Gazz wrote: Well, no. It is flexible but it kills variety.
You play with the numbers, then build / research a single item for the exact purpose you have in mind.
Then you're done with research. Because you have the optimal item.

Implementing research in a way that makes research pointless in the minimum amount of time doesn't sound like a great idea to me. =)

Item could never become optimal because other player/AI can research equivalent 'optimal' item so you should go further and as I said before research will not be cheap (at least it should be more expensive then equivalent procedural node). Also each subsequent improvement should be more expensive then the previous one.

On the jack-of-all-trades question:

Ixos wrote:Doesn't the system lined up actually encourage jack of all trades seeing how what you unlock is random?

It's not entirely random, though. As Josh described it, although what unlocks is somewhat randomized, you still get to choose which unlocked techs to actually research. So, similar to Josh's simplified example in his devlog, a player who wanted to focus on kinetic weapons would be able to immediately point their research (once they're big enough to be able to do research) to Weapons and then on down to the fixed and modifier techs that yield really powerful/small/cheap kinetic weapons.

That's fine by me. I agree with Josh that that seems like a workable balance between player control and surprising options.

What I'm not seeing yet is the value of biasing the semi-random unlocks in favor of whatever you've been researching the most. As Josh put it:

There's a final bit to the system that I want to quickly go over, and it's the bit that rewards specialization and prevents jacks-of-all-trades. It's also quite cool Over time, the procedural unfolding of your tech tree becomes more and more biased towards your previous choices.

I agree that this would promote specialization. What I'm asking is whether that's necessary or desirable.

I'm not sure it's necessary since the players who prefer to specialize can already do exactly that using the basic unlock/discover mechanic described.

And I'm not sure it's desirable because I don't see the value in penalizing a completely reasonable playstyle. As I said, what's wrong with preferring a balanced development of capabilities? Is there some way in which choosing that path would necessarily cause Limit Theory to be less fun to play?

The one way I can see in which that answer might be "yes" is over the course of a very long game. Given enough time, the JOAT approach will wind up being quite good at everything.

But if I've had fun getting there, why is that a problem? Maybe at that point I'm so effective that I might as well start a new game... but wouldn't a specialized player have the same problem, and likely even sooner than the JOAT player (because they've deeply explored only a subset of the base techs)?

I understand the allure of the idea of biasing future unlocks by past choices. "You get better at what you do a lot" is pretty much the basis of Bethesda's RPG system. I'm not opposed to that -- what I'm questioning is why that shouldn't also apply to "doing a little of everything."

Why not bias unlocks for the generalist player so that they more likely to get unlocks in parts of the tech web they haven't yet explored?

Flatfingers wrote:Why not bias unlocks for the generalist player so that they more likely to get unlocks in parts of the tech web they haven't yet explored?

Surely that's not what a generalist would want. A generalist would want the greatest chance of developing their tech tree in an unbiased way, and if they select research nodes in a round robin fashion then they will neutralise Josh's learning algorithm because they won't introduce any effective bias into the research dice role that could be used to infer player specialism.

On the other hand, a player who likes to take a punt on long odds and serendipity would be interested in a dice role biased in favour of unexplored tech branches.

Sorry.. haven't read the total 63 posts

I did just read the recent dev blog though on the topic of research tech etc

One thing I saw there was that you just select what you want to research and it's like it does it all in the background. You just wait for some amount of time and then discoveries are made without any involvement by the player.

What I would really have liked to see is that certain tech research requiring certain materials, minerals and other things found within the game. Just like in real world and sci-fi universe research, they need resources to do the research. During the beginning techs, this wouldn't be much.. but as you go on, the required materials and stuff gets more and more specific as well as rarer (rarer doesn't mean harder to find).

When Josh spoke about research forcing exploration, I thought that he was alluding to this need to go "out there" in order to find these required resources but his dev blog didn't really talk about that. Maybe you could trade for some of the materials, maybe you could find some ancient civilisation that had them, maybe you could find it in some asteroid belt and so on. My point is that it shouldn't really just be as simple as clicking a button and waiting for a certain time for discoveries to be made. It should require some level of resource supply as well.

