AI Economy

[Dubious]

Inspired in part by the discussion of the Planetary Events suggestion in the "Feature requests" thread, I thought it time to revisit the subject of the game's economic engine which was initially brought up during the Demo release in the "Welcome" thread.

There were some definite improvements to the commodities market between the demo and the initial release. Kudos to Ibol for being responsive to early concerns. But he stated the focus of the game design was to encourage advancement into new sectors rather than on trading, so it's just an algorithm designed to adjust prices over time. Observations have shown that while prices do fluctuate, there are not significant changes BETWEEN commodities. The most profitable commodities to buy and sell in a given sector remain stable (with only minor exceptions) throughout the period of observation. Only the commodities with the lowest profit margin are likely to fall out of profitability between one price adjustment and another.

EARLY SECTOR OBSERVATIONS:

In tracking trades across the first six sectors for 20,000+ turns, I have developed a spreadsheet to aid in quickly discovering the best deals within WP drive range (currently 2 sectors). Consequently I have learned I can completely ignore the Best Buy/Sell Sector entries (except when data is stale but not obsolete: about 2100 turns) as they don't mean that commodity is the most PROFITABLE (rarely the case); as well as the Economic Events for the same reason. Shortages and Surpluses are rarely worth noting either if all you want is the single most profitable commodity, though when they coincide with the most profitiable trade you can make a relative killing quickly. By having the spreadsheet simply highlight the best deals (buying and selling) from the current prices (max value) at that station, compared with the adjacent (higher numbered) sector out to the WP single jump limit, I learned the best deals are all you need to know. This what I would expect any basic trader program to produce, and displaying that information the single most important improvement I would suggest to the game. Once you know the best deals for the sectors, they remain so (with minor exceptions for surplus/shortage). Knowing this still doesn't prevent you from making mistakes, such as buying the commodity you are supposed to be selling in that market, but it makes the process faster.

In my (admittedly limited) experience, the most profitable commodity trades tend to remain fairly stable despite price fluctuations. If the highest value is Antimatter bought in sector 3 and sold in Sector 1, and Transuranium bought in Sector 1 and sold in Sector 3, then those trades seem to remain the best among those sectors, regardless of ALL other factors. At least in the first 5-6 sectors, you can ignore all the commodities listed above "Chemicals". They are simply never profitable enough compared to others. "Chemicals" is the most volital of those remaining, falling in and out of "most profitable" on occasion. The rest seem to remain profitable enough to gamble on when necessary, even if your data is quite "stale" and out of date.

Once you have determined the "most profitable" deals in a sector, you can pretty much ignored all subsequent price changes if you don't care about wringing out the very last dollar from every deal by updating the market prices whenever they change. Just watch out for surplus/shortage situations in your "deal" commodities. Those are generally worth taking the time to exploit.

Price changes seem to occur on the same date for a race. That is, all the Terran sectors change their price on roughly the same day (based upon noting the change and jumping to the next sector immediately to find it has changed as well), while the Gruff, Tentaculon, and Firax follow their own dates. (Only observed one sector of each race, so this is conjecture, but their date change was independent of the Terran's, though still seeming to generally be a 2000-2100 turn interval.)

When the "Price/50 supplies" changes, it pre-sages a Major Market Price Shift (probably next trip). This is not infallible, but if near the end of a cycle it's probably a good idea to give up on exploiting any particular market advantage you don't want to get caught holding a large quantity of when the bottom drops out.

-Dubious-

[ibol]

I find your analysis intriguing!
Also, it makes me want to further refine the economics.
You make me want to write a separate program to analyze the economics further..

It is true that I designed the commodities to have a progression from least to most expensive. I can't really envision a time when boxes would be more expensive than anti-matter. but, I could understand some greater fluctuation between commodities closer in relative price...

The first update cycle is nearly complete, so we will have to wait and see what happens in the future.

[Dubious]

Originally Posted by ibol

I find your analysis intriguing!
Also, it makes me want to further refine the economics.
You make me want to write a separate program to analyze the economics further..

It is true that I designed the commodities to have a progression from least to most expensive. I can't really envision a time when boxes would be more expensive than anti-matter. but, I could understand some greater fluctuation between commodities closer in relative price...

