Return to “Suggestions”

Post

Squaring the "Vertical Progression" Circle (maybe)

#1
So there’s been a lot of fooferaw around vertical scaling of tech. I’ve been umming and ahhing over it, and these thoughts aren’t completed, but here’s where I am right now.

A brief summary of what I think are the points of conflict:

Problems with Vertical Progression
  • It creates a treadmill and a need to grind.
  • It can leave you outdated if you don’t research or constantly upgrade.
  • As research trends to higher, the numbers start getting silly. Even with scaling, it’s somewhat dubious to assume that a Level 500 laser is really doing 600x the DPS of a Level 1 laser, and Level 500 armour is really absorbing 600x the DPS. You’d anticipate that there really would be some natural law of diminishing returns.
Problems with no Vertical Progression
(assuming some scaling mechanic - unscaled vertical progression is just Increase T to Win)
  • Markets will become saturated much more quickly without the need to upgrade.
  • ”growth” is necessarily horizontal - i.e. more territory - because there’s nowhere to go upwards.
  • The mechanics for new companies “overtaking” small companies become those of warfare rather than economics. Market Share can’t be won on products because products across companies are essentially uncompetitive. This leads to protectionism, cartelism, actual military fighting to secure markets etc. While replicating 17th century naval empires is certainly not an invalid way to play, it is rather limiting.

TL;DR: limit vertical progression to areas of productive capacity that affect big industrial players, keep it out of the way of smaller Hardenberg players.

The thing is, it appears as if vertical progression makes sense for some technologies, and doesn’t make sense for others.

Weapons, shields and armour - military tech, in other words, seem to be where vertical progression really upsets the Hardenbergs and messes with the gameplay.

Conceptually these things seem to be realistically capped at a certain limited range of inputs and outputs. And also these are the things that most bother people in their constant need to upgrade: “What if I go away for a year and then I get back and people can one-shot me?” Valid concerns - scaling helps, but doesn’t fix “why do I have to upgrade at all though?”

Where it does make more sense is in non-military hardware. Mining, construction, transportation, logistics, space stations, planetary development, etc etc etc. Infrastructural development, basically.

To go back to my prior spreadsheet re: construction drone scaling, for example. I added a “trade off” column of “mass” - amount of materials used by one construction drone.

I made a couple of assumptions with numbers, for the sake of getting something out of the model:
  • I assumed a Level 1 drone would weigh 0.5 tonnes and be able to process 0.3 tonnes of refined materials into a constructed object per minute.
  • I assumed that the construction rate base increase would be 25% per level, and that the mass increase would be 20% per level.
  • I assumed an acceleration exponent of 1.3 to counterbalance diminishing returns.
  • I assumed that cost would scale linearly with mass, so the one could be used as an analogue for the other. This would probably not be true for a whole host of reasons, but I didn’t want to overcomplicate the model.
  • I assumed 100% efficiency, again for simplicity.
At those numbers, one Level 1 construction drone, containing 0.5T of material, can build 18 tonnes worth of ship, space station, asteroid base, mining platform or whatever per hour. Fine for a fighter or bomber sized ship. Increasingly useless at capital ship, space station or planetary megastructure levels. To get a ship containing 60,000T of mass built in a day, you’d need to throw 139 Level 1 drones at it, and these drones would contain 69.5T of mass themselves. Extrapolating rather absurdly for the sake of illustration, you could get the same end result with one Level 130 drone containing mass of 56T.

Obviously that kind of extreme min-maxing wouldn’t be a worthwhile cost saving at all, especially for a single ship. The optimal range is somewhere around level 10 (26 drones with combined mass ~59T to make a 60KT ship in ~1 day). Over an entire manufacturing industry which routinely made ships and stations weighing millions of tonnes, though, it demonstrates that progressing to bigger/faster drones does make sense, and not in an obviously game-breaking way. In fact, as structures get larger, if things like construction time, refining time, production time etc only scale upwards if you add more drones/units, you could end up with hundreds of thousands or even millions of drones. It makes far more sense to buy smaller numbers of bigger/more expensive drones if you want to build a space station than to buy vast numbers of little ones (and also saves the engine from having to model the activity of millions of drones per system).

