Commodity Market

Suggestions for new DLC projects.
User avatar
greene345
Level 3 user
Posts: 82
Joined: Thu Nov 22, 2012 5:43 pm

Re: Commodity Market

Post by greene345 »

Regarding Commodity market, would that be something that players could get in on too? For example, I would love to buy several armored trucks worth of gold and keep it in my basement :p

If we did that though, there should also be a warehousing fee, and perhaps that could be something that should also be added to the game (storage). For example, one of the biggest businesses in the US is the storage industry. In short, paying other people for space to house all our stuff. If we could build storage warehouses, for example, we could have something along the lines of Advertising Budget, Security Budget, Technology Upgrades, and Space.

So, for example, if I wanted to build a consumer warehouse, there could be options to upgrade all of the above (via the sliders). The technology upgrades would cost money and take time (during which time the facility would bring in no income). But, once upgraded there would be a higher income moving forward.
counting
Level 8 user
Posts: 843
Joined: Wed Jan 29, 2014 12:44 am
Been thanked: 1 time

Re: Commodity Market

Post by counting »

greene345 wrote:Regarding Commodity market, would that be something that players could get in on too? For example, I would love to buy several armored trucks worth of gold and keep it in my basement :p

If we did that though, there should also be a warehousing fee, and perhaps that could be something that should also be added to the game (storage). For example, one of the biggest businesses in the US is the storage industry. In short, paying other people for space to house all our stuff. If we could build storage warehouses, for example, we could have something along the lines of Advertising Budget, Security Budget, Technology Upgrades, and Space.

So, for example, if I wanted to build a consumer warehouse, there could be options to upgrade all of the above (via the sliders). The technology upgrades would cost money and take time (during which time the facility would bring in no income). But, once upgraded there would be a higher income moving forward.
We already have a warehouse "fee", it's what warehouse monthly overhead cost and employee salary stand for in the game.

Although I like the idea that warehouse storage space and input/output efficiency can be upgraded via research.
-------------------------------------------------------------------
Twitch channel : twitch.tv/ancientbuilder
Youtube channel : www.youtube.com/user/countingtls
-------------------------------------------------------------------
therealevan
Level 4 user
Posts: 178
Joined: Sun Jan 27, 2013 11:37 pm

Re: Commodity Market

Post by therealevan »

Warehouses should be reworked, stockpiling inventory is absurdly expensive.
Image
counting
Level 8 user
Posts: 843
Joined: Wed Jan 29, 2014 12:44 am
Been thanked: 1 time

Re: Commodity Market

Post by counting »

As long as you are able to sell off the inventory with good enough profit margin, it's just a delay of payment, like in real business you need to have good inventory turnover. And the warehouse maintenance expenses are small enough to be ignored.
-------------------------------------------------------------------
Twitch channel : twitch.tv/ancientbuilder
Youtube channel : www.youtube.com/user/countingtls
-------------------------------------------------------------------
infoscott
Level 3 user
Posts: 64
Joined: Fri Mar 06, 2015 11:04 pm

Re: Commodity Market

Post by infoscott »

In this proposal, I see a blurring of the lines between a forward contract capability and a futures contract (derivative) market. If you just want to implement a supply hedge primarily, with a secondary effect of price guarantees, you could do that just with a forward contract mechanism. If you also want to speculate on commodity spot prices, then yes, a futures market would be nice to have. But keep in mind that most of the futures contracts traded in the real world are speculative and are settled without any product delivery.

My suggestion on how to implement:

A storage unit in a warehouse or inventory unit in a natural resources firm should have a button available to initiate a forward contract. Probably at first it would be easier to implement filled units and not partially filled units. This will allow that unit's goods to be offered "for sale" at either a specified or unspecified future start date at the price of the supplier's choosing. A buyer (AI or player) can choose to enter the contract on a first come, first served basis (like selling a firm). Once the contract is finalized, that unit is locked until the buyer starts taking delivery, the delivery of which acts like an internal sale but only to that external buyer. If the product is eliminated through unit reassignment or firm destruction, the seller has to pay a $ penalty for breaking the supply contract.

A futures market could work independently of and parallel to the forward contract market if the futures instruments are settled in cash and not product delivery.
counting
Level 8 user
Posts: 843
Joined: Wed Jan 29, 2014 12:44 am
Been thanked: 1 time

Re: Commodity Market

Post by counting »

infoscott wrote:In this proposal, I see a blurring of the lines between a forward contract capability and a futures contract (derivative) market. If you just want to implement a supply hedge primarily, with a secondary effect of price guarantees, you could do that just with a forward contract mechanism. If you also want to speculate on commodity spot prices, then yes, a futures market would be nice to have. But keep in mind that most of the futures contracts traded in the real world are speculative and are settled without any product delivery.

