PDA

View Full Version : Heroes 5 AI



Kranyum
08-27-2007, 04:41 AM
I would like to point out the fact that I hope the AI gets an important remake in TotE, because in the state the AI is right now it completely ruins the single player experience.

All single player maps I played proved to be a waste of time as the computer was turtling inside its towns waiting for me to finish him. and that could take hours (on big maps)

While the Multiplayer is fine, because it's a good game (dont know about balance issues) the Single Player and the campaings are a total waste of time. Not to mention the fact that playing the game is no challenge. The AI is no fun to play against and most of the time it ends up wasting my time.

Heroes 3 AI was really good and I enjoyed playing single player maps which always proved entertaining. I see not much of a difference in the adventure map play with heroes 5, so I wonder, why is this one so stupid?

I hope that the AI will be improved in the expansion, cause I doubt there will be many patches after that to adress this issue.

Discuss.

Kranyum
08-27-2007, 04:41 AM
I would like to point out the fact that I hope the AI gets an important remake in TotE, because in the state the AI is right now it completely ruins the single player experience.

All single player maps I played proved to be a waste of time as the computer was turtling inside its towns waiting for me to finish him. and that could take hours (on big maps)

While the Multiplayer is fine, because it's a good game (dont know about balance issues) the Single Player and the campaings are a total waste of time. Not to mention the fact that playing the game is no challenge. The AI is no fun to play against and most of the time it ends up wasting my time.

Heroes 3 AI was really good and I enjoyed playing single player maps which always proved entertaining. I see not much of a difference in the adventure map play with heroes 5, so I wonder, why is this one so stupid?

I hope that the AI will be improved in the expansion, cause I doubt there will be many patches after that to adress this issue.

Discuss.

Elementalist.
08-27-2007, 07:56 AM
I'm with you on that one. There was a thread on HC that shows it's pretty messed up.

I would expect it to be advertised if that was the case.. I don't think we will get a better one even if I consider it one of the top priorities. I can't play single anymore exactly because it's a waste of time.

PhoenixReborn06
08-27-2007, 08:36 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kranyum:
Heroes 3 AI was really good and I enjoyed playing single player maps which always proved entertaining.
Discuss. </div></BLOCKQUOTE>

There's the flaw with your argument. I've seen the heroes 3 ai do dumb things also.

There's always going to be a point where you are stronger and it's just a matter of finishing the map no matter which heroes game it is.

Inferno015
08-27-2007, 08:47 AM
I have to agreee that the AI is somewhat questionable. It doesn't bother me so much in the early game, and it doesn't even bother me in the late game if my enemy is weaker than I am (this is usually because the AI tends to have hundreds of weak heroes instead of a few strong ones). What really aggravates me is when you are storming across the map, and all that your enemy does is switch heroes in his castle or something of that sort. The AI ends up right where it started, but not before wasting about fifteen minutes of your life! In HoF, the caravans make this wait time even longer! For that reason, Kranyum, I completely agree with you that there should be a notable change in the AI in the nearest expansion.

PhoenixReborn06
08-27-2007, 09:34 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Inferno015:
your enemy does is switch heroes in his castle or something of that sort. The AI ends up right where it started, but not before wasting about fifteen minutes of your life! </div></BLOCKQUOTE>

Play on a harder difficulty if you want the ai to be more agressive. Also, I think you must be exaggerating about the fifteen minutes.

I've taken to timing AI turns. On the map six factions (large size 5 enemies) it takes 30-40 seconds for all the ai to move. Less at the beginning and the end.

_____


I think this is a good topic because we know the ai needs improvement and we desire it.

However, we had a thread about ai improvements that chris_sca said they were taking a look at. Hopefully they will make some of the more obvious fixes.

smith-b-d
08-27-2007, 10:09 AM
The AI's always appeared smart enough to someone of my low skill level, however i have noticed that one you beat its main hero its essentially dead for good... but thats no really different to the player now is it? sounds like more of a gameplay design issue to me.

In battles however the AI is definitely not smart at all, i can beat it when severely out numbered.

Kranyum
08-27-2007, 10:28 AM
Heroes 3 did dumb things true, but it kept you playing the game until the very end by being aggressive and attacking with multiple heroes, not to mention retreating and reusing a good hero.

Here it is not the case. Even in Hard and Heroic (cause I play only those), early game things go well, but immediately afterwards it turtles and does nothing. Of course even in the case of heroes 3 there was a point where the game was won, but the difference was the AI did not give up and still proved captivating to play.

On the other hand, regardless of how good or bad heroes 3 was it is clear that the heroes 5 AI is lacking and in the actual state it makes the game practically not enjoyable in single player. And we all agree that SP is an important part of the Heroes series.

So something MUST be done. Unfortunately, where it comes to Ubi and Nival, I am sorry to predict that it won't and the game will still be wasting my time.

BTOG46
08-27-2007, 10:54 AM
You think HoMM has AI problems, you should see the moans we're getting about AI in the Ubi Settlers 6 forums and that's only the open beta and the demo, the game isn't even out yet. http://forums.ubi.com/images/smilies/51.gif

I just can't get it through the dumb noobs heads that settlers 5 and 6 don't have an AI, mp is for human opponents only and in the sp campaigns and maps the enemies actions are just scripted.
http://forums.ubi.com/images/smilies/53.gif

Quantomas
08-29-2007, 04:16 AM
I absolutely agree with the OP. AI is the one thing that makes or breaks this series.