Considering vertical progression in terms of Production Capability also maps onto real world analogues in, for example, the production of automobiles. Considered in terms of embodied technological progress, a Ford Fiesta might as well be a space shuttle compared to a Ford Model T. But the top speed of a Model T was around 45mph, and the top speed of a Fiesta is around 115mph. That’s a jump of only 1 order of magnitude. Further, most cars are generally driven at somewhere between 20-100mph, and we reached the ability to make a 100mph production car in the 1920s. Top speed hasn’t been where the vast technological increase has manifested.

So here we have a potential square to the circle.

If we consider values like “DPS/watt”, “defense per m2 of hull,” “thrust/watt” etc to be sacrosanct and capped to within an order of magnitude, we can visualise “vertical progression” as being something that large-scale, megacorporations need to do to remain competitive with each other, but that small traders, pirates etc needn’t particularly bother with. While the latest model of small fighter may be considered “more advanced,” if you still like your 1967 Mustang there’s no pressing need to upgrade it because it’s still (assuming either no degradation over time or maintenance enabling you to keep component efficiency at 100%) broadly equivalent in performance to those super shiny modern ships.

We can also start thinking about what aspects of materials, weapons, ships, thrusters, computer cores etc can be given an “infinite” component that will provide a tangible benefit while not taking those items outside of our set limits for sensible gameplay. At the moment I got nuthin’, but it’s a valid area of inquiry.

Now, to make an objection to my own work: by itself, this only partially addresses market saturation and horizontal growth concerns. The big vector of competition in the end-user market with more efficient and faster productive capacity is price, which helps somewhat, but not as much as a “scaled treadmill” does because price obviously has a pretty hard set of lower bounds. We need a way to give Hardenberg The Pirate an incentive to buy an upgrade from his 1967 Mustang without making it necessary for him to do so, and that’s a tricky one.

One way of partially addressing this would be to borrow from something I floated over in the Macro Economy thread regarding basic commodities and introducing a form of irrationality to NPC decision making in the form of Novelty Preference.

Consider this to be a single variable between 0 and 1, normally distributed across the population. At 0 an NPC will only spend money to upgrade if the cost-benefit of said upgrade is clear and apparent - Hardenberg with his 1967 Mustang. At 1 an NPC always wants the newest and shiniest thing even if it doesn’t really give them any tangible benefit. Most NPCs, in the range 0.3-0.7, will pay some sort of premium based on the newness of a given item, even if a completely rational cost-benefit analysis might make them skip the transaction.

This partially simulates depreciation as the Novelty Premium diminishes over time. It also provides a built in advantage to smaller players - even if I can only make 100 McDuff-Brand Lasers in a week, and they’re not really much better than the ones MegaCorp Inc can churn out at a rate of 1,000 a day, NPCs with a high novelty preference will prefer mine because they’re new and rare, meaning I can either charge a higher price or quickly capture as much of the market as my productive capacity allows if I keep my price competitive.

These are not complete solutions by any means. But it feels, to my addled brain, like a decent first stab at addressing the potential bottlenecks of a Zero Technological Growth economy while also not forcing smaller players onto a treadmill of Upgrade or Die.
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#3
Well, why do you buy a new car if it doesn't go any faster?

I don't argue (and I don't believe Josh intends) that research makes "no difference". However it's been made pretty clear that there's a definite end point for things like DPS etc, and it's probably within an order of magnitude of the start point. So, I'm working off those assumptions now, and therefore trying to work out a way of maintaining aggregate demand even in a system with capped progression.
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#4
There is yet another aspect not covered here, and that is different functions for certain item types.

Let's assume that there is a rather large corp specializing in weapons. Let's call it ThymineC Munitions, to keep in with the tradition of namecalling. They produce guns in various sizes, and in quite a volume. The guns are technologically well developed, but fall strictly into the high damage, slow ROF, bad tracking category - to paraphrase, they build the equivalent of sniper rifles. Which is fine and dandy, and all their development went into refining this, so they're pretty good sniper rifles.

Now, assume that you need something to deal with fighters. The above setup, no matter how good, has a horrible time dealing with them. Ideally, you want some low-damage, high ROF weapon system with excellent tracking, to keep the little buggers at bay. ThymineC Munitions doesn't have those, and despite being big, their research investment goes into a completely different direction.