The first update cycle is nearly complete, so we will have to wait and see what happens in the future.

Good to hear my timing was on target. I didn't want this to get into the first "bug fix" release, but hoped maybe the next "added features" version.

Re: "I can't really envision a time when boxes would be more expensive than anti-matter." The issue is not "more expensive". The issue is "more profitable". It's not what the buy/sell price is at, it's what the differential is at. People lose track of the fact that a price differential of 10% on commodities is the NOT same in terms of profitability at a price tens of dollars as at a price of thousands of dollars. A "buy/sell" or "profitability" differential of 10% on $10 is only $1, while 10% of %1000 is $100. So if you are applying the same percentage across the board, things tend to stay the same. But "supply and demand" is not an "across the board" static adjustment. A 1% differential in $1000, is +$10. A 1000% differential in $10 is +$100. THAT is a supply/demand difference. It takes less of a difference on high value commodities to have a large impact on profitability.

So, to my mind, what is needed is a way to apply some "supply/demand" with a logic behind it. Now for a bit of theory:

Space stations are expensive and require trade if they aren't just going to be the equivalent of a yacht: "a hole in the water[/space] into which you pour money". All the consumables have to be imported. That means air and water (recycling is never 100% efficient and there is always some "leakage"; apparently this is being ignored or "hand waved away" as a given non-issue), food, and raw materials for any manufacturing. And luxury items and entertainment for inhabitants. To pay for this, there has to be something for the station to export in order to maintain some sort of "balance of trade". And where are those resources (of both import and export types) going to come from? Planets and asteroids. With the possibility (assuming the correct technology exists) of some "mining" of stellar bodies for exotic elements and gases. There has to be enough present in a sector to make building a station feasible. (The game may assume all of that, but it isn't obvious in the market forces at play.)

That's the premise. A station has some resources for export, (which should be based upon the resource population of the sector) and everything else has to be imported. Commodities should be associated with particular celestial bodies: planets, asteroids, stars, etc. Everything available in the sector is exported through the station. If there are multiple bodies of the same type in a sector, then that commodity will be plentiful and relatively cheap to export. But an important function will also be the temporary storage of transit trade goods bound to other destinations. (I assume this function is abstracted into the market price fluctuations.)

The resource population of a sector should establish certain commodity baselines for that sector. If, for instance, there is more than one asteroid belt, then there are not going to be shortages of "raw ore" (except possibly for a very short term) and price fluctuations are going to be minimal. If there are none, then there are not going to be surpluses (except possibly for a very short term), and it is more likely to have shortages and wide fluctuations in price.

Certain resources should be necessary for the manufacture of certain finished goods, like Stim-Ware or Microchips. Again, if they aren't present in the sector, it's unlikely they will be manufactured there. If a sector is surrounded by other sectors with the same resource, it may be more profitable to specialize in processing the raw material into an intermediate commodity than for each resource rich sector to process their own.

All of these are factors that should be obvious without even needing to see the commodities board. The player may need to work out the commodity to resource association, but it should be possible. (If it is there now, it's not discernible.)

Ideally each race would require food associated with a particular type of planet. Multiple types of planets make the sector attractive to those races they can support. If there aren't any, they wouldn't establish a permanent presence except possibly for scientific purposes, in which case they would likely be pretty sparse and restrictive as to available resources for export as their function isn't really "trade". (But they would likely be good consumers of luxury and entertainment goods, depending upon how frequently their sector is visited.)

Which brings up another point: frequency of transit. Ports do not have equal numbers of visits. Like in real estate, it's all about location, location, location! Some are better located than others and get more traffic. High traffic area are more likely to have minor price fluctuations but rarely major ones. Poorly trafficked sectors are much more likely to have wild swings between rare visits as exports pile up and imports get scarcer. Distance also plays a role in both expense of shipping and frequency of transits. "Home worlds" are going to have more traffic than remote colonies. The use of "worm holes" to create short-cuts will make certain sectors more advantagous than others to increasing trade.

