Greetings!
Yesterday we released Elemental: Fallen Enchantress BETA 5-C. This build is our first pass at balance and UI enhancement.
Next week we will have a BETA 5-D which will focus almost exclusively on bug fixing and balancing as well as some new AI counter-strategies so those of you who have posted your game play strategies online (foolish humans! <g>). We are reaching the home stretch. We will also have a Beta 5-E. After that, we'll see where things stand.
Many of you have been in the beta for a very...very long time. I can't imagine the challenge it is to evaluate the game from the perspective of a typical PC gamer. But we're going to ask you, once again, to try to do just that:
Please go to: https://www.elementalgame.com/journals and vote on your impressions of the game in your hands.
We are very excited to see what players think.
These past two years have been wonderful. The community we have built together has been amazing and on behalf of the Stardock team, we really have enjoyed listening to your ideas, impressions and suggestions.
Like all Stardock games, 1.0 isn't the end of our journey. Even as I type, Sins of a Solar Empire: Rebellion v1.1 nears completion. Ideas and code snippets for FE expansions and successors float around. And there's some exciting things we hope to announce in the coming months.
We look forward to your opinions:
https://www.elementalgame.com/journals
Are you really arguing that skyrim UI is not horribad? That the AI doesn't suck? That combat is not horribly unbalanced? That the game wasn't unplayably buggy on the PS3 at lunch?
The plot and dialog were both atrocious and as I said, people are hard up for "epic RPG" which is indeed what skyrim is and what you just described.
Skyrim is definitely engrossing but that is despite the horrid issues that have brought down many games but were not enough to bring down skyrim because of the benefits I listed. Repeating those pluses which I described to me in no way invalidates my assessment or analysis of the game.
Beta 5C is the best beta I've played so far. I love the hero nerf on hitpoints. I actually injured some heroes today!
While some of the monster changes are a bit over the top (mentioned earlier in this thread) I do like the path it is taking. The world comes back alive after having died a bit in the last beta's.
I DO hope (like others) that some of the really 'old' bugs get finally fixed in this polish stage. For me they're not as gamebreaking as for some other people around here, but I do think they need to be fixed.
Definetly agree. Have been playing the whole weekend now, and this game is just pure awesome! I love the changes that was done in this latest version. The game has strategy, it has adventure, and it has charm. Beats everything made earlier of fantasy 4X. I don't dare to imagine how unthinkable great it will be after a few expansion packs. This game deserves to sell good.
NB! And please give it multiplayer. The best fantasy 4X in the world needs it....
I voted excellent.
EFE is really a great come back story. I went through Demigod and WOM and sometimes felt like my warnings and posts were going unheeded. This year, you guys have come out of the gate at full force with SinsReb and now EFE.
For fun, I tried playing a game of Age of Wonders: SM and you can't really appreciate how much of a world EFE delivers until you go back and look at these older games.
I voted fair. Currently, Fallen Enchantress has huge scaling and pacing issues. While it's a 100% balance problem (and as such, fixable by mods and within the scope of beta 5), it still makes the game a whole lot less fun to play to me. It also makes the game content-poorer, because only good balance and pacing gives relevance to all content. If there's no point recruiting Darklings or Ogres, then it's just as if they weren't in the game in the first place.
The AI still needs a significant amount of work, and while I know it's still work in progress, I cannot give a grade based on what I will hope the AI to be in 2 months. And this issue tends to resonate with the pacing/scaling issue - the AI ends up being behind the player in terms of technological and military power, and the scaling issues amplifies how much of a drawback this ends up being.
Poor, the AI is so bad at the moment its hard to tell if the game will be any good. I am still seeing crazy tower building, AI that just stalls out, only gets like 2-3 cities while others get 8-10 from auto-turn games, plus it still doesn't seem to cast a wide array of spells. Since this is a single player only game the ai is crucial and right now its essentially broken. How the tower problem hasn't been fixed yet is just strange. This is a simple conditional/deterministic AI function dont't build x object within x squares how its failing over so many patches is troubling that there is some core engine problem that they can't figure out.
Seriously everyone should enable cheat on a large map with 6 players (no tarth) hit ctrl-u to reveal then ctrl-z to auto-turn and come back in a few hours and look at how the ai is performing over a few hundred turns its pretty bad. The tower layout is crazy and the roads don't make any sense either.