Another local outfit, Katorone Kannons, specializes in high ROF stuff. Quite obviously, they will sell based on the functionality of their guns alone, despite being smaller, having less production capacity and probably, by comparison of stat bonuses, worse weapons. But they fill a niche.

DWMagus Missileworks produces swarmers and small guided missiles. Tombs Torpedoes doesn't do "guided", but will sell you stuff with warheads big enough to demolish a station. Both obviously produce missiles, but cater to different market segments.

It would be an enormous investment of research time and resources for a single manufacturer/faction to cover ALL of the equipment permutations. As opposed to the market for smartphones, which essentially all do the same thing, there is quite a lot of permutations in the spaceship equipment business here.

Do you want large, tough fighters, or would you settle for cheaper and more flimsy ones, if you could get double the amount of them for same hangar bay mass investment? Do you want your construction drones fast or efficient? (less waste during construction, or faster, but with more material used?) Do you want acceleration or top speed, shield capacity or shield recharge rate? Armor hitpoints or armor damage resistance? More cargo capacity or more module slots in your hull?

There's oodles of "different". Adding vertical progression to that further diversifies the market, but keep in mind that the market is local - TanC Hyperdrives might build lighter and faster engines than Flatfinger Dynamics, but that won't help me if the next outlet of them is 35 jumps away and I need to fit a new transport now.

"Better" might not mean statistically better, but rather "more suitable for the task at hand". It can't always be reduced to a raw "item level" type number.
Well, why do you buy a new car if it doesn't go any faster?
'Cause I've got a green-skinned wife and a pouchful of hatchlings now, and my '67 Mustang isn't big enough for all of us to ride in. ;)
Hardenberg was my name
And Terra was my nation
Deep space is my dwelling place
The stars my destination
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#5
Well maybe we wouldn't have to worry about a finite progression system. Maybe eventually if a game progresses long enough, and it does reach a point of basically no research progression, then the universe can go into a collapsed point of all out war. Destroying past technologies and reversing their point in there research branch. This could allow surviving corporations to start on a different branch of research repeating the process all over again. Of course this would be very late in game but it could be a possible mechanic to prolong the game. I'm sure there is an even better solution to this. :problem: :monkey:
IF YOU AINT OUT OF CONTROL, YOU AINT IN CONTROL!
EVE ONLINE FO LYFE
STAR WARS FOR LYFE
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#6
McDuff wrote:I don't argue (and I don't believe Josh intends) that research makes "no difference". However it's been made pretty clear that there's a definite end point for things like DPS etc, and it's probably within an order of magnitude of the start point. So, I'm working off those assumptions now, and therefore trying to work out a way of maintaining aggregate demand even in a system with capped progression.
Perhaps it's just me. But for some reason I really don't understand where your insistence on a limit or end point comes from. I don't think Josh says anything like there being an end point in that post you link you.

My understanding of what Josh is looking for is a universe in which at any given time there is a spread in the "effective power" of technology from low to high, but that over time this power increase at both ends at a slow rate. The point is that things do improve, as they do in real life and over the course of a few years of in-game play the entire market will have turned over so that the toys you have available top play with are totally different.

This whole concept is summed up by this plot:
market_range_player_spec.PNG
market_range_player_spec.PNG (13.38 KiB) Viewed 4699 times
There's no eventual upper limit and things do become obsolete over time. But that period should be long enough to hold onto your favourite gun for long enough before you discard it.

Are we disagreeing over this at a fundamental level, or is you point how difficult it will be to create this kind of universe?
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#8
You know this adds to a good point.

If you do hyper-specialization, maybe only then can you start unlocking some vertical progression. It would make sense that after years (in game) ThymineC Munitions could improve upon the sniper rifle design to have even harder hitting guns without as much of a penalty.

In this sense, it means that vertical progression via research becomes more available to those who would focus on such a thing, but those that wouldn't probably wouldn't need to worry about it anyways.
Image
Early Spring - 1055: Well, I made it to Boatmurdered, and my initial impressions can be set forth in three words: What. The. F*ck.
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#9
mcsven wrote: Perhaps it's just me. But for some reason I really don't understand where your insistence on a limit or end point comes from. I don't think Josh says anything like there being an end point in that post you link you.