Zero Gravitas wrote:A generalist would want the greatest chance of developing their tech tree in an unbiased way, and if they select research nodes in a round robin fashion then they will neutralise Josh's learning algorithm because they won't introduce any effective bias into the research dice role that could be used to infer player specialism.

Well, speaking as just one generalist, gaining access to parts of the tech web beyond the bits I've focused on sounds pretty good to me.

I may be misunderstanding how you're thinking of the term "round robin," but given the number of fixed techs implied in Josh's simple (!) example I wonder if a round robin algorithm wouldn't wind up being more like a fractal: you never unlock more than one or two of the Big Techs because there are so many specialized techs below them to loop through. Or maybe you're thinking of looping through the Big Techs first, then their first-level sub-technologies, and so on?

If the latter, I don't really see a conflict between that process and "across-the-web" unlocks. If my pattern of research is general, then a bias toward greater randomness in which nodes are unlocked would feel useful to me. It would mean I'm getting to see more of the big picture of the complete tech web sooner, rather than a specialist's bias toward deep unlocks in just one corner of the web.

But I don't even need a bonus for generalizing, really. I just don't want to be penalized for it. Heck, maybe that interest is satisfied with a startup option: "Don't tend to unlock techs similar to what you've already discovered."

I'd rather have a benefit for playing as a generalist, though.

light487 wrote:What I would really have liked to see is that certain tech research requiring certain materials, minerals and other things found within the game. ... My point is that it shouldn't really just be as simple as clicking a button and waiting for a certain time for discoveries to be made. It should require some level of resource supply as well.

I could support this as long as it applied to what I called "object" techs, and not "concept" techs.

It sort of makes sense that you'd need uranium to figure out the applied technology of Fission Reactor, or antimatter to discover the tech to build Jump Drives, etc. In fact, needing specialized resources to complete an object tech discovery could create a nice interaction with physical exploration and with strategic level play such as fleet operations, economic development, and high-level diplomacy. If you define key research materials as resource nodes -- planets or really big, unique asteroids -- then you need to apply all your faction's power toward securing worlds with the resources you need for the new technologies you want to research.

This would complicate Josh's nice, clean, simple research mechanic. It would also mean needing to make sure the universe generation algorithm didn't hose the player by putting all the useful resource nodes in the territory of your worst enemy.

But maybe the payoff of integrating research more deeply into multiple areas of strategic play makes the idea of needing resources for applied research worth considering.

Flatfingers wrote:I don't really see a conflict between that process and "across-the-web" unlocks. If my pattern of research is general, then a bias toward greater randomness in which nodes are unlocked would feel useful to me. It would mean I'm getting to see more of the big picture of the complete tech web sooner, rather than a specialist's bias toward deep unlocks in just one corner of the web.

We are saying the same thing. But what I'm trying to convey is that you don't need a bias towards randomness if your choices of research are so broad and varied (as they would be for a generalist) that you aren't introducing any bias into the system in the first place. What Josh is proposing seems to be a fairly basic feedback amplifier – based on relative weighting, make it more likely you'll discover more of the thing you have shown a bias towards previously. If you don't input a bias by repeatedly choosing to research one specific type of technology, there won't be any bias in the output - it will be a random dice role.

Understood, Zero Gravitas.

Now please explain to me how to be sufficiently random that an algorithm looking for patterns doesn't see any.

I'm just having a little fun and not really being snarky. It just seems to me that an option to disable research biasing would be rather safer than hoping code doesn't get confused.

The penalty for generalization is that you're okay at many things but really good at nothing.

Flatfingers wrote:

JoshParnell wrote:There's a final bit to the system that I want to quickly go over, and it's the bit that rewards specialization and prevents jacks-of-all-trades. It's also quite cool Over time, the procedural unfolding of your tech tree becomes more and more biased towards your previous choices.

And I'm not sure it's desirable because I don't see the value in penalizing a completely reasonable playstyle. As I said, what's wrong with preferring a balanced development of capabilities? Is there some way in which choosing that path would necessarily cause Limit Theory to be less fun to play?

How is it penalising you when you are getting more of what you kept choosing over and over?

It may take more effort to go into the completely opposite direction when you have preferred "heavy" technologies for the last 200 picks but I don't believe that a bias would ever be heavy enough to completely bar you from researching something else.

