V19 Experimental Build 4 available!

[Mytheos]

@Chris

I know this isnt the place for it, but since this has derailed to this point...

Maybe you could have a Map Contest, where every Month/Week you add a Community Made map to the experimental builds that was voted on by the Community.

Might create a spark for more maps to be made, and might increase user awareness of community map usage.

[thothkins]

The community map pack is full of mapping goodies, and a number of these are being put into the final game. I think there was quite a long hiatus of community map building due to a wish to get to Beta where they would have more chance of being final versions. But there's progress being made.

[ThunderGr]

As I said the UFO Series (Aftermath, Aftershock, Afterlight) had random maps and you could select landing locations. So, nice try, Chris, but modern games have already done it.

[Chris]

Afaik, they are not tile-based games like the original X-Com and therefore it's nowhere near as problematic to do it for them. They use a 3D engine where units move more freely rather than from tile to tile, as RTS games generally do.

Using square tiles is the biggest limitation regarding randomisation in maps, as both us and Firaxis have found.

Even if I'm not correct about that, being so unrelentingly negative about everything we do really isn't achieving anything. If it's as easy to make a perfect remake of X-Com as you appear to believe, the success of this game and Firaxis' title suggests that you could do very well out of doing so.

I strongly suspect what you'll actually find is that what you believe to be very simple solutions will have ramifications for the project elsewhere, and the reason why people have made "incorrect" choices is because that not making them would cause even more issues.

And even if I am wrong, you can point it out without being sarcastic. I'm not quite sure why you believe paying $20 for a game entitles you to be rude to the developers?

[ThunderGr]

Even if I'm not correct about that, being so unrelentingly negative about everything we do really isn't achieving anything. If it's as easy to make a perfect remake of X-Com as you appear to believe, the success of this game and Firaxis' title suggests that you could do very well out of doing so.

I am only being negative when you do things that simplify the game and make it less challenging and less interesting. Is it my fault that this is what you have been doing lately?

OK, If the UFO Series does not cover you, how about the Civilization IV and V? The use tile movement. They have random maps.

I understand the problems but, from what Max has said about random maps, you only need to improve the validation step in the random map generation process. Meaning, you are almost there, why not finish it?

BTW, the O.G. addressed the buildings problem by providing a fixed set of buildings that were used during the map generation. People do not mind that, because it is common sense that building architecture is similar.

I strongly suspect what you'll actually find is that what you believe to be very simple solutions will have ramifications for the project elsewhere, and the reason why people have made "incorrect" choices is because that not making them would cause even more issues.

They are still incorrect choices. Sorry. Avoiding the problem is hardly ever a solution.

Edited July 14, 2013 by ThunderGr

[Went]

The original X-Com had very basic graphics, and if we used very basic graphics and had no logic in our map generation then we'd be able to use fully randomised maps.

Nobody is talking about completely random maps. Original XCom has big tiles (farms, buildings, lakes, parks etc) of same size. And these tiles were combined in random order to obtain pseudo-random environment. And of course, these tiles has generalized constraints to avoid inadequate combinations. Can you say to me as "game programmer to programmer", why this is impossible in Xenonauts game?

[ThunderGr]

Original XCom has big tiles (farms, buildings, lakes, parks etc) of same size. And these tiles were combined in random order to obtain pseudo-random environment. And of course, these tiles has generalized constraints to avoid inadequate combinations.

...and that. But I thought it was common knowledge...

[Chris]

Went - I never said that was impossible, simply because it's actually already possible in our game. All you'd have to do is make sure all the the sub-maps had ground edge textures that matched up (for example, farm fields all had grass edges), make them all the same size, then put them in a single folder.

Once you've done that, you can just paint an entire map with that "tile" and the game will automatically randomly fill each of the "tiles" with a random sub-map from the folder.

In actuality, the effects are a bit ugly as they don't seem to have much logic to them. As I said in the previous post, any kind of urban area tends to have more "logic" to it than you might think and the randomised versions look unconvincing. We decided to use semi-randomised maps instead, so they had a more convincing appearance.

ThunderGr - from my point of view, it looks like you're being negative for the sake of it. It might just had been the sarcastic "nice try" that made it come across like that. It's annoying to take the time to type a perfectly reasonable explanation of the issues at hand and then have someone be rude in reply.

The answer to your question is above, in my reply to Went - that is actually already possible in the game but the end results look crap. I wrote that in my original reply.

Re: the "incorrect decisions", you're missing the point. If we use the After*** games as an example, we could presumably have the same randomised map system as they use if we weren't a tile based game. But then we wouldn't be a tile-based game, and X-Com veterans would (rightly) be complaining that we'd changed a fundamental aspect for only marginal gain.