My understanding of what Josh is looking for is a universe in which at any given time there is a spread in the "effective power" of technology from low to high, but that over time this power increase at both ends at a slow rate. The point is that things do improve, as they do in real life and over the course of a few years of in-game play the entire market will have turned over so that the toys you have available top play with are totally different.
Diminishing returns equates to, in real terms, an effective end limit.

Suppose the rate of progression at T=1000 has slowed to the point that 1Bn investment gives me a marginal increase of 1%.

Even though the progress hasn't stopped, it's highly unlikely that any rational actor will consider this to be worth the investment. The cost benefit equation is all out of whack, and I can get a better ROI from purchasing more items at the current technological level than investing in research to try to get myself to that higher level.

Without a significant marginal reward for my investment in research, I have no reason to do it.

Unless every other branch of the research tree is full, my temporal accrual of research points is better invested where it will yield more reward.

If we're talking about making vertical progression slower or faster, we're left with the same problem Hardenberg pointed out. If it's quick enough to be economically efficient for a viable economics sim, it's too fast for people who don't want to be outpaced on an "upgrade or die" treadmill, and vice versa.

"Just slow it down or turn it off" was previously rejected as a solution to this.
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#10
Hardenberg wrote: There's oodles of "different"....
"Better" might not mean statistically better, but rather "more suitable for the task at hand". It can't always be reduced to a raw "item level" type number.
However, you still hit limits if you're always working in terms of trade offs.

If I'm trading off power for range, there's a certain point at which I'm making stuff that only works over distances of metres, even if it's immensely powerful. Useful for repelling boarders but only if they try to get in *really close* to my guns.

If I'm trading off range for power, after a while I'll end up being able to hit you from a light year away with something as effective as a stern telling off.

As T increases, you'll end up reaching an effective limit to any given set of trade offs, and diminishing the value of new research.

Remember we're starting the game at an approximation of T=100 years. We can fudge around the matter, kinda, but for the history generation to work the research mechanic can't be broken after this time.

ETA: also, while there may be oodles of different, there's an infinity of space. Oodles seems very very small compared to infinity.

Across a hundred systems you might have 10,000 NPCs. As branchy and cloudy as the research system is, there'll still be an awful lot of overlap in a population that big.

Big numbers fuck a lot of things up. I'm trying to work out how it can be made to work when the numbers get really big as well as when they're really small.
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#11
"Just slow it down or turn it off" was previously rejected as a solution to this.
Then why not turn the picture around and look at it from another angle? Why not just have Josh include an option to toggle realistic economic simulation on?
I know not what life is, nor death.
Year in year out-all but a dream.
Both Heaven and Hell are left behind;
I stand in the moonlit dawn,
Free from clouds of attachment.
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#12
Because as far as I can understand, without the economic simulation there's no game. Even if you never touch the economics yourself, it's what's needed to give you jobs, contracts, people to fight, people to ally with. The difference between a universe that feels like a living, breathing space, and a universe full of procedurally generated NPCs who need to go fight procedurally generated enemies with no history and no context.

What would you replace it with?
Post

Re: Squaring the "Vertical Progression" Circle (maybe)

#15
McDuff wrote:Because as far as I can understand, without the economic simulation there's no game. Even if you never touch the economics yourself, it's what's needed to give you jobs, contracts, people to fight, people to ally with. The difference between a universe that feels like a living, breathing space, and a universe full of procedurally generated NPCs who need to go fight procedurally generated enemies with no history and no context.

What would you replace it with?
My preference, like yours, is that a fully competitive and dynamic economic system as the modus vivendi of the game universe be at least a possibility. A realistically competitive economic (and development) landscape does raise valid gameplay concerns. As such, the default setting should, and likely will, aim for something a bit more sedate and less cutthroat.

However, with a toggle here and a slider there it may be possible to render a Limit Theory universe a bit more Gekko-like.
I know not what life is, nor death.
Year in year out-all but a dream.
Both Heaven and Hell are left behind;
I stand in the moonlit dawn,
Free from clouds of attachment.

Online Now

Users browsing this forum: No registered users and 15 guests

cron