An often-discussed topic with X3 is damage per shot vs DPS. I am convinced that this is a religious matter given how strongly people insist that one is far more important than the other. =)
If you went for "high damage per shot" items every time, wouldn't it be an advantage if the game didn't keep wasting your time by offering something you clearly don't want?

Gazz wrote:How is it penalising you when you are getting more of what you kept choosing over and over? ... If you went for "high damage per shot" items every time, wouldn't it be an advantage if the game didn't keep wasting your time by offering something you clearly don't want?

Again, I have zero objection to that. As I said, I like the idea of a game being designed to give players more of what they like.

But what if balanced, jack-of-all-trades play is what some players like? That's why I'm hoping "rewards specialization and prevents jacks-of-all-trades" will get some further consideration.

Certainly no game can be all things to all possible people. But that has never been a good argument against looking for simple ways to satisfy more kinds of players as long as those features don't degrade anyone else's preferred style of play.

Grumblesaur wrote:The penalty for generalization is that you're okay at many things but really good at nothing.

True. But being at least okay at many things also is -- or should be -- the great benefit of being human.

Robert A. Heinlein wrote:A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects.

At this point I feel like I've made the case for not "prevent[ing] jacks-of-all-trades" play about the best I can, and I don't want to belabor the point, so I'll leave it at that.

I'm curious what folks here think about the other comments and questions I raised in my laundry list. In particular, what about the idea that technologies should be about more than just shipbuilding?

You =/= your machines, unless you're the protagonist from Max Barry's "Machine Man."

Sure, it is a great ideal to want to be excellent at everything, you will notice when you walk outside and meet other people (or even simply click on the TV or peer into the mirror), you will see individuals who are absolute garbage at programming computers, but cook the best damn pancakes you've ever had. You'll see individuals who can solve a polynomial equation in their head without batting an eyelash, and people who suck at math, but can build a weather-resistant lean-to of leaves and sticks in a few hour's time.

Nobody is good at everything. From a philosophical standpoint, maybe everybody should strive to be good at as many things as possible. Or rather, as many things as are pertinent to their personal situation. But again the root problem is that we're talking about equipment. We're talking about guns, engines, shields, and other things. Demanding that a person be good at everything is akin to demanding that a gun tell you the time of day, cook you breakfast, and heal wounds.

Beside that, if you could go jack-of-all-trades and not have some penalty for it (i.e., being exceptionally mediocre) and then have a vertical progression where every aspect of your jack-of-all-trades turns you into a king-of-all-trades, then you've officially made the game boring.

As a fictional trope, characters who have remarkable abilities with no penalties that they have to deal with (and thus are not subject to character development) are called Mary Sues/Gary Stus, and anyone will tell you that Sues and Stus are boring. The game will lose its longevity if jack-of-all-trades is a legitimate thoroughfare to success.

I like Skyrim's skill progression, personally. You are rewarded for using skills with increased abilities in those skills. You don't have to specialize, but trying to keep all your skill points spread evenly is detrimental to your success. Of course, dumping them all into one skill is equally unwise, since you might run into a boss that's exceptionally weak to fire magic, but you've only ever touched two-handed weapons in your entire playthrough.

I like the idea of (partial) specialization being an important aspect of gameplay. In Skyrim I use one-handed weapons and usually restoration or destruction magic. I use heavy armor. The advantage is that I can hit really hard with a mace and keep myself healed with a glowy ball in my other hand. That's more than enough to get me through most fights. If for some reason I decide that archery and stealth is the way to go, it would be better for me to start that in another file, because I'll be compromising myself in using a weapon type that is not my character's forte.

You don't have to have it all in one save file, and why should you, anyway? If you get the entire experience of a procedural game in one go-around, doesn't that defeat the purpose of it being infinitely variable in the first place? I support specialization mechanics because they add to the replay value of a game, which means that the game is more worth your money in the first place.

If you try to play with all your birthday presents at once, you'll make a mess and possibly even break something. One thing at a time.

Much of this discussion appears to centre around how to interpret Josh's use of the term "Jack of all trades". It's possible he meant "Master of all trades", which at least to me has a signficantly different meaning.