I put "incorrect" in quotation marks because in this example, the decision we made there wasn't the incorrect one, but it may look like it to someone that isn't familiar with how every part of the game works and isn't aware that it's a choice of nice-looking map randomisation or tile-based mechanics. That leads to a situation where a developer ends up being criticised for making the correct decision, and if you look at it from my point of view for a moment, you can probably see why it's a frustrating position to be in.

The Civ games are at a drastically different scale (a square is 1.6m in Xeno, it's a whole city in Civ), so don't suffer from anywhere near the level of edge blending issues as you get at a lesser scale.

[ThunderGr]

ThunderGr - from my point of view, it looks like you're being negative for the sake of it. It might just had been the sarcastic "nice try" that made it come across like that. It's annoying to take the time to type a perfectly reasonable explanation of the issues at hand and then have someone be rude in reply.

Not my intention. I have played enough tile-based in my life to know you can have perfectly playable random maps. The level of "convincingness" depends on the tile-blocks you use. In fact, there are a few posts over the net that provide basic information on how to make a random map generator for tile-based games. here is an example.

Besides, I have already said, "you have nearly done it", so, why disregard this piece of positive appreciation of your work?

Re: the "incorrect decisions", you're missing the point[...]

I put "incorrect" in quotation marks because in this example, the decision we made there wasn't the incorrect one, but it may look like it to someone that isn't familiar with how every part of the game works and isn't aware that it's a choice of nice-looking map randomisation or tile-based mechanics. That leads to a situation where a developer ends up being criticised for making the correct decision, and if you look at it from my point of view for a moment, you can probably see why it's a frustrating position to be in.

I took that part of your post as a more general reference to the game development, this is why I phrased it like that. Still, I do not think that avoiding to solve a problem is a better solution. Why give up on it when you are so close?

If you make the edges uni-blend for each tileset, you will avoid all blending problems that come from the fact the graphics are more detailed.

One of the major problems, as pointed out by Max, is that the UFO/Dropship tiles can be overwritten.

Your position is a difficult position since you always have to be the target of frustrating comments and unsatisfied players. It needs a lot of strength to do it and I do not underestimate that.

I, for one, understand you perfectly. Still, as a player, I do not wish this game to have the fate of UFO: ET, nor this company to have the fate of the company that created said title. You are deeply involved in the development problems and you try to make it to release by minimizing said problems. At the same time, the solutions you pick can make a huge difference in game experience and people will complain about it. Managing to handle the compromise without failing to keep your fan base is the challenge you have to meet. A hell of a job, for certain but this is how things are, the way I see it.

Now, since I am not part of the development team, I cannot know the details for the specific issue. Still, I cannot but point out that the random maps were given as a feature in the advertized game. And this is what the whole fuss is about.

@Max

Yes, you are absolutely right. Only UFO: Aftermath had the random maps(no, not the part about you being a pain....). It was an unjustified generalization from my part, I am sorry. They had the map-validation step where they were checking for path continuity and adjusting procedure if there were deadends.

As for the X-COM validation, it is obvious that, since we never got deadends or unplayable maps, there was a validation step. Never occurred to me a map that had my dropship at the top of trees or in a building. Nor the UFO ever spawned at a similar invalid location.

Apart from all of that, I want to point out that the dropship/UFO relative locations were completely unknown. It could be that you landed directly next to it or at the other side of the map or anywhere in-between. That is awesome. That made the shroud count .

Edited July 14, 2013 by ThunderGr

[Slish]

Hello,

First of all, I love this game. And been playing it a bit. It certainly looks great and had some fun with the game already.

I have 1 question though, is the "Hidden Movement" screen fixed in 19.4 or not? It mentions it in fixes, not sure what it means though.

For me nothing seems to have been changed compared to 19.3.

Fixes

- Camera during AI turn should no longer focus the "hidden movement" screen on the location of the "hidden" movement

- Camera should also not jump around so erratically when following shots (more camera changes to come, not done yet!)

Nobody seems to mention it in here as well...

[Hadrian]

Wait what? Who said the UFO series had random maps? It didnt, and never did. I have every game right here and its bespoke maps in every one of them. In fact I know pretty much every single map in Aftershock and Afterlight like the back of my hand and have them memorised because they are not random.

[Simmo753]

no hadrian the original series before those games had random maps mate.

[ThunderGr]

Wait what? Who said the UFO series had random maps? It didnt, and never did. I have every game right here and its bespoke maps in every one of them. In fact I know pretty much every single map in Aftershock and Afterlight like the back of my hand and have them memorised because they are not random.

UFO: Aftermath had them, Hadrian. I know. I have finished it dozens of times.

I also have Aftershock and Afterlight, which did not have random maps. It was a slip of tongue and I apologized.