I'm really enjoying the latest beta, the balance improvements have actually made me choose different technologies I wouldn't normally pick, such as getting basic weapons! I think the early game balance has improved alot, but I think mid to late game still needs work. And you need to look at bugs such as the AI not picking up equipment and also monsters not casting magic during auto-combat (such as fighting the dark sorcerer).
I've been trying to keep from playing the beta so I can do 1st impression vids. Things were looking like they were moving along the right lines at last play through, though. I'll see if I can crank out a vid over the next 2-3 days or so.
Skyrim's UI is a freakin' disaster, worst UI ever really, but, first, it needs to run on consoles too, second, the modding community is so large and dependable that you could count on a "fix" within the first month from release. At least initially that won't be the case for FE and I'm not sure its UI will be moddable anyway.
About the need for an epic RPG, I disagree. Surely, people were asking for one, but if the game sucked people would have stayed clear of it anyway. Something similar happened to EWoM after all, did it not?
I'm not saying your points were wrong, just "partial". Skyrim didn't win so many players just because of cool dragons. Many games have those. The atmosphere, the setting, the music, the world, the cities, the fitting armors and weapons... so many things were spot on already at release. Of course, you might put it all together under the item "coolness factor" but it'd seem simplistic to me.
That's not to say a little more of "cool" wouldn't help FE: I've been advocating a "more magical" magic system since day one, for instance. But it's not just that, it's about the mix being more than the sum of its parts (as a well integrated experience) - while currently, as another poster said in this thread, I think it's actually a little less. Of course, it's not an exact science, people will disagree.
More prominent terraforming of the beautiful cloth map would be a blast, and fit in very well with the background story.
Let's give the raise/lower land spells to all Sovs and change their casting cost to 5, then screen-cast that game for a trailer.
You find Fallen Enchantress's AI to be "so bad"? So which strategy game do you think has a good AI? I can't beat it above challenging. I come across well constructed cities, good army coordination, decent use of spells and intelligent expansion when I play. Maybe it's just relative. So what game has good AI in your opinion that I should be comparing this game to?
We are not even in the same ballpark. Go play any modern 4x the ai is competent enough to at least expand at a reasonable rate. There are still "stalling" problems with the AI, it'll get to 3 or cities and basically turn itself off. I am not even approching the realm of whether its competent enough to beat a player I am talking it really can't even play the game. Like I said go watch the ai play itself with ctrl-z you will see some freaky behavior.
Go Play SOTS1 the ai will eat you alive, granted thats 6 years of evolved ai.
The thing about FE is that the AI is using a more complex structure. It will seem dumb in development, but once it is at 90%, it will simply be the best AI ever. This AI plays like a human, which means it will be a long time before it's perfect. The other difference is that this AI can learn and be taught by modders. That means new ideas and constant improvement is almost guaranteed. As long as the rest of the game is good.
Huh?
He's basically saying that when they are almost finish fixing the AI (at 90% done), it will be the best AI.
Ah, sucks that we are hoping for an AI that will be good later down the road if we mean after release. I am curious as to how it is playing more like a human and less like any other TBS AI.
Presumably it can perceive a variety of simple strategies that produce positive results, or variations on a strategy it already knows that produce positive results, and change variables accordingly. So that while remaining a rules-based AI, it knows how to adapt a bit.
Basically, every AI pretty much cheats nonstop. The devs do that and then focus on pathing and tactics. Those AI's seem better, but really they suck at playing the game and just cheat to make up for it. This AI plays the game and attempts to do so as well as a normal player would. If you don't follow other game's AI development, that won't mean much. It means alot to me though, as the only other AI that has achieved this is the GalCiv2 AI. As I recall that AI was alot worse at this stage of development. I am pretty confident this one will be at least at 90% perfection by release. The biggest issues right now are priority, which can be easily fixed when the balancing is complete. That is the hard thing about programming. You need the game to be finished before you can really set things up right. Changing just a few balance factors will totally change how the game is played.
This game does some other stuff that I find particularly groundbreaking. You see, the AI has many personalities and each one will play the game completely different. Every aspect of strategy and choice can be hooked up to a personality with different values for different play styles. So you can have a tech turtler, a warmonger rusher, a manic diplomat, a dark wizard and a slaver personality all play with drastically different style and optimization. The kicker is that now the user can define a personality and share it with others via mods, making the AI potentially limitless in variation and skill.
