With Legendary Heroes out, I can finally start reading how people are playing the “finished” game. Kael’s gotten me tons of reports and I’ve started tackling them one by one.
I am hoping to have the first batch ready for the next update (perhaps as early as next week). But that won’t be the end of it. I plan to take care of some low hanging fruit strategy suggestions first and then go on and deal with the more challenging aspects that will require a lot more play testing to do.
Stay tuned!
In the meantime, check out 8 out of 8’s video impressions of Legendary Heroes:
the 4X games are usually fine in terms of design
the problem is AI
developers are absolutely clueless about how to play games, let alone write AI for them
you cannot have a rewarding symmetrical game where your choices matter in the beginning and in the end unless you play on crazy difficulties (which make the game asymmetrical)
if you made the right decision, you should be rewarded for it with a huge advantage... taking that away trivializes the strategic elements of a game
the only way to keep it is to make sure decisions are so hard that the player never gets that huge advantage... that needs good AI
i've never played a 4x computer game that did not turn into a slog end game. That means that the design is not fine.
Er, Legendary Heroes doesn't require conquering the entire world. You can win by quest and you can win by building the towers. AI players will surrender if you're overwhelmingly more powerful than them. I agree it can be a bit a slog at the end, but I actually think Legendary Heroes does reasonably well at reducing the slog. If nothing else, there are normally a few Epic quests to give a bit of variety as you're crushing the AI players.
We obviously disagree about how good Civ V is which is fine, everyone likes different things. But I take strong issue with the idea that Civ V's problems come from over-polishing. We must be using different definitions of the word. Civ V's problems (relative to Civ IV at least) are that they tried to simplify too many elements, happiness being the most glaring problem but there are others, and ended up with a game which has some poor core mechanics. Simplification is NOT the same as polishing.
I would define polishing a game as making so that the game design is taut (ie decisions are such that multiple choices are valid and the choice made does make a difference), the game is stable (doesn't crash), core mechanics and systems don't have any visible bugs or major exploits (LH problems in this area include the save/load problems, instability after generating several maps and the growth/food mechanic being broken by queuing pioneers) and visible minor bugs or UI glitches during typical gameplay are few and far between.
I think Stardock have done a good job of making the game stable, a reasonable job of making the game design taut but a poor job of eliminating major and minor bugs and exploits. I say the last partially in the context of this being effectively the 3rd release of the game (whether you view it as 3 separate games or a base game with 2 expansions makes little difference) and many of the major bugs/exploits have been known for a very long time...
I used to play for PGL and used to be in diamond league in Starcraft and was the top ranked player in Total Annihilation's Boneyards's service. I like to think I can play.
However, writing "good" AI for a game like this is very difficult because players often play the game in drastically different ways and often times, spells and other game changing effects are added that I'm not familiar with (the tornado spell for instance).
It's not a matter of knowing how to play the games. It's about making the choice between designing a game with many many different strategies that causes the AI to have blind spots or one with a very very narrow range of player decisions.
Ultimately, since FE/LH is a single player game, some of the onus (imo) is on the player to choose not to make use of exploitative strategies. For example, you can exploit the game by training pioneers and then canceling them. Is that going to get "fixed"? Probably not because it's a single player game and that's a self-inflicted exploit.
There are a lot of straight forward AI strategies that can be enhanced (and hopefully v1.2 will help on that). There is generally little motivation to go and address things in the game that players intentionally and willfully do for themselves that make the game "easier". Players have a lot of say over how challenging the AI is for them.
None of what I say should be construed as me saying that AI work shouldn't continue to be done to the game to ensure that it continues to challenge experienced players. I am just emphasizing that player exploits tend to be low on the priority queue.
Correct me if I'm wrong on these two points, but the AI only surrenders if you are beating them up and the AI cannot use the master quest victory.
That's not really that important though. I think my terminology was inaccurate, and there is less maintenance per turn than a game like Civ late game. That's a good thing.