Nival, you deserve credit for a gorgeous game. When I first heard that you were going to revive the Heroes series, I wasn't sure what to expect. When I received my Heroes V box shortly after release, I was anxious to see what you have done, after all there were so many things that could break the game, the maps could feel too cramped, the tactical battlefield could have been stunted ... . As it turned out, you got all these things right, almost perfectly right. Congrats on the new magic and battle system, you did an awesome job on balancing the factions and the overland spells. In particular the new town portal and dimension door spell designs demonstrate how far you have gotten in understanding the gameplay. And to top it all off, you gave us absolutely gorgeous artwork, a marvel to behold.

There is that. The first time I raised an eyebrow was when I realized that the campaign maps relied on scripted events to throw a challenge at the player instead of making the maps genuinely challenging. You ask what's the difference? In the former case, you have to acquire the means to meet the challenge and the result is one or maybe a few battles in which you have to be victorious. In the latter case, you have to explore the map systematically to understand the strategic challenge it presents, and then to seek a way to succeed, possibly even against near impossible odds, requiring you to use all the means at your disposal and drawing on every trick you know. This is the kind of fun that made Heroes popular in the first place and garnered it such a large community. I bought all the expansion packs, not so much for new visuals, but because it was fun to find out how the new tweaks and creatures created new strategic challenges. For this to work, the AI is crucial.

Heroes III had a mathematically balanced combat AI, a masterpiece written by a real wiz. Every action and spell was mathematically evaluated to its impact on the battle, and that was no meager feat considering how different some of the actions are. Is it better to hit an enemy directly, place a stack so as to prevent a ranged attacker from shooting arrows, or cast a spell like slow or shield? This AI was a masterpiece and virtually always proofed competent. Who does not remember the enemy hero casting a spell or placing a stack in a way that you hadn't thought of but turned out quite effective? Still, I don't begrudge Heroes V's combat AI, it's more heuristic but fairly serviceable. The thing is Heroes III had the same mathematically balanced AI for the overland strategy, whereas Heroes V has what? What a terrific deep game it could be if it had an AI on par with Heroes III. It makes you groan in frustration, not really sure whether to laugh or cry, with how gorgeous Heroes V really is, but on the other hand hardly playable because of AI foolishness.

Enter Heroes IV. This is the one lesson that broke 3DO. I remember playing one of the campaign maps on the hardest difficulty and wondered why I only found weakly garrisoned enemy towns. Until I found the tomb stones of enemy heroes who had perished against neutrals scattered about. Combat was equally unbalanced, once you figured out how to use multiple heroes in one army equipped with potions and artefacts to hold your opponents at bay, you had a definitive edge in most battles. Some claim the downfall of Heroes IV was from this imbalance, but the reality is that the people who played the game felt that the developers lost touch and had no concept of what makes the game fun. This led the community to depart in droves. This same lack of conceptual understanding is what people now perceive in Heroes V's AI. To exemplify this, there is one map in the original campaign that requires you to travel over an inferno landscape with no towns of your own. I choose some patient check-it-out-all strategy and just for good measure took out one heavily garrisoned inferno town along the way. All I got in return was a scripted animation that obliterated the town and a message saying that I must have cheated. It wasn't that hard to do, it only required to recruit creatures diligently on the way and preserve troop strength with the resurrection spell. For a developer to assume what the players are capable of shows not only lack of imagination but also an arrogance by design, in this case it pertains to the AI and map design. Just because you are the developer it doesn't mean that you are smarter than the player.

Ubisoft, you are the one publisher renowned for being smart. You have a good sense for business and a track record that proofs that your understanding of the market is superior than your competitor's. Picking up the 3DO franchise was a smart decision. Don't led it be botched by an AI design that drags down an otherwise excellent product. Fix this, and you will have an incredible asset with a large community.

Nival, Ubisoft, come on fix the AI. You know who I am (your forum admin has my profile), and working professionally on AI for games you hardly can do much wrong to listen at least to what I have to say. If there is any doubt about mathematically balancing the AI for Heroes V drop me a line. It can be done in short order, e.g. four weeks, if you know what to do.

Kranyum, you are absolutely right the AI taking a quarter hour and more for a turn. Anyone, doubting this, try to create a huge random map with eight players and two levels. At first, everything appears fine, but turn processing grows gradually as the enemy heroes discover more ground. Eventually it reaches the point when turn processing time reaches unacceptable levels, which essentially makes the random map feature with large map sizes useless.

Quantomas

PhoenixReborn06
08-29-2007, 05:03 PM
That is all well and good but how about some specific examples on how to fix the ai.

Perhaps some formulae?

KingAlamar
08-30-2007, 05:45 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by PhoenixReborn06:
That is all well and good but how about some specific examples on how to fix the ai.

