Connection ErrorBad RequestThis page went staleSession ExpiredSign in requiredForbiddenNot FoundToo Many RequestsServer ErrorBad GatewayService UnavailableGateway TimeoutHTTP Version Not Supported
Session Expired
Your session has expired. Please log in to continue.
Sign in required
You need an account to do that. Sign in or create one — it only takes a minute.
This page went stale
Your security token was rejected, usually because the page sat open too long or you signed in elsewhere. Nothing was saved. Reload the page and try again.
Request Failed
We encountered an issue while processing your request.
Please check your internet connection and try again.
>therefore, at the moment, I honestly can't imagine how and in what form this will turn out The AI already receives enough bonuses on the strategic map, depending on the difficulty level.
Well, well, it often runs back and forth, builds haphazardly, and not according to the situation and the map. It also divides troops between heroes not very wisely, so if you're going to improve it, you need to improve everything, systematically and gradually, but absolutely everything.
I think that if you stop controlling hundreds of zombies with a single fake hero, teach the AI to use dispel during battles, utilize defensive abilities, avoid unnecessary risks, and manage resource stacks efficiently, then yes, the AI will become a bit more challenging, especially if it can "deal more damage."
It's unlikely that this will significantly help the AI player (if it continues to play with only 1-2 heroes without using hero combinations), but it will make fighting neutral creatures much more difficult (as it used to be).
And I overlooked the most important thing: the development strategy. The AI plays very poorly in this regard, building some nonsense and then heading to the Capitol (thereby wasting a lot of precious time). In other words, building cavalry in the city/castle on the first day is almost an impossible task for it. For humans, it's the standard approach. And if it's Jebus, then it's angels, no matter what.
Doomsday The main thing is, if you start, finish it. At least create a patch for the battles. I sincerely wish you success!
(Dirty, could you move the topic to the Heroes of Might and Magic III online section? Only 3.5 people will read it here, and you know that yourself)
Сейчас балуюсь этим https://vcmi.eu/download/ За то что нейтралы не бегают за единичкой, для меня это уже стоит того
On the one hand, the idea is a priori unpromising.
On the other hand, there is an opportunity to structure and comprehend the system of priorities and values for development, at least for oneself. And then try to transfer it to an AI algorithm.
In my opinion, 3 directions for rationalizing companion behavior:
1. Battle tactics.
2. System for controlling available territory and objects of interest on the map (chains of heroes, army movement, building a sequence for capturing objects with the right hero) and increasing the zone and quality of control.
3. Priorities in developing castles depending on point 2 (for example, you need to build a little bit of meat on level 1 to break the defenses on level 2, etc.).
.... шло время, а ситуация не менялась. Мясо продолжало визжать и пищать и главное НИЧЕГО не делать!!!
In my opinion, what's the point of fixing it? The goal of the computer in the game is to give you the opportunity to understand the game and hone your skills. But you become a master by playing with other people. Besides, the game hasn't become so iconic that people would keep coming back to it.
The increased intelligence of neutral creatures will lead to a lower threshold for the profitability of battles; reliably winning a battle with acceptable losses will only be possible if there is an overwhelming advantage in forces. The game will become several times longer. Battles with neutral creatures will be reduced to exploiting a few "foolproof" methods of overwhelming the enemy with superior forces and standard safeguards against surprises from intelligent neutral creatures. The value of a single turn will sharply decrease (since the player already has a significant advantage). No one will be interested in such "tactics," repeated countless times; in practice, everyone will use the auto-battle feature (and the more intelligent the neutrals, the more often they will use it), as it also saves time. This type of gameplay (if someone enjoys it) can already be achieved by simply agreeing to play with mandatory auto-battles. AI upgrades are usually done to make the computer a more interesting opponent, not to change the game's profile. To achieve this, it is not necessary to change the algorithms of neutral creatures, but rather to train the computer to deal with them as effectively as a human player, using feints, removing guards, employing advanced strategies for capturing dungeons, and so on. And it should not be afraid to attack, and somehow build/buy/upgrade/combine troops more intelligently. If it is not possible to teach it chains of actions and strategies, then at least it can be taught to select targets on the map based on the build order, rather than grabbing everything that is poorly defended; this can be taught.