[Simmo753]

i got upset with the after*** series and only played the first one alittle before giving it up. never played the others.

[Selvhan]

Keep on the good work ! It's nice to see new build every 2 weeks or so !

[jazzglands]

They are still incorrect choices. Sorry. Avoiding the problem is hardly ever a solution.

I used to believe as you did, before I became a professional programmer who had to actually sit down and design a complicated project. It's clear you are at best a student of the field and at worst an armchair designer with no experience or expertise and a greatly inflated sense of his own knowledge.

If I am wrong and you do have this level of expertise, then I hope you make your own X-COM clone. For a developer of your obvious skill and experience I would expect it no later than September of this year. I look forward to enjoying it.

[4Aces]

UFO: Aftermath had them, Hadrian. I know. I have finished it dozens of times.

I distinctly remember seeing the same maps over and over. I could probably mock up a few diagrams of them right now, but that is too much work so I went searching: here is a link. They devs claimed the maps were random, but they did not actually mention that it was just some placeable objects and aliens that were random.

As for your link to procedurally generating 2D tiles, I am afraid that while there is some useful information there it does not apply to these tiles. The two games that were discussed within two links of that page, were exceptionally basic art (and the rest were 3D). What this amount to is comparing the graphical style of Xenonauts to a stick man. The problem is not that the tiles cannot be randomly generated, it is that they need to be specifically placed, or else they will look weird. If this were more like the demo that Chris just posted the pic of, then it would be easy to do. Just make sure that a random cabbage is not in the middle of a street or wall, and you are good to go (well not quite that easy but relatively close to it).

Xenonauts uses highly detailed bitmaps and they require blending where dot art or procedural (aka bland) bitmaps do not. Imagine a straight line with two tiles. Now put grass tile A beside grass tile G, and you will see a seam where they meet as the bitmaps do not blend across them as a whole. Only tile A can be placed beside tile B to get no seam. The hard part comes from the fact that only tile G may have an edge border that will look appropriate to start a new tile type (or it may even require a specific type of tile). So you have to have the entire grass cube in a specific order, and it may dictate what kind of tiles can go beside it (probably not, but it is possible). So you can have a cascade effect that takes over the map, and large sections could be self-forming from a single tile placement. That is what Max_Caine was talking about. Semi-random maps that require a boatload of work to get the code just right.

It is far easier to make large modules (e.g. 1/16th of the map) that are combined either manually, or procedurally, and then layer all the placeable objects on it. Considering the modding tools that allow users to create their own maps, and the fact that the game will have a semi-randomized map selection, there will be far more maps than Aftermath ever had. Having manual modules would also allow tweaking of existing maps very easily. Just clone the module, rename, alter, and swap it on the map (or make it into a new map). Once semi-random alien placement is also nailed down, it will be highly acceptable.

Thanks for the link!

[DNK]

RE: Random Maps

And, yeah, I'm talking about "random modules/submaps", not fully procedural maps. I may have been imprecise in my earlier wording.

I think the devs are overvaluing how "totally realistic urban spaces" are important (sure, they are), and severely undervaluing how "replayability" and "the excitement of not knowing what the map is like" are important. I find next to NO replayability from non-random maps and they severely break immersion (far moreso than "unrealistic cityscapes"), and this is a big reason I straight skipped the Firaxis remake.

For single-player games, map randomization is a huge selling point, and easily makes or breaks a decision for me. I'm glad we have community maps that are partially random (which is sufficient for me to consider replayable), but I'm also sick of having to wade through countless repeats of "official" maps, which just bore and tire me of the game. The devs need a LOT, lot more non-random maps to overcome the issues of "map fatigue" for players, and it would take a lot less time to plan out some semi-random maps that avoided whatever they think is "unrealistic" for the cities/farms than to make like 40 variants of each map type.

RE: Snipers / Ground Combat

3) Scout/sniper rather than being "diversity of tactics" is "the only show in town worth seeing" if you can see across the whole map. And it's broken/OP for a sniper weapon to be good at being long range?

Broken for it to be amazingly better than anything else at killing in geneal, yes. These aren't "snipers" so much as "designated marksmen" rifles (going by the picture). They should be better than normal rifles, but not so much as to totally unbalance the game in their favor. I would suggest:

Make

two separate sniper rifles

: one with the same damage/recoil as a rifle but with increased range (at a far higher TU cost), another with much higher damage and increased range/TU but also a ton more recoil/weight (so that they are fairly inaccurate for soldiers to use unless they are high strength). The latter needs someone who has both high strength and accuracy to wield, which (look below), should NOT be that easy to come by. The former seems better, and will be on open maps, but at the same time is really weak in shorter range situations where it can't fire as much and can't burst fire, so if you get a "random" map with a lot of CQC, your sniper squad just became weak as hell. This forces balancing on squad loadouts and severely reduces the OP nature of snipers currently.

