Posts from Creating a random map generator for Heroes of Might and Magic II.

I would like PReDS, Captain, Kama, Brennan, and of course Dust to share their valuable opinions.
Here's my reasoning. In the third game, they play without water. So, maybe we don't need water either. Water, on the one hand, adds variety to the strategy, which seems good, but on the other hand, players of the third game have accumulated vast experience and play without water (and there are special magic and artifacts for this!). The same goes for barriers/tents. It's a very ambiguous thing, especially in generation. Sphinxes are also unnecessary. Also, there shouldn't be burial mounds (where you can hire ghosts). Because finding a burial mound makes the rest of the game 95% meaningless. Now, about teleports. As you know, in the second game (unlike the third), there is a tactical trick called "teleport block." Plus, there is no way to control the teleport with the spacebar. Therefore, I would also cut down on teleports.
I repeat: all of this is just my IMHO.
The opinions of experienced online players are extremely needed.

At Qwerty's invitation, I'm sharing my opinion on this here...

I would leave the water, after all, this isn't the third game, and secondly, an entire class (of which there are only 6) (and not just a couple of heroes, as in the third game) specializes specifically as a seafarer, and this greatly increases the strategic possibilities of the game...

Tents/barriers - I agree, let's get rid of them, along with sphinxes and burial mounds....

But regarding teleports, I would leave them... Because they expand the strategy too much; otherwise, it's just a simple run and fight, which isn't as interesting. Plus, the possibility of opponents clashing on some island early in the game, fighting for gold mines, is also interesting, as you can dive into a teleport and spend almost all your warriors for the gold or even die, or you can go wandering, gather a company, find loot, or something else tasty... The only thing is that something needs to be done about the locking mechanism. Because then whoever is the host wins, or whoever is closer. Mike and I, when we played a map with teleports on an island a long time ago, simply agreed not to lock them (well, regarding the limitation of this, there are many options, or lock them, but not for more than a day, etc.)

Teleporters are just different; they don't work like in Heroes of Might and Magic III. You teleport, and then you fight. But what will be generated on the other end is a question mark. However, they do expand the possibilities, just like water, and I agree with that.

Here's another thought about water... If it generates on an island without a wood mine, it probably results in a flawed map... Or does it mean a mainland surrounded by water, without islands?
In any case, it probably makes sense to consider the experience from the third installment. That is, several templates. By the way, you don't need many of them. Plus, the map size should probably be taken into account. And the possibility of alliances is definitely needed, IMHO.

Is there a well-developed road network, or not? In other words, do all cities need to be connected to each other by roads? Do roads need to lead to all nearby mines, or just to the lumber mill and the mine? And also, is symmetry good or bad? :) Symmetrical maps are easier to create.

Is there a well-developed road network, or not? In other words, do all cities need to be connected to each other by roads? Do roads need to lead to all nearby mines, or just to the lumber mill and the mine? And also, is symmetry good or bad? :) Symmetrical maps are easier to create.

I think roads are necessary, at least as a guide, so you don't get completely lost on the map while searching for enemies. They can connect castles, villages, teleporters, etc. The main thing is that there shouldn't be just one road on the map, leading simply from one castle to another. A network of roads is needed. I'm in favor of symmetrical maps! Not perfect symmetry, of course, but in moderation. The starting area should definitely contain at least 2 mines in easy access: stone and wood. It would be good if each of them had a small pile of the corresponding resource. And all the other mines should be in the nearby area, perhaps near a native village or a neutral city. I also think that, to add a bit of randomness, the map should have more hidden troop locations, dragon utopias, so that you can collect decent neutral armies, like in the random maps of Heroes 3 (Vivienne, Angels). In our case, these could be large clusters of medusa lairs, thief camps, dwarves, nomads, and even genie lamps(!), which would be a real highlight of the randomness of Heroes 2. If possible, it would be nice to add about five interesting objects. To add extra interest to the game. Objects similar to those in Heroes 3, giving high-level magic, some kind of Pandora's boxes, dragon utopias with artifacts and money, etc. These goodies should mainly be generated in separate zones - treasure areas. Where good objects, high-level artifacts, etc., are gathered, with strong guards that are not found in ordinary, starting areas. Ask away, I'm always happy to share my thoughts.