The size of a commodity in terms of the amount of cargo space it takes up is also a factor. Small volume but high value items are every trader's dream cargo. But most have to settle for mixed loads. The intrinsic value of a large commodity like "heavy equipment" might be low at the point of manufacture, but high at the destination. But the shear size of the commodity makes fewer able to be transported at a time. That also has a tendency to push the selling price up at the destination that NEEDS that equipment while traders want more profit on the smaller number of units they can fit on board. In fact, some ships (like the "assault scout" or "stealth runner") might not even have enough room to load that commodity in combination with anything else. In which case, they may have to settle for a less profitable load. And why should all ports have an infinite quantity of all commodities available at any time? Only bulk carrier ships today are likely to have only one commodity in their holds.

The nature of those transiting also come into play. Sigurns are traders, so they will tend to bring down prices. Tentaculons are scientists, so they tend to not be trading except in "data". But all races should have some sort of trade they are engaged in, as they all have a "balance of trade" issue.

Just some thoughts.

-Dubious-

[plugger]

Hi,

I like the trading aspect to.

However, when I run the numbers, it's hard to make it pay 'cause of the cost of supplies used to warp and to travel within sectors.

You can buff your ship and give it a trading slant and you can take advantage of shortages but it'd be nice to be able to more of a trading 'vibe' on offer.

Perhaps you could be offered a contract with a MegaCorp ('Evil-is-Us') at a bankers station occasionally. Taking the contract gives you a % discount for a particular commodity 'cause that's what Evil Corp deals in.

You can retain your contract for as long as you keep trucking a certain amount of goods back to Evil Corp HQ in sector 'x'. You'd be able to get cheap supplies in return making the trading runs more viable.

Maybe occasionally they'll offer you a 'special', under the table, contract. Flying around with the Evil Corp logo on the side of your ship might cause a few problems with certain embassies.

Should have taken the contract with 'Interstellar Donuts' instead...

Cheers,
Plugger

[Waltorious]

Originally Posted by plugger

Hi,
Perhaps you could be offered a contract with a MegaCorp ('Evil-is-Us') at a bankers station occasionally. Taking the contract gives you a % discount for a particular commodity 'cause that's what Evil Corp deals in.

Another way to do something like this is to simply have a "delivery" quest, kind of like the banker delivery quests, except the quest requires buying and transporting a certain number and type of goods to a specific sector. That way there'd be a cash reward on top of the money from the trade. This wouldn't be as free-form as true trading but it would give ships with a lot of cargo space another way to make good money.

Some races have similar quests already, but I'm imagining something more specific, like "transport 12 units of Chemicals from sector 4 to sector 6". You could give some justification, e.g. there's a shortage which resulted in a rush order or something and they need a courier.

I think it would also be pretty cool to have the opportunity to transport contraband, and engage in smuggling operations. Maybe the Pirates could offer quests related to this? (I've not tried any Pirate quests yet, so I'm not sure what they're like currently.) This would be harder to implement as you'd probably need to add some devices and effects that help one get illegal goods into stations without being detected.

Actually, contraband might not fit with the lore that well, since the Bankers would be happy to trade any kind of good as long as it's profitable. So maybe the Bankers could offer missions to smuggle goods to other races? "The Limoquee have foolish restrictions on trading stim-ware. Find a way around their patrols and deliver this stim-ware shipment to a Limoquee spacehome." Something like that? The pirates might offer similar, especially since Limoquee shoot pirates on sight, so the player could be their "undercover" delivery ship.

[Dubious]

Originally Posted by Waltorious

Another way to do something like this is to simply have a "delivery" quest, kind of like the banker delivery quests, except the quest requires buying and transporting a certain number and type of goods to a specific sector.

+1 to this idea.

Originally Posted by Waltorious

I think it would also be pretty cool to have the opportunity to transport contraband, and engage in smuggling operations.

You are aware that a number of commodities are already "tagged" as "illegal"? They are:

Stim-Ware
Weapons
Transuranium
Narcotics
AntiMatter

All of which tend to be the most profitable, especially in the early game. I'm fully expecting to have some "police" race show up in the later game.

Of course, currently there is no mechanism (to my limited knowledge) to attempt to "hide" such contraband from the authorities. Though a device for that purpose should be simple enough to add to the game. Perhaps as the result of a pirate quest?