Perhaps some formulae? </div></BLOCKQUOTE>

A specific formula for the adventure AI is going to be a tough nut to crack http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif

At this point tweaks are about all that I'd want / expect to see. For example making sure the AI builds better heroes.

You could do this via the use of templates [multiple per hero]. If you don't like using multiple templates per hero then tweaking the % chance for getting appropriate skills based on the hero would be a decent alternative.

PhoenixReborn06
08-30-2007, 07:43 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by KingAlamar:
A specific formula for the adventure AI is going to be a tough nut to crack http://forums.ubi.com/groupee_common/emoticons/icon_wink.gif
</div></BLOCKQUOTE>

Exactly my point.

It's very easy to write this:

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">The thing is Heroes III had the same mathematically balanced AI for the overland strategy, whereas Heroes V has what? What a terrific deep game it could be if it had an AI on par with Heroes III. </div></BLOCKQUOTE>

And I have to agree that it's true. The adventure AI was pretty good in heroes 3 (though I disagree that the combat ai makes the proper decisions).

But like I've said before it's obvious Nival can't do it or won't spend the money to do it whatever the reason.

Finally:

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">AI is the one thing that makes or breaks this series. </div></BLOCKQUOTE>

No it isn't. People love the games irregardless of the quality of the ai.

A great AI would give the game more longevity.

Pitsu
08-30-2007, 10:19 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by PhoenixReborn06:
That is all well and good but how about some specific examples on how to fix the ai.

Perhaps some formulae? </div></BLOCKQUOTE>

Without knowing the existing AI structure and code it is hard, and often pointless, to suggest a detailed fix. If someone complains about Windows stability, you cannot ask him to code his own patch for Windows. Feedback like "crash occurs when doing bla/bla" or "AI behaves dumb in situation bla-bla" is what a customer should give. Proposing solution to the problem is developers work. Particularly without knowing the source code customer cannot go deeper than " Should try to save best hero by fleeing/surrendering", "should send out more scouts to distract player" etc.

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">
No it isn't. People love the games irregardless of the quality of the ai.
</div></BLOCKQUOTE>

Wow. http://forums.ubi.com/images/smilies/blink.gif You mean some hardcore fan freaks or "people" in general?

smith-b-d
08-30-2007, 10:26 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Pitsu:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">
No it isn't. People love the games irregardless of the quality of the ai.
</div></BLOCKQUOTE>

Wow. http://forums.ubi.com/images/smilies/blink.gif You mean some hardcore fan freaks or "people" in general? </div></BLOCKQUOTE>
Total Annihilation was awesome depite the pathetic ai... but who ever played against it anyway? multiplayer lan bashs were the way to go.

Quantomas
08-31-2007, 03:02 AM
Formulae? That depends, it's more of an algorithmic challenge than a straightforward formula.

The principle is probably best explained by looking at the tactical battles. A straightforward implementation will do two things: it will look for a way to deal out maximum damage, while on the other hand trying to preserve its own troops. A heuristic approach, the way things are, does exactly that turn by turn, and if the algorithm looks far enough ahead, comes to a reasonably good solution. An algebraic approach factorizes each possible action, i.e. assigns it a value how much it shifts the balance of power between the two factions, and builds a chain of actions and reactions from start to finish of the battle, choosing such a chain that appears most reasonable based on the known strengths of both armies. As the combat progresses turn by turn, the chain of actions is constantly reevaluated based on the tactical constraints that develop. On the surface both approaches look pretty much comparable. The important difference though is that the latter links the objective(s) with the chosen course of action. Consider siege combat. For the human player it's kind of obvious that at times it makes sense to sacrifice, lets say your stack of black dragons to take out the enemy catapult, if it's clear that the enemy flyers and ranged attackers can be dealt with and the battle won if the walls hold. How is the AI to make such a choice? In fact, there are more elements to siege combat than simple damage optimization: creatures can damage the catapult, the catapult can damage walls and towers, walls can delay and prevent attacks, and towers again can damage creatures. Not to mention that there are advanced tactics to defend with a partially breached wall, e.g. immobilizing opponents with Dendroid Arches in the moat or blinding creatures that stand in the breach. As it is, simple damage optimization is also the reason for the one gremlin or archer left bug, when an attacking knight starts to cast benediction instead of finishing off the last standing defender while the towers take down his troops . It's also part of the reason while attacking heroes fail to retreat, whereas in Heroes III is was difficult to capture an enemy hero with his artefacts. A factorized solution that chooses actions linked to objectives is capable of significantly more, employing more versatile tactics is only part of it. For example objectives that stem from the overall strategic requirements, e.g. to preserve certain types of troops, could be factored in, or different types of heroes could be made to favour different combat styles, pacing and tactics. But as said earlier, Heroes V's combat AI is servicable and fun to play against, and there should be no top priority to overhaul it. It feels a bit like playing against a brute, but then brute force is at the core of warfare.