Also, the former is basically a 'scoped' version of a rifle, so it should

cost

the same, while the latter (for any tech tier) would require a whole reworking of the weapon to produce high-powered, high-accuracy shots, so the weapons should be much, much more expensive to boot (and with far reduced ammo counts, moreso than currently).

Also,

limit soldier progression

a bit. A big reason snipers are so OP is that you can end up with soldiers that are really OP, with both the strength to handle any weapon and the accuracy to hit a penny at 2km. Stats need stronger caps, with soldiers have a wide randomness to their max caps, so you end up

having

to specialize them, instead of being able to plop a CMD or CLN into anything you want with everything you want, being able to hit anything you want at any range, and being able to run a mile and shoot 5 aimed shots in one turn because they have infinity TUs.

Also, in general, make

weapons less accurate

so that

tactics

(flanking, advancing under covering fire/smoke) become more important than having 5 snipers in a line 20 tiles away playing "firing squad" with the enemy (which is unchallenging and boring and unrealistic and unbalanced). Even with a "sniper", you shouldn't be able to get 80-100% hit rates at 20 tiles.

These are all things that devs can change to rebalance the game and really create a much better ground combat game, which seems to be their focus anyway.

Oh, and of course making an actual battle AI for aliens that has them using cover instead of running in circles in open fields when under fire would go a long, long way.

Edited July 15, 2013 by DNK

[GizmoGomez]

Ooh, limiting the max cap on a per-stat basis sounds interesting. Maybe making it so that strength, reaction, and accuracy were randomized? TUs would increase for everyone, I'd think, and so would health and resilience. Well, maybe not resilience. Put a random cap on that too, I guess.

[DNK]

I think this is how it was in XCOM, but if not, whatever it's what I want (in more detail):

When a soldier is generated, two numbers are generated for each stat, the "starting stat" and the "max stat" (additionally, you could have a rate of progression randomized for each). You would want that, mostly, the starting stat and max stat would be related, so that a high starting stat gives a high likelihood for a high max stat, but not always so. The starting stats would, as currently, all be within certain fairly tight bounds (tighter than XCOM's, and higher on the whole). The max stats would be more variable, though. So, someone might have a max strength of 67, while another might max out at 82. Same for others.

That with a reduced (by half at least) progression rate for stats would make the game more challenging and negate the fact that open maps allow "supermen" to just mop up aliens in the open without difficulty or risk.

There are other ideas in the forums I like, like reducing progression per mission but making it easier (so you don't have to do lots of little bs).

[Chris]

Yeah, I'll speak to Aaron and we'll have the soldier progression reduced a bit for the next build.

We'll probably limit gains to 1 point per mission in each stat (currently 2), but make it twice as easy to get that one point as is currently the case. We'll also reduce all the stat caps to 99, from 100 for most stats at present or 120 as APs.

[Skitso]

I think this is how it was in XCOM, but if not, whatever it's what I want (in more detail):

When a soldier is generated, two numbers are generated for each stat, the "starting stat" and the "max stat" (additionally, you could have a rate of progression randomized for each). You would want that, mostly, the starting stat and max stat would be related, so that a high starting stat gives a high likelihood for a high max stat, but not always so. The starting stats would, as currently, all be within certain fairly tight bounds (tighter than XCOM's, and higher on the whole). The max stats would be more variable, though. So, someone might have a max strength of 67, while another might max out at 82. Same for others.

That with a reduced (by half at least) progression rate for stats would make the game more challenging and negate the fact that open maps allow "supermen" to just mop up aliens in the open without difficulty or risk.

There are other ideas in the forums I like, like reducing progression per mission but making it easier (so you don't have to do lots of little bs).

Soldiers could maybe have one additional 'progression' or 'potential' stat that would show how much and how fast one could gain experience?

[4Aces]

From the manual: "Soldiers gain experience in battle and level up their attributes (they gain stat points by using those stats, so for example you gain Accuracy points by firing your weapons at aliens)."

With reduced requirements to get the stat bonus, you will find that they just come from regular combat. Currently you have to 'train', by clearing most of the map, then having everyone run back an forth (for TUs), or shoot at stationary objects (Accuracy), which is what DNK was referring to.

[DNK]

Ultimately, just keeping the stats lower would be super. That would be my #1 priority. I mean, really, how much better are experienced cream of the crop soldiers going to get fighting a short campaign? Are the Xenonauts hiring fresh recruits that double their abilities in a few months?

[Nickboom]

I say leave ap around a 100 or the current number because at 99 you some really weird moving