-Dubious-

[Dubious]

More thoughts on trading:

One of the ways to make the life of a trader more "interesting", is to make it so the player has to put together combinations of consignments of cargoes into a profitable trip. As it is now, each commodity takes up a fixed amount of "cargo space" for the market price, and that cargo can be sold anywhere the trader cares to dispose of it.

This is a way of abstracting the current "intermodal shipping container" [http://en.wikipedia.org/wiki/Shipping_container], which are reusable. These come in nominal standard 20 and 40 foot long metal "boxes" (or "cans" in Sci-Fi space trading lingo): one 40 foot container is the standard unit moved between truck, rail, or ship. Two 20 foot containers are placed on a rail or truck "spine car" which keeps them together until they reach their respective destinations. These "half-cans" are generally the smallest container handled. Smaller packages would be packed into either a full or a half-can. A nominal 20 foot "half-can" has an internal volume of 1169 cubic feet (imperial) or 33.1 cubic meters (metric) [http://en.wikipedia.org/wiki/Intermodal_container]. There are variation in these sizes and volumes, but this is merely to give an idea of general capacity. Remember, those are the volumes of a 20 foot "half-can". Roughly double that for the normal 40 foot "full-can".

The big advantage of using "cans" is that it standardizes the cargo size of the "commodities". But not all "commodities" are the same size and this can get forgotten. If a piece of "heavy equipment" takes up one entire can, then the market price is for delivery of one "unit" (which just happens to be "one can") of cargo. On the other hand, one unit of "boxes of boxes" is a "can" of boxes. A full can would be roughly 66 CUBIC meters of boxes. Assuming the market price is an abstracted "unit value" (i.e. $1 = $1,000), that is still a lot of boxes in even one "full-can" of cargo. Sectors will still need them (for packing commodities into "cans"), but not likely a repeating trade loop of an entire cargo ship's capacity worth. But "half-cans" leave "gaps" in a ship's cargo hold to fit odd lots into that may otherwise not appeal. That sort of "dribble" is what keeps those lesser profitable commodities from becoming urgent shortages and surpluses in the supply/demand market.

Also, many "cans" contain "mixed lots" of commodities destined for the same end-point. (Think of "household goods" for a family moving to another sector, or a consignment for a department store or laboratory.)

The consequence of thinking in terms of capacity is that a trader then has to juggle current available capacity with cargo profitability. It's extremely easy when all the cargo is going to one destination: pack the hold with the single most profitable item and sell everything at the destination. This is the economic model in use now, and the only really practical one when starting out. But most cargo is not going to just one port. The port itself is merely a transhipment point. Trader ships are "space trains" hauling loads of "cans" from one point to another. If the destination is 8 sectors away and the trader can only jump 2 sectors at a time, then the "running costs" of docking fees and re-supply at the 3 intermediate stops eat into the available cash until the payday at the end of the run. It normally will make sense to leave at least some room for cargo bound for the intermediate stops, especially if tight on cash to begin with. But until contracted, profitable "long hauls" become available, this will not be a popular course of action.

The game already uses the concept of "cargo space" (e.g. "full-cans"), and variation in how much certain items take up. The current problem is that it is unknown how much cargo space is required for each item until after it is "bought". The careful trader would need to buy one of a commodity or item, and then track how much cargo space it took up. It's not clear if "half-cans" are already in the picture, as all that is given is the amount of remaining available cargo space (which may always be rounded down, giving a false picture of actual requirements). The "shipping specialist" skill suggests this capability is being abstracted. But while an increase of available cargo space increases profitability, it doesn't increase interest in the trading game. Making the default assumption that one unit of cargo space is a "half-can", thus allowing half cargo space units, increases the possible combinations without changing the remainder of the game. But leaving the assumption "as is" and changing the size of various commodities can have the same effect.

To increase interest in trade, there is a need to have two kinds of cargo to be available: for sale, and for delivery.

The current market place is a "for sale" market, which encourages the trader to buy commodities on speculation of being able to make a profit at a specific destination. By having varying quantities available at each port, the profitability question becomes less "cut-and-dried" than "buy a full load of this here and sell it there".