One reason why the original Heroes of Might and Magic was so popular, was that the gameplay challenged your wits while never being unfair. This kind of gameplay made you come back and back again. An algebraic approach to the overland AI can achieve exactly that. There are a number of differences to the tactical combat but the principle is the same. Every action on the map is factorized, whether it's capturing a mine or building up your town, the value assigned to each action reflecting how much a power shift it represents in comparison to the other factions in the game. Essentially, heroes are the projection of your power in the realm. Having them idle costs you and is rarely justified, except when protecting an asset. The way the AI judges if something like that is justified, or any action for that matter, is based on its objectives to explore and conquer the realm. Fighting creatures, leveling up heroes, collecting treasure and building a power base, have all [EDIT:] comparable values assigned to their actions, which enable the AI to choose between different paths in the context of the current balance of power. Powerful heroes appearing at the borders, threats to bases and so on, can all be evaluated in a system of this kind.

I could go on and describe how proper strategic behaviour results from this design, but that would probably be beyond this topic. Implementing a system like this is absolutely non-trivial and requires experience. But people with the right skill can, judging from experience and assuming all the necessary hooks for evaluating objects on the overland map, hero stats etc are in place, implement this approach in approximately one month.

Seeing that Heroes V is a full-fledged product, I have no doubt that all the necessary hooks are in place. What I can't judge is Nival's level of experience with this kind of AI. You see, I really love this series and was saddened to see its demise with Heroes IV. I was all amazed when I finally had Heroes V in my hands and to see that they recreated it with an indepth understanding of the balance issues at the core of the gameplay, let alone the gorgeous artwork. And then all aghast that they got it almost perfectly right, but busted it with the strategic AI. So the idea of my post was to give them a nudge. I might even be tempted to help them in one way or another, if need be. The Heroes franchise traditionally had the best AI on the market. Combined with the clear cut game play, simple straightforward treasure-hunting conquest, it's just pure joy to play. I prefer it much over a game like Civ IV because this is way too much fettered by real-world paralleles, and strategic choices being subject to moral imperatives (to go to war or not, for example). But right now a Heroes V with a fixed AI just appears a dream.

KingAlamar
08-31-2007, 05:48 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by PhoenixReborn06:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">AI is the one thing that makes or breaks this series. </div></BLOCKQUOTE>

No it isn't. People love the games irregardless of the quality of the ai.

A great AI would give the game more longevity. </div></BLOCKQUOTE>

I would say that great maps & stories are what makes or breaks a HoMM game AND gives [primarily] contributes to HoMM's longevity.

I do believe, for SP, that the AI contributes much more than you imply though. I know I don't play H5 nearly as much as the other HoMMs and a large part of it has to do with the principles behind the AI and its implementation.

EDIT: That being said I wonder why doesn't Nival put in at least a token amount of time to make some cosmetic tweaks to the AI http://forums.ubi.com/groupee_common/emoticons/icon_frown.gif

Sorapak
08-31-2007, 06:45 PM
The A.I. is Heroes V biggest drawback,especialy if u play single player only,like me.

Kranyum
09-01-2007, 09:51 AM
@Quantomas

Your intentions to help Nival improve their AI are honorable but I'm afraid this will not happen.

The problem is they are not listening to fan suggestions and do as they wish.

I hope something will be changed but honestly I doubt it.

C_h_u_c_k_l_e_s
09-01-2007, 09:55 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kranyum:
@Quantomas

Your intentions to help Nival improve their AI are honorable but I'm afraid this will not happen.

The problem is they are not listening to fan suggestions and do as they wish.

I hope something will be changed but honestly I doubt it. </div></BLOCKQUOTE>

Actually, the problem is one of language. Nothing more, nothing less. Nival DOES converse with fans...Russian ones. As for english suggestions, those get through too...only they get filtered through Ubisoft first. Once you see the content in TotE you'll understand what I mean about fan suggestions. A LOT of them got included. However, I highly doubt you'll see AI improvements. Those would be made to the original game and I simply don't see them making sweeping, monumental changes to the AI at this late of a date. Their AI focus should be on H6, not H5 at this point.

Elodin
09-01-2007, 02:27 PM
The AI is set up not to play as well on lower difficulty levels.

If you want the AI to cast level 5 spells and play the best it can you must play on heroic.

Elementalist.
09-01-2007, 04:38 PM
I still find the language barrier a bit strange, I mean they know no English? http://forums.ubi.com/images/smilies/blink.gif I will agree that if there's going to be an H6 the focus on AI should be there. Sure that would please many fans to see one but it makes little practical sense after all this time.

Of course Nival had often seemingly plentiful time to do something right as say the original HoF RMG...And miserably failed but I hope they learnt from their mistakes.

I guess the (possible)amount of sloppiness we'll see in tote will be a sign of what is to come.

smith-b-d
09-01-2007, 07:06 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Elodin:
If you want the AI to cast level 5 spells and play the best it can you must play on heroic. </div></BLOCKQUOTE>
Hard is the most balanced according to the files... virtually the only difference between hard & heroic is how many resources you start off with.

Theoretically its possible to mod the AI to make better decisions on the adventure map, but skill choices & battle map is all hard coded i think.

C_h_u_c_k_l_e_s
09-01-2007, 09:27 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Elementalist.:
I still find the language barrier a bit strange, I mean they know no English? </div></BLOCKQUOTE>

It's more of a problem of SOME of the Nival people knowing broken english and others not knowing it at all. So Fabrice was trying to talk to the partial speaking english reps at nival (and english isn't Fabrice's native tongue either) and then the reps are having to translate to the people who do the work. Lots of little things get lost in translation.