A "Master of all trades" outcome could make the game boring; but I think that's really in the eye of the beholder. To some players however that may just be the "end game" (about which we know little when it comes to LT). In any case, I wouldn't have a problem with it as long as it took a really long time to get to that place. In any event, given the freedom Josh is aiming for, it seems that with enough cash you could setup research capabilities all over the place that are specialised centres in their own right: one each for the major "fixed" nodes. This would make it highly likely that a master of all trades will crop up eventually regardless.

I guess my view is that this is a game balancing question really, and only knowable as a result of a play-test by a large group of players with different styles.

I won't do a full, in-depth response at the moment, but just quickly touch on the subject of specialist vs. generalist, which I suppose I worded poorly in my log.

You are not punished for being a generalist insofar as you will lose research efficiency, or anything like that. Only that you won't have time to spread yourself everywhere in a system that is as...broad as this one. In other words, it's not like a simple 5-skill system where you can easily dump an equal amount of time into each of the five skills. Since the tech tree has an unbounded depth, depth and breadth will always be at odds with one another. And that's what I really like. It means you can be as much of a specialist as you want (imagine drilling 20 nodes down into a specialized propulsion technology, just because you really like traveling in style). But of course, it does not prevent you from being a generalist. It's just that being a true generalist in a system with such a high branching factor is going to mean spreading yourself very thin.

Contrast, for example, with TES, where we generally end up with uber jacks-of-all-trades by the end-game (especially in Skyrim ). Probably because the level of specialization is quite low. There are a good many skills (more so in Morrowind, and I would say that I am much more prone to having a specialist character in Morrowind), but it's always just a linear progression and the depth isn't that high. Imagine if, at level 50 conjuration you had the option to explicitly delve into "necromancy" vs "elemental conjuration" vs "supernatural conjuration" vs "inanimate object conjuration." Skyrim beats around that bush with the perk tree but it's not really much of a choice. In the end there are so few nodes and they tend to converge into some master node, rather than further branching to allow more depth / specialization. It's just a beefed-up linear progression. In that kind of system, being a generalist is a viable - and in fact usually the best - strategy. In LT the progression looks a lot more exponential and a lot less linear (because it's a real "tree" of tech progression, not just a few linear chains). So being a generalist will mean a lot of time spent in the shallow nodes. That's not a penalty, it's life!

But. I still haven't made the most fundamental, critical point, which is this: from a "value" standpoint it makes no difference whether you specialize or not, you are not explicitly "rewarded" or "penalized" either way (sorry if I worded it poorly in the log). The speed of your research is not affected by what you choose to research, so in the end, you are always performing the same amount of "work" on your tech tree, regardless of whether that work is spread over a huge number of branches or a single, deep branch. Think of it like conservation of energy. Conservation of research power! This means that as a generalist, ultimately you will still have the same value acquired through researching as a specialist. But! As a generalist, your value will be spread across many items, which means it is likely that none of them will be as "good" (assuming some degree of vertical progression) as those of a specialist, who has funneled all of his research "energy" into a few select lines of research.

The whole bit about the research tree becoming biased towards previous choices is just a mechanism to help you explore more of what you want to explore. It does not influence the "value" produced by research.

There's really very little that's "gamey" about this system, and that's part of why I like it so much - it feels very reflective of reality. If you choose to study many fields, it is true that your energy will be spread among them, as opposed to focusing on a single field. Simple math And as for the research bias, it makes total sense - I'm spent my life studying graphics programming, so it's much more likely that I'll have a great idea related to graphics today than...that I'll have a great idea related to biochemistry Just makes sense!

JoshParnell wrote:... vs "inanimate object conjuration."...

Yes, that's right. Imagine being able to summon giant mushrooms to confuse your opponent. Or perhaps mountains of sweet rolls to distract and tempt them. Or rocks!! "Summon Alfred the Lonely Rock". Now there's a spell I would buy.

JoshParnell wrote:In LT the progression looks a lot more exponential and a lot less linear (because it's a real "tree" of tech progression, not just a few linear chains). So being a generalist will mean a lot of time spent in the shallow nodes. That's not a penalty, it's life!

Isn't it still true that you will have diminishing returns the deeper you go into each branch of specialiation?

This should according to my logic mean a generalist can pick up more of the "easy" shallow bonuses across the board and gain a higher total %modifier boost but spread thin across all areas instead.

In my eyes it's a very fair trade off and balanced system