My suggestion on how to implement:

A storage unit in a warehouse or inventory unit in a natural resources firm should have a button available to initiate a forward contract. Probably at first it would be easier to implement filled units and not partially filled units. This will allow that unit's goods to be offered "for sale" at either a specified or unspecified future start date at the price of the supplier's choosing. A buyer (AI or player) can choose to enter the contract on a first come, first served basis (like selling a firm). Once the contract is finalized, that unit is locked until the buyer starts taking delivery, the delivery of which acts like an internal sale but only to that external buyer. If the product is eliminated through unit reassignment or firm destruction, the seller has to pay a $ penalty for breaking the supply contract.

A futures market could work independently of and parallel to the forward contract market if the futures instruments are settled in cash and not product delivery.
I agree this could work hand-in-hand with futures-exchange-like mechanism, but a lot of works need to be done for contracts to behave, let alone become financial commodity to be traded.

Non-standard OTC commodity forwards are still contracts in nature, so the performance of this mechanism depends heavily on AI behavior. From what we know about current AI present tender offer, it would be more or less like putting real estate on-sale, where players would rarely do that due to the unfavorable terms AIs proposed. Also a lot of codes need to be added for current AIs since they don't actually have any clue about using warehouses, hence before AI can "understand" how to do it, they won't be able to "sell" it.
-------------------------------------------------------------------
Twitch channel : twitch.tv/ancientbuilder
Youtube channel : www.youtube.com/user/countingtls
-------------------------------------------------------------------
infoscott
Level 3 user
Posts: 64
Joined: Fri Mar 06, 2015 11:04 pm

Re: Commodity Market

Post by infoscott »

counting wrote:
infoscott wrote:In this proposal, I see a blurring of the lines between a forward contract capability and a futures contract (derivative) market. If you just want to implement a supply hedge primarily, with a secondary effect of price guarantees, you could do that just with a forward contract mechanism. If you also want to speculate on commodity spot prices, then yes, a futures market would be nice to have. But keep in mind that most of the futures contracts traded in the real world are speculative and are settled without any product delivery.

My suggestion on how to implement:

A storage unit in a warehouse or inventory unit in a natural resources firm should have a button available to initiate a forward contract. Probably at first it would be easier to implement filled units and not partially filled units. This will allow that unit's goods to be offered "for sale" at either a specified or unspecified future start date at the price of the supplier's choosing. A buyer (AI or player) can choose to enter the contract on a first come, first served basis (like selling a firm). Once the contract is finalized, that unit is locked until the buyer starts taking delivery, the delivery of which acts like an internal sale but only to that external buyer. If the product is eliminated through unit reassignment or firm destruction, the seller has to pay a $ penalty for breaking the supply contract.

A futures market could work independently of and parallel to the forward contract market if the futures instruments are settled in cash and not product delivery.
I agree this could work hand-in-hand with futures-exchange-like mechanism, but a lot of works need to be done for contracts to behave, let alone become financial commodity to be traded.

Non-standard OTC commodity forwards are still contracts in nature, so the performance of this mechanism depends heavily on AI behavior. From what we know about current AI present tender offer, it would be more or less like putting real estate on-sale, where players would rarely do that due to the unfavorable terms AIs proposed. Also a lot of codes need to be added for current AIs since they don't actually have any clue about using warehouses, hence before AI can "understand" how to do it, they won't be able to "sell" it.
I think if we keep the futures separate as cash settled trades, it's not much different than the AI trying to min-max the stock market. The trick may be to get the sim engine tuned to fluctuate commodity prices in a meaningful way.

AI and warehouses...hmm. If they are only looking at sales units, that's a problem. There would need to be some goal seeking code to look upstream from a sales unit to see if there is a locked warehouse that could satisfy future need. Do you know if the AI does any future logistics projections at all when it determines suppliers?
counting
Level 8 user
Posts: 843
Joined: Wed Jan 29, 2014 12:44 am
Been thanked: 1 time

Re: Commodity Market

Post by counting »

infoscott wrote:
AI and warehouses...hmm. If they are only looking at sales units, that's a problem. There would need to be some goal seeking code to look upstream from a sales unit to see if there is a locked warehouse that could satisfy future need. Do you know if the AI does any future logistics projections at all when it determines suppliers?
From what I gathered over the years, AI tends to consider supply and demand great deal when choosing suppliers. After that, if there are multiple sources available, it would choose the highest rating one (in current city first, than considering globally), like what a factory would link to if allowing auto-linking (I assume the programming code is the same). And I don't think there is any regression or statistical analysis behind choosing a supplier, but simply by some weight average index comparing with a threshold value. I don't think current AIs have any capability assessing future risks either. And human players can play the same trick over-and-over to the same AI as many times as they like. They only adapt to immediate changes, but never learn from past experiences.