But for me, the biggest gripe I have is that the people they have doing the english text for the game...it's not thier native tongue either. Hence all the typos. http://forums.ubi.com/groupee_common/emoticons/icon_frown.gif

Quantomas
09-01-2007, 10:57 PM
Kranyum, thank you for your friendly words. Makes me think, though.

Chuckles, that is a valuable bit of intelligence. Maybe as a community we can organize some sort of concerted effort. What do you think? The way I see it, Ubisoft is content to have just another franchise on their hands that generates them a modest but reasonable revenue stream. In a way very much like what they did with The Settlers franchise, they downsized the developer, and are content to sell with each installment a couple of hundred thousand copies in Germany and make a profit on that.

But in our case we have Heroes of Might and Magic, one of the most popular franchises in the turn-based genre. Why should it sell any less copies than Civilization IV, for example? The turn-based crowd is huge, essentially the same people who play CivIV also play GalCiv or HoMM for that matter. You invest a lot of time in these games and carefully choose what you play, and I wager, one of the top criteria is that the challenge is worthy of your time. How would it look if your rivals in CivIV didn't bother to search tribal villages or wouldn't use their naval forces properly? You would probably play it once and tell your friends something like "Well, it did look nice enough, but I don't know, somehow it didn't feel right." Heroes V made the transition to fully imagined 3D just as well, if not better, than CivIV, and HoMMV's user interface and gameplay is some of the best around. The way Nival brought the animations to live is outstanding and makes you chuckle with delight, and yet the game is shunned by a large part of the crowd. Why is that, hmm? Speaking in business terms Heroes V vastly undersells. Imagine Heroes V + expansions with a sophisticated and versatile AI that offers a depth of gameplay no other title can offer. What will be the effect on sales? An additional 2 million copies? That would be 30 million bucks of extra revenue, even for Ubisoft this is serious money.

Now, Chuckles is right, normally redoing an AI is a monumental effort. But most of this effort comes from the team learning to find a solution, not from the size of the required code base. So my suggestion is to take a proven solution and adapt it, and if they need pointers, or even a blueprint, all they have to do is to ask. In the worst case it will cost one of their employees a month of work, but if things turn out right, it's a win-win-situation for all. It can earn Ubisoft a ton of money, take Nival out of their current predicament, and the fans get the game they wish for.

So, are there any Russion speaking fans among us who could rub this message under the right noses? Chuckles? Do you speak Russian, or did you learn this information from others in the community?

PhoenixReborn06
09-01-2007, 11:13 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Quantomas:
and yet the game is shunned by a large part of the crowd. Why is that, hmm? Speaking in business terms Heroes V vastly undersells.
</div></BLOCKQUOTE>

You have made me curious. Can you point me to some data about sales?

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">
Imagine Heroes V + expansions with a sophisticated and versatile AI that offers a depth of gameplay no other title can offer.
</div></BLOCKQUOTE>

Sounds good.

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">
So, are there any Russion speaking fans among us who could rub this message under the right noses? Chuckles? Do you speak Russian, or did you learn this information from others in the community? </div></BLOCKQUOTE>

I can suggest things to russian speaking members at the heroescommunity forum but I'm not 100% clear on what the message is.

Quantomas
09-02-2007, 12:38 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by PhoenixReborn06:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Quantomas:
and yet the game is shunned by a large part of the crowd. Why is that, hmm? Speaking in business terms Heroes V vastly undersells.
</div></BLOCKQUOTE>

You have made me curious. Can you point me to some data about sales? </div></BLOCKQUOTE>

The NPD sales figures show how good titles are doing in North America. They are typically published weekly and echoed around games and news sites like IGN. CivIV made it to the top positions and stuck there for a good time. Knowing a bit about the overall sales volumes my best guess is that they sold a good sized seven-figure. There are of course other sources, like the publishers quarterly reports (2K in this case), but you have to know a bit of their portfolios because they tend to compound sales figures. Normally, if you know enough about the market and if you have input from different sources, you reasonably can extrapolate sales figures, and of course you hear sometimes figures from collegues or friends who got some insider figures.

With Heroes V things were not so fortunate, never really made the top spots and sales appeared to have flagged. Rumoured staff reductions at Nival also don't help this impression.

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">
Imagine Heroes V + expansions with a sophisticated and versatile AI that offers a depth of gameplay no other title can offer.
</div></BLOCKQUOTE>

Well, that's the message. In my earlier posts in this thread I described an approach to AI, they could adapt to Heroes V in short order, e.g. four weeks, if they are interested. If they have no familiarity with this type of AI, I can point them in the right direction or even provide them with a blueprint. I my eyes it's just the will that is required, and they can't do much wrong at least to listen. The implementation time of four weeks is not a wild guess, but comes from professional experience of working with this type of AI, and knowing how a project like Heroes V is structured and implemented. If they have not done something extraordinary out of order the quote is reliable.