If anyone is interested...
Many people don't play random maps because the maps often turn out to be poorly designed and inconvenient.
However, it seems there's a chance to fix this by "fine-tuning" the generated maps to make them more enjoyable to play.
Specifically, after many years of playing random maps, I've decided to create my own alternative map generator for Heroes 2.
I haven't decided yet whether it will be just an add-on that modifies what the existing editor has already created, or a program that writes the map from scratch (the latter is more difficult for me to imagine at the moment).
In any case, I need people who represent the majority on the forum – idea generators :). I'll handle all the programming myself, as I'm only writing at a dilettante, non-team level. That's not the problem; what's needed is the logic that the standard editor uses to generate maps. And coming up with that myself is time-consuming and difficult. If you're interested, please write.
I can't guarantee that the project will reach a successful conclusion, but if it progresses, it will become an excellent addition to Heroes 2.
I will report on the development stage as it progresses.

Practice has shown the viability of the following system. There are several so-called "templates" - general outlines of maps, water and mountain areas, etc. Then, some generation principles are added - for example, fairy dwellings only appear in neutral areas or near a fairy (i.e., not near a Necropolis). And then there are, for example, 3 templates for Haven, 5 for Academy, 7 for Fortress, and 7 for Stronghold. All resulting maps of the same template will be similar, but will also differ.