What sean is refering to is deterministic versus non-deterministic AI. If you take all the potential agency of a person or a machine in a given system you can come up with a set of possibilities or "events". I.E. an AI in the game can move its units, order buildings being built, etc just like the player. There are alot more events than that, especially once you add in diplomacy/technology/combat etc. Programtically you can almost reduce it to all the potential "clicks" you can make. I could build an A.I. that pushes all those buttons as the player would based off of 100% pure randomization I call a rand() function, each event in the set is assigned a number, number comes up the program "clicks" the button.
Obviously if you watched an AI that was purely random it would look schizophrenic so you have to give it some deterministic behavior pattern based off conditional logic and/or scripting. This means if you know the inputs the output will ALWAYS be the same.
IF TURN=1 SETTLE IMMEDIATE SQUARE.
That is purely deterministic ai behavior, or "scripted". Non deterministic behavior is you give the AI a list of potential events
IF TURN=1 {SETTLE,MOVE,RECRUIT,FIGHT,GRABGOODIE}
Maybe you could limit that down by doing a range check, for instance there might not be a FIGHT,RECRUIT,GRABGOODIE in range.
Eventually you will weight those decisions and then generate a random number, so maybe SETTLE will have a weight coming out to a 80% chance, etc.
If you start going through all the potential events, all the potential restrictions like range, and add in AI behaviors that change the weight numbers around you start realising how ridiculously large a process this is. Thats not to even mention all the probability and statistical math that goes into creating those weight values. Most game designers run straight back to purely deterministic/scripted/cheating AI.
You can also make the AI "evolve", that is save statistics at end game, gauge outcomes of choices then go back and adjust the weight given to those particular events based on success.
Then as well you can give the AI long term goals that will heavily change the weight numbers over a given number of turns or until an objective is accomplished (the paradox AI is pretty good at this).
So what Sean is really refering to is that Stardock makes its AI's far more non-deterministic. There really is no such thing a truely non-deterministic AI, there are also problems with true mathmatical randomization, especially entropy, and the potential of a given AI to learn through a evolutionary algorithm is subjective to the programmer(s). Stardock however makes a real effort to actually make the AI "play" the game.
Setting all those weights, rules, events, and making the algortihms genetic, factoring in all the variables (stuff like map size and average turns til victory have a heavy influence) and writing all the logic is a massive undertaking. Thats not even mentioning adding in all the underlying code that calls all the other core function of the game. Kerberos has the same kind of mostly non-deterministic AI.
That being said I have been watching the AI play itself now for a few patches and its pretty bad, Brad needs at least 6 months to get it to a point I would consider it on par with say a Civ 4 AI. Obviously just an opinion but the current AI does some really whacky stuff.
http://www.youtube.com/watch?v=8K-g7ipAqYY
Hence, Master of Mana is still more flavorful and engrossing than this
This is not 100% correct.
This is only an issue if you have scripted AI that uses build orders and other static decision making tools. If you have a dynamic AI which understands all the different game mechanics and has a system for assigning value to objects and situations based on a personality bias, then the AI should be able to play well regardless of balance changes.
If the AI was written well, it should be able to figure out that its faster to build a workshop and then build 2 pioneers than the other way around. It shouldn't follow a scripted build order at all. The AI should also figure out that getting the Tower of Erog early in the game is a waste of time and that focusing on civ technology is better in the long run.
I really crave to fight an AI that puts up a decent challenge by understanding the game but FE is not going to achieve that. The best we can hope for is a game thats fun and balanced, with cheating AI available for those who want to feel pressure.
I voted "good".
I won't list what I like because it's most of it , I'll just concentrate on what I'd like to see improved (or added), in no particular order:
- spells - in the current game I have 4 fire shards and the spell Pillar of Flame is ABSOLUTELY devastating on my enemies' armies that enter my territory. So much so that I've been wondering if that's a bug. It just obliterates them or leaves the army on a few HP. Is there any way the AI can counter that? Is there an enchantment to boost fire (or any other) resistance of a unit significantly and can AI learn to use it?