What I meant is that in every 4x game where the goal is conquest, there is a point where the balance of the game tips from challenging game play to effortless cleanup. This is due to the snowballing effect of power, and it's something that games like Civ try to design around with systems such as happiness, corruption and war weariness. In the case of Civ games, this means that you usually end up conquering, consolidating, conquering, consolidating, etc. It's not perfect, but it does keep the ramp from going insane.
The only system that tries to affect this at all in LH is unrest, and due to the ease of overcoming it, it's mostly a non-factor.
As a result we have a game where the player has effectively won after 2 hours of playing, but must spend 6 hours cleaning up (on large maps), which is very bad. Granted, part of this is due to AI not wanting to quit and not being nearly as strong as its score suggests, and the AI improvements earlier mentioned will improve both sides of this equation (takes longer to effectively win, and less time to clean up).
Ideally I'd love to see some sort of mechanic cause the game to be over by around 4 hours, with 0 hours of cleanup.
The game should end quickly when it is clear who is going to win. I only say one thing: Singularity Planetbuster
Noticed a bug, a screen shot would have wrapped it up, but...yeah sorry.
Anyways I assume everyone noticed this already but...
If you line up your troops all right next to each other and there isnt space to walk around them. The AI does nothing and constantly passes on their turn.
And epic army can be defeated by a single archer doing 1 damage a shot if they have a line of whatever in front of them.
| = border, X = Troop Square
|XXXXXX| = AI stops doing anything and just skips turn and lets you pick them off at range.
Its a common tactic to line up your troops like this due to swarm tactics, and I hate having to do...
|XXX_XXX|
|___X___|
In order to not have the AI sit there and die to ranged.
What do you mean? That an exploit that would give you, say, unlimited population growth, would not be a prioritized fix? If you are hoping to get any sort of competitive single player community (GOTM and the like), game-breaking exploits should be fixed (or, at a minimum, detectable/loggable).If, however, you mean that one can increase growth somewhat by tactically queuing pioneers at the right points in time (which you currently can) I agree they can be left in.
For this reason, I would also like to see autosave fixed, and ideally the seed for the RNG should've be taken from the save file so as battles and drops are repeatable (as in Civ).
I have to disagree with this approach, serious exploits should certainly be fixed as a pretty high priority IMO. Particularly the pioneer exploit which is so easy to use and so powerful.
You're giving the player an awful lot of credit if you think they can sit there with their capital not growing due to maxed out food and NOT think, mmm well, I might want a pioneer in the future, would it hurt if I just queued one up, I'm not doing it because of that exploit of course. However maybe I should queue two, cause you know, there are lots of places I could put outposts down...
Abracadabra! Magic, suddenly the capital is growing again. And then you are sitting there wondering why everyone is finding the game too easy and trying to find ways to improve the AI to match... fix the exploits (even if players are only using them accidentally) and you'll help the problem.
It's your game but obvious and overpowered exploits are a big turn off for me...
I'm seriously disappointed if fixing known exploits/bugs is a low priority just because it's a single player game. That's just sloppy.
Brad, you guys need to create a game where you're a game development studio. You'd have to allocate resources to different areas (marketing, developers, etc.), design the game, communicate to forum users, etc. Would remind me of the DLC Quest game on Steam.
It seems to me that creating and supporting a game is just like playing one- you have limited resources and need to find the best way to optimize them.
I can't wait to see how they implement people bitching and moaning on the forums over every little thing you say.
I think it depends on the type of exploit. I really couldn't care less about custom faction exploits, for example. No claims are made about them being balanced. If you find them unbalanced, don't use them.
I don't care too much about "exploits" which essentially revolve around playing the game optimally. There can be an issue with the AI using them properly, but in theory that can be programmed in.
I care quite a lot about exploits which reduce the suspension of disbelief, or seem broken. For example, not requiring resources other than gold when upgrading unit size doesn't seem right, and encourages an obvious exploit where you build smaller units and then upgrade them. Queuing pioneers to exploit population growth also seems odd.
I also care quite a lot about exploits which seem so powerful that they make the rest of the game pointless. Using a horse to steal Insane loot from undefended lairs isn't particularly against the spirit of the game and it doesn't break any game mechanic. But it's too powerful, and has odd effects (like higher monster difficulty levels actually making the game easier when using this tactic). Hopefully this will be fixed in the next patch.
