combat skills (archery, offense, armorer) and magic essential skills as earth, air, resistance and intelligence. He receives slow, shield, haste, cure and dispel also because at least he knows how to use them and when.

  4. I noticed that AI lose a valuable time to visit obelisks then he is enough dumb to send his main hero lose endless turns and dig. Therefore in my mod, all obelisks are set to visited for AI.

  5. As AI does not know as human does how to deal with different towns, in Better AI, each first day of the week every town one AI owns will be converted to its starting faction.

  6. Of course, obvious bonuses are granted every week, money, resources and extra creatures, depending on the difficulty level selected.

  7. Every AI hero receives a movement bonus each day, up to 6000 points (again, depends on the difficulty level). This bonus is independent and growths from start, so an AI recruited month 3 for example, will receive full bonus, to help leveling faster.

  8. Obviously, AI heroes are marked and not available to human.

  9. At level 15, each AI receives town portal because he does not know where to search for and which town has better chance to give.

  10. Naturally, mana is fully recovered each day for AI. We know when to stop in town and save time, hard to teach him that, so help it.

  11. Based on number of AI present on map, he has a chance to get a town for free every day 1 of the week (and that town will be automatically converted to his faction).
    For example, in a game Human/1 AI, chance will be 100%, in a game Human/2 AI will be 50% and so on.

  12. There are other tweaks I did, as removing every week joiners from his garrisons, as they slow the game and he does not use them etc.

I think, that even with a better coded AI, still he will need such bonuses to keep the game challenging. So far I am very pleased with results. Of course, in WoG there are other problems as AI not knowing how to use special structures and being overwhelmed at the end by special abilities. But that’s another problem.

And of course, each changes should be parameterizable depending on difficulty. Once a random generator created, maybe will not be very hard to implement a difficulty selection when loading map, proposing several options.

One could say: you want fun game, play human. While this is true, it is also very constraining. Time, energy and full game played in one session is not what every one can afford.

Valery what about Htoseat Multiplayer then?