Firstly, the number of neutral creatures can be reduced.
Secondly, neutral armies consist of one or one and a half types of troops (one and a half if there are un-upgraded and upgraded stacks). So, even with (hypothetically) equal intelligence to the player, the AI will still have much less variability in tactics.
This type of gameplay (if someone likes it) can be achieved now by simply agreeing to play with mandatory auto-combat.
But I don't know about that. I'm afraid that auto-combat "fool vs fool" is not the same as auto-combat "smart vs smart". In addition, when it comes to neutral creatures, it seems easier to create a more or less intelligent AI than to do the same for an entire army with a hero. This should also affect the value of auto-combat.
Some kind of madness. I tried to read it many times to understand what was written here, I was looking for a grain of truth, but alas, I didn't find it.
in practice, everyone will use autobattle
1. What made you think that autobattle will fight better than a person? Because you heard the word "chess"? 2. If, as you say, autobattle will be used more often, and the losses will really be acceptable, how will this prolong the game in real time? Do you use autobattle when you have 2 angels against a horde of boars? Or not? This is actually a plus, as it will leave more time to think about the global strategy and the game itself pvp.
It is possible to achieve a similar game format (if someone likes it) even now, just by agreeing to play with mandatory autobattle.
What are you talking about? Have you tried playing what you are advising? There are such games in GR, only 1 minute, quick combat. Try it :D
I'm afraid that autobattle "fool vs fool" is not the same as autobattle "smart vs smart"
Exactly! That's right!
Battles with neutrals will be reduced to abusing several "ironclad" ways to overwhelm the opponent with superior forces and standard protection against surprises from smart neutrals.
So, according to you, is it normal when you take 1 dragonfly and start abusing neutrals, leading them around the map, shooting and spitting at them? Is this the level of play? It is expected that some will cling to this, presenting it as the holy grail, but no one is forcing you guys to play anything else. Play the classic. Enjoy leading a horde of golems back and forth, it's so skillful. Otherwise, God forbid, all the schemes that have been spent years studying will stop working, no, we don't need that.
make the AI a more interesting opponent
a truly worthy goal. But its implementation (I am sure) is too complex. But I am also sure that the current version is the result of the developers' miscalculation, and not a deep plan to create the most interesting gameplay. When they abuse zombie farming. This is especially evident in the TE/VT versions, where the neutrals have also had their morale taken away! In general, it's terrible, they took away the stick that the blind man was leaning on. And all the units on the map turn into amoral golems/elementals/undead.
In general, I don't believe in the success of this project. The author of the topic did not leave any contact information: who he is, where their website is located, how to track the progress, etc. This is just a conversation for the sake of conversation.
Сейчас балуюсь этим https://vcmi.eu/download/ За то что нейтралы не бегают за единичкой, для меня это уже стоит того
By "longer," I meant not the actual time, but the in-game time.
Besides, ranged unit spamming will suffer less, as ranged units still have an objective advantage. Tactics like "overwhelming everyone with lightning" will suffer even less. But complex strategies that require skill will die out (battles with stacks of level 1 units, breaking through ranged units without tactics, Moon, Magog/Fireball/Ring of Cold) because losses are inevitable, and you'll think twice before attacking. You'll also have to forget about beautiful crypts or ranged units in Dwarf battles.
In reality, a practical, working tactic will look like this: take a reliable number of tough units (in case of a head-on battle) and conduct that same head-on battle in one way or another. Of course, you can try to outsmart the AI here, but the effect won't be worth the effort (I often see pointless battles in guides now, and what will it be like when you have to think like in chess in every battle to save a dozen Trolls?). You position yourself optimally, let the mobs approach, and press Q, because there's no point in doing anything else.