alias-hw
09-02-2007, 01:16 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Quantomas:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by PhoenixReborn06:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Quantomas:
and yet the game is shunned by a large part of the crowd. Why is that, hmm? Speaking in business terms Heroes V vastly undersells.
</div></BLOCKQUOTE>

You have made me curious. Can you point me to some data about sales? </div></BLOCKQUOTE>

The NPD sales figures show how good titles are doing in North America. They are typically published weekly and echoed around games and news sites like IGN. CivIV made it to the top positions and stuck there for a good time. Knowing a bit about the overall sales volumes my best guess is that they sold a good sized seven-figure. There are of course other sources, like the publishers quarterly reports (2K in this case), but you have to know a bit of their portfolios because they tend to compound sales figures. Normally, if you know enough about the market and if you have input from different sources, you reasonably can extrapolate sales figures, and of course you hear sometimes figures from collegues or friends who got some insider figures.

With Heroes V things were not so fortunate, never really made the top spots and sales appeared to have flagged. Rumoured staff reductions at Nival also don't help this impression.

</div></BLOCKQUOTE>

There is more than sales numbers compared to other games that decides if a game is a success or a flop.

Ubi never really expected Heroes 5 to be a great hit. They expected some decent sales and (this next part I assume) to get the name out there to that people remember the Might and Magic name again so that Ubi could launch Dark Messiah which they did expect/hope to become a great hit.

They never expected H5 to be a great hit, yet already about 6 months ago H5 already sold more copies than H3 every did in its entire existence. So, in the eyes of Ubi H5 was a HUGE success that did FAR better then ever expected.

So, H5 is not underselling; it is actually overselling. Just that, yes, compared to some other games H5 sells a lot less copies. But then again, compared to yet other games H5 has sales numbers that those game makes can only dream about.

And the staff reduction at Nival is not a rumor but a fact. However, it is also already explained here on this forum by the H5 producer why this is the case. And if you think about it business wise, I would also have fired people. Short version is that people were hired to do a job. That job is now done and Nival moves on to other parts. So those people are no longer needed and they are fired. Hard, cold, but fair and normal business.

Mind you, the above is pointing out sales aren't as bad as some people would like have us believe. On the matter of the IA, I agree that it could be better. I said from the beginning that the AI of H5 is underperforming. Hell, the AI from H3 was better than the AI from H5. Yes, it made some stupid mistakes, but it did continue to fight until the last minute while the AI of H5 stops playing when the game is about 75% done and then just goes and sits in its last town sulking and playing a I don't want to come out and play anymore' game. Yes, some people went and explained that the AI of H3 only SEEMED better and that is actually cheated even more behind the scene. But, I don't care about that. I want to have fun playing and if that involves an AI doing all kinds or weird cheating I don't care. I much rather (incorrectly) think that I'm being challenged than have an AI that just runs home and then sits in its town and waits for you for that last battle... which you will win anyway since you have taken over other towns as well and have a HUGE army.

But... talking about the AI of H5, and even TotE now is really of no use.
Why? Because H5 is already old news for a game maker like Ubi. As soon as a game comes out, it's considered over and done with and they move on. The support that we had (well, support...) is only done so that you are kept happy enough to buy the next product in line.

TotE... again too late to do anything about it. This close to the release date they aren't going to mess with it anymore. If they didn't already incorporate a better AI (which I doubt) they aren't going to create it now. TotE should be considered a finished product and you can forget about things being added at this stage. Only thing you can do now is wait and see what will actually be in the game.

Really, the only thing that might still be changed if enough noise is made is H6. Maybe we get really lucky and they keep some of the major complains, like the AI, in mind for H6.

Maybe, but I doubt it.

H.W.

Elementalist.
09-02-2007, 01:29 AM
@Chuckles
Oh, I forgot that either of them has no native english language. However ubi should have made better choices about who was charged with writing the english texts.

C_h_u_c_k_l_e_s
09-02-2007, 07:09 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Quantomas:
So, are there any Russion speaking fans among us who could rub this message under the right noses? Chuckles? Do you speak Russian, or did you learn this information from others in the community? </div></BLOCKQUOTE>

No, I'm in the US and I don't speak Russian. I've not been in direct contact with anyone from Nival. However, I'm in touch with Fabrice daily (well...almost. The kids and work DO interfere sometimes. *grin*). I do gather requests and suggestions and complaints here on this forum and pass them along to Fabrice who, in turn, passes what he considers important on to Nival. Nival IS aware of the AI shortcomings, as is Fabrice. The question is a matter of priorities. Nival has deadlines they must meet for Ubisoft. And they've got a long list of items they want to address (things to change and things to add). And given the time constraints (and the budgetary ones), the AI is rather low on that list for H5. In fact, IIRC, Fabrice stated that a major AI overhaul isn't an issue for a patch. Now I'm no tech expert, but I think if they were going to address the AI in a major way...they would've done it BEFORE either expansion came out. H6 plans are already in the works, so I'd assume that any major looks into the AI would be focused towards that.

Concerning the sales of H5....I'm sorry, but you're mistaken. First off, the Turn-based market is one of (if not THE) smallest markets in PC gaming (or gaming altogether). Second, Heroes has traditionally RULED that market. Third, H5 sold far above their own expectations. If it hadn't been a success, we wouldn't have seen an expansion...much less two and a followup game. I've seen the actual sales figures...but I don't recall what they were right now. I'll have to dig around and see if I can find them again.