- spells again - Summoner - if I'm a summoner I'd like to have access to some special summons, not accessible through the normal spellbooks. Also, a few to start with and also that Summoner II, III etc actually mean a BIG improvement in summoned creatures, not just a few more HP and attack power increased by few points. Needless to say, summons should be tactical but also strategic (which means permanent) with some upkeep of course. Golems, drakes, living armours, flying swords, undead horde... there are many possible ideas to choose from. Let me (and AI) summon 3-4 bears who will level if we manage to keep them alive up in time and become REALLY scary. Don't restrict me to one shadow wolf who will get 2 AP for the Summoner II . If you think it's not easy to balance, increase the upkeep when they levelup, or increase the cost of summoning each successive monster.
- lack of food for city growth is annoying. It shouldn't be so much of a matter.
- in one of my recent games I played against Yithril who had Juggernaut armies (LOTS of them). That game felt great because I really had a sense of fighting a strong enemy. And I had to really choose my battles. I miss that with most of other adversaries since they just don't feel so different, aside from Resoln with spiders and magic. I'd like more of that , more of diverse AI armies which will have strong creatures.I countered Juggernauts with archers. Yithril should've countered that with armored troops but didn't.
- as someone else mentioned, champions who stay in a city don't progress. For admin champions, that's completely sensless. Even for "normal" champions. They should get some XP, less than battling but still.
- make the old monsters dangerous. If I leave an Ogre wandering around for whatever reason, let it progress, levelup, become a real city crusher (in time).If a lair matures for 100 turns, let it produce something terrifying.
- Wildland allies, Wildings, Darklings. By the time I build their settlements, they are almost useless. Make them upgreadable according to my tech level, or something like that. Why would I want an army of Trolls that dies to a single Shaman?
- building cities still feels boring. It's better, but still I don't see some REAL differences among Fortress, Town and Conclave. I know numbers prove me wrong since obviously Conclave is better in research if you build everything but I don't really FEEL it. I still build everything I can, just change the priorities. Decrease the number of buildings in various settlements (make it so that conclave really focuses on research and is unable to build city improvements connected to gildar or troops). Make it so that you cannot produce some troops in town or in conclave, or that your mages are significantly weaker if not produced in Conclave.
- Outposts. Too generic. Too :"I have to bulid an outpost to secure the resources". But it should be "I WANT to build an outpost because that will give me a fortified position in that area." Make the outpost contain a built-in guard force so that it won't change hands so easily or be destroyed if a monster stumbles by. Make the upgrades enhance that force. There are lot of things that could be done with outposts.
- more interaction with AI. Make it contact you, exchange information (for example map information). Make it evaluate and threaten (if you're kingdom and he's empire) or offer mutually beneficial deals if you're of the same alignment.
There's more but I can't remember it all now...
Newbie view, 28 hours played total, 8 hours with latest patch. Few things that I find might need some change
1) Cannot upgrade number of troops in a unit as tech unlocks become available. Pretty annoying as it takes some of the motivation to level up units and thus takes away from one key experiences of rpg's IMO.
2) Units that level up should be listed in the battle info log after the battle.
3) Lighly armored/armed units become useless as tech levels rise. Damage delivered falls too quickly to "1 damage" level. Full array of troop types should be used through the whole game timecycle. Some ideas (although I know this isn't going to change by release if ever) to fix:
3.1) Higher movement/dodge values for light troops
3.2) Different (nonlinear) progression of damage delivered
3.3) Flanking etc (more attacks to target lowers it's armor abilities) bonuses to increase the meaning of numbers over quality
4) Want Random Spawns. Cleared areas are too empty.
Overall I gave "Good" as the game is just a few inches from excellent. Progressions over the last month has been amazing though.
The UI is a disaster for consoles too.
And yes, the modding community did step up to it http://skyrim.nexusmods.com/mods/3863
But that doesn't mean that this wasn't a flaw in the game.
It didn't suck, it coupled the worse UI ever made for an AAA title with a bunch of other flaws and some really good pluses like the hand crafted dungeons and open world and the awesome-cool dragons. The sum is greater then the parts and you ended up with a good game that contain flaws that historically have proved fatal to many games. Skyrim survived those fatal wounds.
Although honoestly I never finished the game, I stopped playing at the "peace summit" where the plot just got too godawful for me to stomach
Great post, every one of your points deserves attention.
There are many great features available to you once you register, including:
Sign in or Create Account