This does leave a lot of grey area, but in general it's only the big, suspension of disbelief destroying exploits, that seem to make other parts of the game pointless, which I really object to.
Frogboy, I.. Don't really have much to add to your OP, other than.. Keep up the good work. Also if you are working on the Elemental RPG (Baulder's Gate style) and Galciv 3, or for the love of all things holy, a campaign for Sins (yeah, yeah I know, sandbox.. but I want to know so badly.. who are they running from!?), feel free to pm me some details, you'd have my full and undivi ded attention.
Now onto a comment about single player exploits.. I think you should disable loading while in a game, unless of course you are playing as the Oracle. After all, she is the only one who can see into the future right? I mean I just lost a huge battle that could mean ill tidings for my entire kingdom, and a dragon just ate one of my key food-producing towns! Damn, if only I had seen this coming, oh.. I don't know, ten days ago? Well, it is a good thing that my scribe keeps a daily record of the on-goings of his kind, just, and loving sovereign.. Scribe!! *scribble..scribble* Take this note to the Doc, and tell him to take it (queue breaking fourth wall) back to me!
I know of quite a few exploits in the game, I just don't use them. Except pretending I am the Oracle no matter what character I play. But be warned dear reader.. Knowing the future is to be trapped by it!
Only Diamond leauge? You should be at least Grandmaster, and play LH 24/7 every day so we don't have any bugs or exploits!
No JK, you guys are doing fine, but some people can't help but want everything and your soul
I am not the reader.
+1, I actually subscribed to these forums initially just to post a thread on this topic. The subject may not get much light, but it has some strong supportive arguments.
Hah, fun idea. +2
The GUI propels me to do this^.
I still find it incredibly frustrating to watch the computer run up to danger areas (say, right beside Lord of the Flame and build an outpost, and then have said Lord walk around the outpost to harass me. I don't know why this happens sometimes as I had a game just happen like this where there were no resources at all near the built outpost. I certainly can't take a Deadly monster that early in the game, yet the computer keeps on building near them so they roam and wreak havoc. At least if I EVER saw them attack the computer's stuff I would feel it was just the computer being dumb and laugh at them dying, but to have it come after me, walking past outposts and pioneers and armies of the computer along the way, it feels like I'm being punished alone for the computer's bad decisions.
Irritates me no end too.
The AI is still as of 1.2 targeting caravans with spells like wither.
One suggested idea: make caravans immune magic. Will keep the AI from doing this. Either that or make the AI think caravans are immune magic.
I do think the AI is playing better game overall though. AI raiding seems to be more advanced,using throwaway champs to raid (where you don't like them anyway, more injuries is fine) is probably a good idea.
I wish there was a way they could use steamworks to watch how people play and then incorporate the strategies into the AI.
Feedback:
Something I notice a lot vs AI in LH, they usually go for different targets with random scattered armies here and there, but rarely do that "doom" stack appear with the Sovereign. Is it possible to make doom stacks more common? As in they will only declare when they have one ready, and then proceed towards capitol while of course sending out those harass parties. I remember in FE vanilla that on Insane I got such huge stacks sent against me (20+ sized full army (9)), and I had to sparely use mana and resources etc. to beat them because they were so big and the challenge was upped a bit.
I wasn't really addressing the idea of "good" AI, but rather competent AI as it relates to the problems with these type of games that people described in earlier posts
This isn't about you or Elemental, but really any symmetrical game where one side can get an advantage... It's just mathematically impossible to reward the early game without making the mid/late game one-sided.
People keep wishing for something that is simply not going to happen unless symmetry is gone or the AI is competent.
when all the "broken" 4X games can work fine in a multiplayer setting, I think it's pretty clear that AI is the weak part
I think the balance issues with pretty much all 4X games give a pretty good indication of how well the developers understand the strategies, but that isn't even the point. Even if people do learn how to play 4X games, they still don't know how to write strong AI for complex problems
This is a really good catch (the caravans).
There are many great features available to you once you register, including:
Sign in or Create Account