"For delivery" could be implemented (as suggested by Waltorious) in the form of "quests". Though ideally this would be in the form of multiple quest offers of specific quantities going to different sectors, with specific "delivery by" dates. Then the trader needs to plan out the route and juggle capacities at various stages along it. The trader would not be "buying/selling" the commodities, but rather earning a "Transportation Fee" of a fixed amount: a "sure thing" rather than a speculation, that hopefully leaves a profit after running expenses. Not generally considered the way to "get rich", but usually a sustainable, slow growth method. That will encourage leaving some free space for "spec loads". Some smaller ships may not have the available capacity to take on many delivery quests, and remain forced in the speculation mode. OTOH, such small ships should be more flexible in taking multiple high-profit, small volume loads to different sectors. More like "large cargo space ships = railroads, small cargo space ships = trucks".

-Dubious-

[Kazeto]

Hmm ... I'll say I like the idea. It would require reworking how the market works but generally repeatable trading quests—in spite of delivery quests generally being monotonous—would probably be less bothersome to people because it would keep them doing something different.

That being said, if such an option were to be implemented I propose adding a module that would increase cargo capacity even more than a normal module but at the cost of only being usable for those missions; you needn't even be creative with the name because calling it “Shipping Container” or “Delivery Module” would probably work.

[ibol]

OK, I have now read all of this... It is a lot to take in, but very interesting.

Commodities should be associated with particular celestial bodies: planets, asteroids, stars, etc. Everything available in the sector is exported through the station

(and the other ideas associated with this line of thought)

This sounds more like the outline for an entirely new game, rather than an addition to AI. It is interesting, but in my opinion, not within the scope of (this version of) the game. Perhaps AI2?

You are right that a lot is abstracted. What you see in the sector maps is just the stuff that you can interact with. You can assume there are colony worlds, other ships, non-interesting space features, etc. scattered everywhere, but they are of no gameplay consequence. The original design of AI was very concerned with ease of interaction and instant rewards. Things that don't do anything immediate generally aren't shown.

Here is something your economic analysis has not mentioned, but I feel safe in revealing: The game keeps track of what you buy and sell, and adjusts prices accordingly. Buy a lot, prices go up, sell a lot, prices go down. Perhaps the effect is too small to notice. Perhaps you won't notice it until you find a shipwreck full of transuranium, plunder it, sell it off, and then the price drops.

All that said, I do intend to refine the trading aspect. Some good ideas have been put forth here. I also plan on a special type of quest that will affect the actual prices of commodities once completed

[Dubious]

Can't say I am surprised to hear this is too radical a change for a mere "feature addition". Something for AI2 consideration is entirely reasonable.

Abstraction is not a bad thing. Most people are not into heavy amounts of detail in a "rogue-like" game. The best I hope for is to discern some logic in the abstraction. That the "most profitable" deal for a sector remains pretty constant would in fact be a way of abstracting "commodity association", which led me to hope perhaps there was some of that going on. But the focus on the high end of the market made me wonder if I was reading too much into it.

I had noticed that buying and selling affected the price of those particular commodities only (they would be the only items that changed price), and found it very reasonable. Of course that I only had 5 available cargo space for most of my trades meant I wasn't seeing the impact of large volumes at the time, but have seen more now that I have upgraded to a cargo vessel. The only suggestion on that would be a timer on how long between those particular types of trade (i.e. how long between one "sell" and another "sell", and another for the "buys") affecting how big an impact. That would be a way of abstracting the amount of "trade" in that commodity, though the player may be having too much of an impact if only their actions affect the timer. And I have seen some evidence that perhaps such a timer already exists.

One more late thought I forgot to add: "perishables". Most of the commodities are either manufactured goods or raw materials, which tend to be less time sensitive in getting to market. Only a few, like "food", and "narcotics" have a "shelf life". (Stim-ware falls in-between.) The longer such perishable items remain on a station, the more their value drops, so they become more urgent to move to their market before it becomes impossible to make any profit on them. Makes for another "trade quest" rationale.

Oh, and I ought to mention that many of my ideas on commodities come from playing "railroad" games like "Railroad Tycoon" as well as reading.

Looking forward to seeing what you can come up with.

-Dubious-