C_h_u_c_k_l_e_s
09-02-2007, 07:14 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Quantomas:
With Heroes V things were not so fortunate, never really made the top spots and sales appeared to have flagged. Rumoured staff reductions at Nival also don't help this impression. </div></BLOCKQUOTE>

What really makes H5 sales hard to track is the amount of sales in Europe, Russia and china. A VERY large portion of sales came from Russia. In order to get a decent estimation of sales, you have to look world-wide.

In regards to the staff reductions....that wasn't a rumor. It DID occur. However, it was a long time in the planning. They were positions that were not critical to game development. Mostly along the lines of story writing and a few artists. Ubisoft will take over some duties as will others at Nival. It won't and it hasn't slowed down production/development at all. People have made WAY too much of this story.

C_h_u_c_k_l_e_s
09-02-2007, 07:22 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Elementalist.:
@Chuckles
Oh, I forgot that either of them has no native english language. However ubi should have made better choices about who was charged with writing the english texts. </div></BLOCKQUOTE>

I partially blame myself for this. During the late development phase of H5 original, Fabrice was asking around for translating help. Not really full translating, but more along the lines of double-checking the translations. Unfortunately, I'm in the US and as most US gamers, we're under the weird delusion that all games are made in English BY English-speaking people. NWC games never had to worry about translation. So when I heard him announce the need, I was thinking they needed help to translate the game FROM English into other languages...such as French or Russian. If I'd REALLY thought about it, I would've realized that it was WRITTEN in French & Russian and they needed help translating into OTHER languages...including English. But with no double checking, they went with their own non-native English translator. I won't let it happen again. *sigh*

KingAlamar
09-02-2007, 07:47 AM
Is it only me or with H5 decent sales figures + decent expansion sales [it sold more than H3?? .... wow] means that we're quite likely to get more of the same?

I mean why bother spending a lot of $, time, etc. on an AI when obviously a H5 AI is good enough to drive decent sales of the game.

The only way that I think that we can make our point is through our purchases and to let people know why a game wasn't purchased.

Basically as long as HoMM is selling well why would it be in a company's interests to make a better mousetrap?

I know it sounds bad but I hope that King's Bounty blows the socks off of what we got from H5 BECAUSE that competition may force UBI/Nival to release a better product for H6.

[God I hope market forces actually work this time ....]

C_h_u_c_k_l_e_s
09-02-2007, 10:52 AM
No, you're looking at it from the wrong angle. With higher than expected profits, they'll put a larger budget into place for H6. AND with Nival having one entire game + expansions under their belt, they'll work more efficiently on h6 and be allowed more freedom for unique ideas and innovations. They were pretty handcuffed with the development of h5...all things considered, h6 should be the true benchmark on what Nival/Ubisoft can do with the Heroes series.

Elementalist.
09-02-2007, 11:01 AM
Here's hoping that this is a good thing http://forums.ubi.com/groupee_common/emoticons/icon_biggrin.gif
As for the translation part that's an interesting anecdote but one can never know and it doesn't matter anymore.

KingAlamar
09-02-2007, 11:12 AM
How's this Chuckles: I hope you're right and I also hope that King's Bounty is a roaring success to fuel that extra bit of competitive fire in UBI's stomach.

Either way H6 should be the benchmark that we look to for things like documentation, map editor, AI, etc.

Quantomas
09-02-2007, 07:37 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by C_h_u_c_k_l_e_s:
What really makes H5 sales hard to track is the amount of sales in Europe, Russia and china. A VERY large portion of sales came from Russia. In order to get a decent estimation of sales, you have to look world-wide. </div></BLOCKQUOTE>

Interesting. I haven't payed much attention to the younger markets yet, and so far just assumed that a lower average income there is reflected in the retail pricing, so that these markets effectively contribute only a small part to a publisher's total revenue. Makes me curious about how much a box of Heroes V sells in Russia.

Regarding the size of the turn-based community in North America and Europe alone, people who actually spend bucks on a box, it's traditionally more than 2 million. It's true there are much larger market segments, but it's still a huge crowd. Civ4 did a good job to sell up to its potential whereas Heroes V did not. Still it's good to hear that the revenue was deemed good enough to be a success and that Nival will build on it.

Considering all things you guys have said, it seems indeed a bit unlikely that anyone will go and upgrade the AI in Heroes V right now, even if it is as simple and straightforward as I say. I guess I have to blame just myself, I could have said something way earlier. I keep my offer standing, maybe my message will reach the folks at Nival sooner or later and it's useful for Heroes 6 just as well. The thing I am suggesting is called an algebraic factorizing AI, there is not much documentation on the web, it's mainly been hatched by a number of AI programmers working for different companies. It's more than worth to check it out though, it's straightforward to use and the AI truly powerful, in short it can offer the Heroes franchise the one distinction it needs to claim back its crown. Don't sit contently on your gold mine when you can be a hero and reign supreme over the land.

Kranyum
09-04-2007, 09:24 AM
Well, that's that...

The AI won't improve, and the single player will remain boring at best.
And as Multiplayer lasts to long to be playable by lots of people, I guess the game will remain at most mediocre.

PhoenixReborn06
09-05-2007, 12:11 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kranyum:
And as Multiplayer lasts to long to be playable by lots of people, I guess the game will remain at most mediocre. </div></BLOCKQUOTE>

I don't think that's a great argument. How long do you spend playing a single player game?

It's possible to save and reload so it isn't that tough to play a multiplayer game over a couple of days. My games usually take between 4 and 6 hours but I wouldn't playing longer if someone wanted to play a bigger map spread out over some days.

The expansions include simultaneous turns which cuts down on some time. They really should add that to the base game.

Meanwhile playing with the files can yield interesting results on the ai. We are all awaiting the results of smith_bd's ai mod.

smith-b-d
09-05-2007, 01:40 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by PhoenixReborn06:
Meanwhile playing with the files can yield interesting results on the ai. We are all awaiting the results of smith_bd's ai mod. </div></BLOCKQUOTE>
If it doesn't kill me first. Its giving me the sh!ts at the moment, changing stuff seems to make everything more erratic/random... its hard to tell if somethings a improvement or not.
I'm thinking of uploading a few different files and letting others decide which works best.

Edo_555
09-05-2007, 01:55 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Elementalist.:
I still find the language barrier a bit strange, I mean they know no English? http://forums.ubi.com/images/smilies/blink.gif I will agree that if there's going to be an H6 the focus on AI should be there. Sure that would please many fans to see one but it makes little practical sense after all this time.

Of course Nival had often seemingly plentiful time to do something right as say the original HoF RMG...And miserably failed but I hope they learnt from their mistakes.

I guess the (possible)amount of sloppiness we'll see in tote will be a sign of what is to come. </div></BLOCKQUOTE>
Realy if you wan't to know programming, you must alsou known english. In russian forum in thread dedicated to AI one of the developers once posted, that they sometimes watch this forum for any AI isues. He was AI developer.

But if Nival arn't international comunity friendlie, they shound expect, taht comunity will be frienly to them.

If Ubisoft would expect Heroes to be AAA hit, then they probably would picked different developer and gived more time/money.

Kranyum
09-05-2007, 02:34 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by PhoenixReborn06:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kranyum:
And as Multiplayer lasts to long to be playable by lots of people, I guess the game will remain at most mediocre. </div></BLOCKQUOTE>

I don't think that's a great argument. How long do you spend playing a single player game?

It's possible to save and reload so it isn't that tough to play a multiplayer game over a couple of days. My games usually take between 4 and 6 hours but I wouldn't playing longer if someone wanted to play a bigger map spread out over some days.

The expansions include simultaneous turns which cuts down on some time. They really should add that to the base game.

Meanwhile playing with the files can yield interesting results on the ai. We are all awaiting the results of smith_bd's ai mod. </div></BLOCKQUOTE>

Well, simply put, I played Heroes 3 on single player a lot of time. A few times per single player map.

But here, it's pointless to do that as it's not that challenging...

PhoenixReborn06
09-05-2007, 10:25 PM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kranyum:
Well, simply put, I played Heroes 3 on single player a lot of time. A few times per single player map.
</div></BLOCKQUOTE>

And how often do you win?

Kranyum
09-06-2007, 05:17 AM
90% of the time ... but before you say what I think you'll say, I'll point out that the game is fun from the first minute to the very last, and makes me want to try the map again with another faction or position.

PhoenixReborn06
09-06-2007, 08:59 AM
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Kranyum:
90% of the time ... but before you say what I think you'll say, I'll point out that the game is fun from the first minute to the very last, and makes me want to try the map again with another faction or position. </div></BLOCKQUOTE>

I'm going to say it anyway. But I think we're on the same page here. Due to threads here I have noticed that the ai turtles up in heroes v after a certain point. And I have also noticed that heroes 3 ai acts more interesting on the adventure map. When you know the games inside and out you can win all the time no matter what.

That's why King Alamar' suggestions are pretty good, cosmetic changes. I've been saying from the start I wish they stopped doing the obvious dumb stuff even if they aren't going to do algebraic expansion or whatever...

ZzzombieBunny
09-06-2007, 09:44 AM
Im just curious, what difficulty setting AI are we talking? Anything below Hard (or even Heroic to be honest) cant really be comparable.

The differences between Hard and Heroic are bigger than some might be aware, its not just starting resources that differ between the various difficulties. Its more about the differences between the decision behaviour, movement planning, restrictions on town building and town defense.

It is kinda odd how the AI settings have been setup, the fact that an Easy setting was added a few patches ago leads me to think that the overal AI settings are adjusted towards a more shallow and predictable AI, rather than give it wider playing field.

For example the AI has values that are rather restricting on allowing it to plan ahead more than a day or two. Its also rather scared of the Human player, avoiding confontration as much as possible untill having a superior force. On the lower difficulties its however more prone to attacking, although with less consideration towards if it even has a chance or not to win. This makes for a rather boring gameplay on both the lower and higher difficulties, because you pretty much always know what to expect.