In our last devlog, we showed you the arena mode. It is only natural to continue with what is most likely the hidden gem of the Heroes of Might & Magic franchise. That is, of course, the random map generator and the templates it provides. Our approach, while very similar to what you might be used to from past Heroes games, focuses more on the quality of the maps, rather than the quantity. But fear not! There will be a very vast template pool you can choose from in Olden Era! Be advised, this is all very technical and it could change in the future!
Keep in mind, all names are subject to change, and there will be many more templates added! We are planning to create not only very different templates, but also different sizes for each of these templates.
Why do we need a good map gen?
In essence, a good map generator is one of the primary tools used by the community to ensure replayability of the older games in the series such as Heroes 3, Heroes 5 or Heroes 7. We can see that in cases where such a tool was missing, such as Heroes 4 and 6, which are far less popular than the other three and have a much smaller following as a result. Having a well-established and well-adjusted map generator means that players can play the game over and over again without feeling like they’re continuously running into the same routine.
Furthermore, out of all the feedback we received, all of the interviews we conducted, and all the talks we had, we discovered some issues with Heroes3's map generator (and all the other Heroes of Might & Magic games, not just Heroes3): Our rough estimates say about 15% of the matches played are borderline unwinnable due to the randomly generated map being too unfavorable. Some of the points here are:
- Crossroads too far from the starting castle and with uneven distance for all players (e.g. one player has it within the reach of their town, another has to spend 2 turns to get there)
- Impossible to form a unified strike force to eliminate guards because the spawn zone doesn't have enough external dwellings
- The main hero can't be upgraded to a high enough level because not enough experience and attributes are available in the spawn zone
- Roads shifted to the edges of the zone, making significant parts of the area difficult to access due to terrain penalties
- Paths blocked by very strong neutral squads
So what did we do to improve this?
Regarding the direct comparison between Heroes 3 RMG and ours. We generally have a similar feature set (i. e. biomes), and we tried to both make it more robust and expand it. Here are a few examples:
Our RMG:
- Guarantees that paths won't be blocked by a random enemy.
- Provides control over ease of accessibility of different objects. E.g. one could ask for a sawmill to be placed near the starting city, for a road between two cities or for a powerful item tucked away in a corner.
- Has a flexible system for mandatory content. The original has something similar for mines, and they are indeed the most obvious target for such a system — after all, the player needs a sawmill no matter what. But we can and do use it for much more — dwellings, hero stats, visibility — you name it. It is a key instrument for designing fair map templates.
- Ensures that distance between the starting town and the higher value biomes (ex. treasure zones) is kept similar for all players.
- Allows you to select what types of Pandora Boxes can be found in each biome.
- Allows full control over what artifacts can and cannot show up. This also includes spells, units, heroes and skills.
- Lets you disable and enable biome types that will be generated (e.g. no lava)
- Mandatory objects will be customizable. For instance we’ll be able to set a certain amount of dwellings, or mines for a biome.
Technical side of RMG made easy for you!
To start with we have tooling to analyze map templates. It is possible to examine a generated map in much greater detail than from a player’s perspective (for instance to see object value distribution, which objects are random and which are mandatory, exact zone boundaries etc.). More importantly we have tools to analyze the maps statistically. One could simulate a thousand generations and calculate the average number of Dragon Utopias in a specific zone. These tools offer us a much more precise generation than we were used to having in previous heroes installments, which also means that any anomalies you would normally find will be quickly dealt with (e.g. waterfall decoration spam in Conflux biomes or excess amount of singular mountain peaks).
On the variety and control
Let's go with a simple example. Say you want to generate pictures of cats in funny costumes. Your artist drew a bunch of cats as well as hats, mustaches, glasses, and whatnot. You then might just throw a random number of random items on a cat. It will probably look rough — two pairs of glasses, one shoe and a wig peeking through a beret. Unless you are into that sort of look, you establish some rules: either zero or two shoes, no more than one pair of glasses, no tall wigs with berets. But then if you add too many such rules, establishing a rigid cat dress code, you will end up with just a handful of different cats.
What is the point of generating them randomly then? Generating HoMM maps is not that different from this perspective. For every big and little thing you usually have to ask yourself, how random should it be? It goes for zone borders, elevation regions, city placement, road graphs, quantities of resources and strength of the guards. This is what we believe is the key difficulty in any sort of procedural generation. You want things to be varied and diverse, but you also want them to make sense.
These goals are somewhat contradictory, so one has to balance them.
Name of the templates is subject to change
Mandatory content is something that has to appear in a zone no matter what. As in one of the examples above: you want the player to have access to wood early. You can then ask for a mandatory sawmill — and it will appear in the zone no matter what. Better yet, you can ask for that sawmill to be near the player’s castle spawn and the generator will try to oblige. You can also ask for a mandatory object from a list. You could ask for [sawmill, quarry, gold mine], and one of them will appear. On top of that, this list can be expanded — so for example if you want, say, four tier 6 dwellings to show up in the starting area, you can do that. Random content is the rest of the map. We have a bag of things that might appear. And the generator takes things from that bag randomly. This bag can be different for each zone and the items in it may obey certain rules. For instance a treasure zone bag can have a larger percentage of expensive items.
An important part of the design of our generator is flexibility. Usually when we asked the game design specialists, if we want a certain thing to work in one of two ways they said: let's do a customizable switch in the config. For instance one could customize directly in a template the win conditions of the map, whether the player is allowed to hire new heroes, whether or not certain items are allowed to appear. We believe this customizability will make the RMG useful for a single player setting. On top of it all, it is possible to generate a map of any size supported by the game (and even some unsupported).
So how will you play Olden Era?
We will provide many templates to straight up pick from, so you will always be able to discover something new. However, we have our own map editor for you to create and save your own custom maps, with your own requirements! On this though, there will be a separate devlog in the future, as it is still a work in progress, and we are trying to improve on it as much as possible before sharing it entirely!
Speaking of singleplayer, something highly requested is the campaign, which will also have its own separate devlog. We are aiming to give you the first few missions of the campaign, as well as the map editor itself, in EA. So do expect a lot more information about these as we get closer to EA!
Follow news and updates: