Objectives and progress

A project needs a justification to exist. Sometimes it is as simple as having fun tinkering and having a few people to embrace it.

Developing AI, which is still a frontier for humankind and the games industry, can't be done with tinkering. You would get nowhere. I love working on AI, pushing the edge, so it is natural for me. But it also means I see something different in this project.

We had a disagreement here. Some people thought the purpose of this project is to curate Heroes V, caring for its lore, polishing the graphics, adding to the gameplay, and topping it off with some great content. This isn't wrong. But for me Heroes V is a strategy game. It has been designed to cater to many different audiences and as such made compromises, campaign missions that are very simple, the development focus on selling content and the game, with AI falling short. No other game in the industry has ever received a major AI overhaul after release, but I am going to change this because I believe Heroes V is worth this effort – as a strategy game. Once this is done, or if we have the resources, we can take care of all the other issues.

Naturally some things go hand in hand. It makes sense to overhaul the vanilla campaigns to showcase what heights a strategy game can reach, or to produce individual maps with the same goal.

But if the discussion descends into whether dwarves must be underground, or whether it is permissible that orcs appear in the first mission of the original Heroes V campaign, or whether a good strategic map layout should be scrapped in favour of telling a scripted story, I cannot follow you.

In principle Nival has gifted us with a map editor that, if you can look past its user-unfriendlyness, is truly an amazing tool. You can shape astonishingly creative landscapes and outstanding strategic challenges, if you know how to. The game mechanics only add to it, you can have snow and lush islands even underground that tell stories to match.

If you see the game only as an immersive experience, I beg you to open your eyes. There is an entire world to discover in the strategic core each map offers and to methodically plan your opponents' defeat. Once you have tasted it, you know that there is an infinitely greater game to be had than merely going from adventure map site to adventure map site. And the best thing is, once you begin to see it, you can pick up the rest while playing. Strategic planning is exceptional fun and begins with surveying the map and learning about its strategic features. To give all interested a point of entry, I will explain more about strategic planning in the next posts.

Coming back to this project's main objective, it is the creation of an AI that transforms Heroes V into a truly dedicated strategy game. Everything else follows from it logically.

We are now nearing the point where we get our first full release. This means the AI has to perform at least to some minimum standard required by a good strategy game, i.e. it must be non-cheating and capable to adapt to different situations. It won't be perfect at start, but it has the foundation in place to be upgraded step by step until it is meeting our goals fully.

It also means that the game has to be reasonably polished. The beta constraints will be removed, scripts will come back, the debug dumping will be stripped etc.

The interface and gameplay will also need polishing. To this end, the guarded treasure vaults and high-level dwellings have passed another iteration cycle and are mostly in their final state now. v3128 has a better diversity of guards, the rewards are balanced downwards to avoid exploits but still provide you with enough compensation to move forward, and the reported bugs have been fixed.

Remember, these fights are meant to be tough. In principle these fights against equally strong heroes give you something to test your mettle. If you are unprepared, you will loose too much of your forces. You will discover that each type of guarded treasure vault offers a unique fight now, against defenders with vastly different playing styles.

As this is an advanced feature, you have to use your wits. The hero opposing you has the same level as you and an equally powerful army. As a human you have the advantage to perform somewhat better in combat. If you combine this with smart skill selection to arrive at a good hero build plus come equipped with artifacts that give you an edge, you should do fine. You may even discover what a difference visiting a fountain of fortune beforehand makes.

If you are relatively new to HoMM, I recommend that you enter the treasure vaults and high-level dwellings only if you have troops to spare.

If you haven't planned your weekend already, the books 1 and 2 of Marzhin's great Legends of the Ancients (LotA) campaign should be playable now with the non-cheating AI v3128. Though they have still to see thorough testing. You find the download link here.