Regarding "the result of the developers' calculations, not a deep plan to create the most interesting gameplay" - this is obvious, but we have interesting gameplay, even if it's the result of the developers' calculations. Stupid neutral creatures are practically a mandatory element of gameplay in any successful game where these neutral creatures exist. But the opponent is a different matter. Hunting should be easier than war.
In general, any decent strategist can write an adequate algorithm for the AI's strategic behavior – it won't be more complex than a tactical algorithm. We're not talking about multi-day plans and chains, but about simple things that will at least double its strength. I've also watched a lot of gosolo, and it's completely nonsensical there
I understand that professionals have been bored for a long time. But have you thought about the newcomers? What will happen if an AI immediately defeats a novice without giving them any chance? Lider1
1. What makes you think that the autobattle will fight better than a human? Just because you heard the word "chess"? 2. If the autobattle is used more often, as you say, and the losses are really acceptable, how will this prolong the game in real time? Do you use the autobattle when you have 2 angels against a horde of boars? Or not? This is actually a plus, as it will leave more time to think about the global strategy and the game itself, especially in PvP.
And there will be. To develop skill, you need something to practice on, but how will you learn when you are being practiced on?! Perhaps, for "farming," you could create a separate, more advanced logic for those who are really bored – that would be a good idea. But not more than that.
They thought about you a long time ago. Play the original. No one is forcing you to play anything else. mrshch
And complex options that require skill will die out.
Not knowing how it will actually be, to say in advance that "complex tactics will die out" is a bit reckless, don't you think? Perhaps more sophisticated tactics will appear than there were before. Who knows. mrshch
In general, even a mediocre programmer can write an adequate algorithm for the strategic behavior of the AI - it won't be more complex than a tactical algorithm.
This is assuming that the programmer knows how to program. You can be a mediocre strategist and not be able to write anything because you lack programming knowledge.
And somehow, no one has written anything in all this time..
And I repeat: no one is forcing you. And there's nothing to force. There is no real subject matter here.
Сейчас балуюсь этим https://vcmi.eu/download/ За то что нейтралы не бегают за единичкой, для меня это уже стоит того
This is assuming the player knows how to program. You can be an average player and not write anything because you lack programming knowledge.
And yet, no one has written anything in all this time.
I meant simply to abstractly write an algorithm. A step-by-step instruction on what and how to do. This is not difficult for a skilled player. The challenge lies in translating it into a programming language and implementing it in the game – this requires being a very good programmer.
I meant to simply write an algorithm in an abstract way. A step-by-step instruction on what and how to do. This is not difficult for a skilled player.
To write an algorithm, you need an algorithmic understanding of the development of actions and tactics and strategy – this skill is very rare... maybe only a few out of a thousand or more possess it. Developing skills in intuitive behavior on the map, without realizing "why I did exactly that!" – anyone can do that.
Let's say, among the current players present on the forum, this skill might be possessed by Vik and me. I don't have the skills or desire to develop online gaming skills. But Vik could be "challenged"... he seems like a motivated person.
In any case, writing such an algorithm is a huge task that requires a lot of time and dedication from the performer.
.... шло время, а ситуация не менялась. Мясо продолжало визжать и пищать и главное НИЧЕГО не делать!!!
In any case, writing such an algorithm is a huge task that requires a lot of time and dedication from the developer.
This is if the goal is to "teach the AI to play like an average Heroes player." But if the goal is to "play better than it plays now" - it's not difficult. Even just changing the build order can lead to a significant improvement.
This is if the task is to "teach the AI to play like an average Heroes player." But if the task is to "play better than it plays now," it's not difficult at all.
I agree. "Achieving a result" is a difficult and complex process, while creating the appearance of action is "not difficult at all."
Even if you just change the build order, there will be a sharp, qualitative leap upwards.
Uh-huh... and find enough Heroes players who would play all sorts of mediocre games without losing the community...
.... шло время, а ситуация не менялась. Мясо продолжало визжать и пищать и главное НИЧЕГО не делать!!!
[igrik]