Lux in tenebris, how did you manage to get this program working? It immediately crashes on my Windows system. Work on the generator is progressing slowly, but it is. I think it's still better than nothing. I'm not posting screenshots of the generated maps yet because they are still far from being as good as even a regular generator, let alone an advanced one, which many are waiting for. I'm not focusing on testing because I'm spending more time on the (de)generation algorithm right now. I'm providing a rough plan for map creation and typical problems with the standard randomizer below, in case anyone is interested in the process. (Sorry, I couldn't find how to remove the text by category here.) ------ Castles ------ 1) Depending on the player's instructions regarding the location of their castle, a section is determined (9 squares: north-west, north, north-east west, center, east south-west, south, south-east) 2) If Random is selected, the program chooses a random square and a position within it that is not too close to the edge of the map, always with some space on the sides. 3) Subsequent castles should not: - be closer than n positions to each other. Enemy castles are placed at the maximum permissible distance or close to it to avoid "childish frustration." 4) If the map is small or there are many castles close to each other, the generator marks them as "open," which then need to be "closed" with terrain (mountains, trees, etc.). This also means that approaching the castle will take some time. Castle placement options (templates): If there are 2 castles (initially occupied by players), they are placed in opposite squares (north-west and south-east, for example) or simply on opposite sides (north and south, it doesn't matter where exactly). 3 players and 3 castles - options that have the longest path between them. The castles of 2 out of 3 players should not be located too close to each other, although there will also be no symmetry. 4 players can be located in such a way that one pair can be close to each other. In general, there is no strict distribution. Random positions. As the number of players and their castles increases, a landscape labyrinth system needs to be implemented, i.e., the path from castle to castle should be increased by adding detours. ----- Land ----- 1) The land around the castles is placed according to the options set by the players. If not, it is randomized. 2) First, templates are reviewed. Based on the presence of water on the map (land or island system), a group of suitable templates is selected. Everything is clear with land; islands are taken from templates + transitions between land are generated and the space of the islands is expanded/reduced. 3) To avoid glitches with single cells and isolated areas, the generation of land is not done separately, cell by cell, but according to the principle of expanding an area in different directions (the central cell analyzes each cell clockwise around it, and something is placed there). For example, several islands can grow and merge into one large continent. 4) The generation of land and islands is subject to the following rules: - Do not place single, isolated cells (an island of 9 cells in a 3x3 row is not serious). - Ensure that all land is used, provide transitions, and avoid "dead," meaningless zones and long, narrow bridges. - Islands consisting of fewer than n cells are meaningless. Such an island can be connected to the main continent with a "bridge." - Create transitions from Dirt type land to connect "pure" types of terrain and Beach only for the coastline. ----- Paths and Roads ----- 1) The transport network (explicit and implicit roads: asphalt, paths, passages, trails, etc.) is created based on a large assortment of templates, which are shuffled with each other to produce a final mixed result. 2) At the same time: - Each castle (if desired; except for neutral ones) receives access to the main (explicit) road. - The templates use not individual cells-positions of the road, but rather large areas (to avoid inconsistencies). All elements of this "puzzle" are connected to each other by generated transitions. 3) The second option for building a transport network is to generate the road "on the fly," i.e., by analyzing the route of the path based on the 8 cells closest to the center. This was rejected because it does not guarantee harmony and integrity of the map. ----- Landscape ----- 1) Individual, small areas of landscape taken from library templates (a forest of several trees, a rocky ridge, other ensembles, etc.) are placed on the map in accordance with the surrounding area (the principle of analyzing neighboring areas for compatibility of nature). 2) At the same time: - Landscape elements (LE) are placed WHOLE, depending on the adopted transport system, to ensure normal passage through the land. - LE should create convenient hiding places and be as involved as possible in the game. That is, if it is a forest, then there should be a sawmill somewhere in it; if the rocks do nothing, then there should be a mine or fort or any other object under them, etc. ---------- And a couple more flaws that the standard generator allows (minor, but annoying): ---------- 1) Resources and, even worse, artifacts are easily accessible from behind monsters. :) I think everyone has used this, but it's not quite right. Visible solution: place the "monster + treasure" combination with the treasure in an inaccessible area without passing the battle with the monster. Again, try to analyze the landscape and "adapt" the entire treasure to it. At the same time, if the monster is at the beginning of a transition, behind which there is a resource and (possibly) an enemy castle, close the transition on both sides with monsters of equal strength. I think that's fair. 2) Temporary unavailability of essential mines. Perhaps problem number 1, which reduces the beginning of the game to a stupid search for resources with troops of the first levels. Visible solution: analyze the area adjacent to the castle for a suitable place to place a mine or sawmill. At the same time, the path to them should not cause serious difficulties (although, it depends on the difficulty level). The guard of the mine, in the case when the latter is located very close to the castle, is symbolic. Further - a freebie, i.e., without a guard on the medium difficulty level. 3) One or two active elements on the map (buildings, artifacts, monsters, resources, etc.) are sometimes located in dead zones, where it is either impossible to get to, or there is no point in it. There is no guaranteed solution yet. Most likely, somehow analyze the landscape and everything related to the place where we are going to put the element. 4) Uneven distribution of encountered resources and monsters. Lack of a sufficient number of initial resources or some kind of mine in your area. Lack of mines with specific resources needed to build your type of city. Uneven distribution of mines on the map, when one castle has a lot of them, and the rest have very few.

Ty3o-ny3o, what about that... maybe it's just not possible to make it neutral...?
2) At the same time:

- Each castle (if desired; except for neutral ones) gets access to the main (obvious) road.

Good day to all, esteemed fans of Heroes 2! I am one of you and will help in any way I can. What can I do? I am a programmer... And I've even created something for Heroes 2, namely, I developed a program that fully decodes a Heroes map and can save it as an image (when I participated in a tournament on this wonderful resource, it turned out to be very convenient to print the entire map and visually plan a strategy). Well, I'll also mention that I know the formats and internal structure of saves, images, animations, etc. What else? Hmm... I can write part of the code, but for that, we need to decide what the game will be based on. --- P.S. I am a rather busy person and therefore won't be able to devote much time to this project, but I will provide as much help as I can! Feel free to contact me!

In this thread, we discuss ideas and, hopefully, the progress of creating a random map generator for Heroes of Might and Magic II.

It seems like no progress has been made in creating a generator in the past 3 years. If it's not possible to create a generator on a voluntary basis, maybe it's worth hiring a coder? People who are interested in this could chip in 200-300 rubles each, and I don't think that would be too much of a burden. What do you think?