However, AIs sometimes do exhibit certain game theory co-operative behaviors between each others. Like choosing separate products in the same class if both have the same expertise, so they can both be monopoly in some products to gain higher profit, rather than go toe-to-toe in all products. And although obviously willing to go full on price war reacting to human players, but left for AI's own choosing, sometimes AIs will settle to a relative profitable position regarding to their production capacity. Sometimes less aggressive AI setting, make AI corporation perform better if no human try to squeeze them out. I also think a lot of the rule-based conditions are specifically targeting human player's instead of maximizing profit (hence the suicide tendency a lot of players have observed)
-------------------------------------------------------------------
Twitch channel : twitch.tv/ancientbuilder
Youtube channel : www.youtube.com/user/countingtls
-------------------------------------------------------------------
infoscott
Level 3 user
Posts: 64
Joined: Fri Mar 06, 2015 11:04 pm

Re: Commodity Market

Post by infoscott »

counting wrote:
infoscott wrote:
AI and warehouses...hmm. If they are only looking at sales units, that's a problem. There would need to be some goal seeking code to look upstream from a sales unit to see if there is a locked warehouse that could satisfy future need. Do you know if the AI does any future logistics projections at all when it determines suppliers?
From what I gathered over the years, AI tends to consider supply and demand great deal when choosing suppliers. After that, if there are multiple sources available, it would choose the highest rating one (in current city first, than considering globally), like what a factory would link to if allowing auto-linking (I assume the programming code is the same). And I don't think there is any regression or statistical analysis behind choosing a supplier, but simply by some weight average index comparing with a threshold value. I don't think current AIs have any capability assessing future risks either. And human players can play the same trick over-and-over to the same AI as many times as they like. They only adapt to immediate changes, but never learn from past experiences.

However, AIs sometimes do exhibit certain game theory co-operative behaviors between each others. Like choosing separate products in the same class if both have the same expertise, so they can both be monopoly in some products to gain higher profit, rather than go toe-to-toe in all products. And although obviously willing to go full on price war reacting to human players, but left for AI's own choosing, sometimes AIs will settle to a relative profitable position regarding to their production capacity. Sometimes less aggressive AI setting, make AI corporation perform better if no human try to squeeze them out. I also think a lot of the rule-based conditions are specifically targeting human player's instead of maximizing profit (hence the suicide tendency a lot of players have observed)
Hmmf. It would be a far better simulator if the AI could do its own demand forecasting and then upstream supplier delivery forecasting. In the real world "line down" for manufacturing is a huge deal. I don't see how the AI could play adequately without that kind of forecasting.

If ES's coders could implement that artificial intelligence into the simulator, I would think it would make for a much better executed product, and worthy to the academic community to simulate logistical challenges. :geek:
counting
Level 8 user
Posts: 843
Joined: Wed Jan 29, 2014 12:44 am
Been thanked: 1 time

Re: Commodity Market

Post by counting »

infoscott wrote: Hmmf. It would be a far better simulator if the AI could do its own demand forecasting and then upstream supplier delivery forecasting. In the real world "line down" for manufacturing is a huge deal. I don't see how the AI could play adequately without that kind of forecasting.

If ES's coders could implement that artificial intelligence into the simulator, I would think it would make for a much better executed product, and worthy to the academic community to simulate logistical challenges. :geek:
Well, it's part of the reason why B2C vertical integrated AIs are those who can survived, instead of those "flat" AIs in the long run. A lot of time you could observe AIs with huge profit margin, but less impressive revenue due to them being "specialized" in certain high profit margin product. But when their starting sources run dry, and if no "human players" are kind enough to provide enough upstream, they go bankrupt very quickly (hence the famous all-AIs-go-bankrupt-after-several-decades bug in the original Cap II). This short coming is partially solved by allowing friendly merger (if two AIs have high enough trade between them will have a chance of auto-merger no matter their stock holding position), hence in the long run, the surviving AI corporations are always vertically integrated.

In the real worlds with hundreds, due to thousands of participants, and information isn't free, it would create a much more complex environment. I don't think a single player simulation game is capable of handling it. But regarding more "capable" AIs, I do agree there are long ways to go. It's not easy to compose an AIs to see the "big picture" (not even an easy task for human), but there are ways to integrate machine learning models to improved them (even applied evolutionary algorithm to make them adaptive). Also from an academic perspective, I can surely say that even the most complicated simulation that can yield a testable result, is still just a shallow shadow of the real world conditions. And most of the time we run millions of agents just to test some insignificant details in a theory. It would eat up quite a bit of computing resources just to handle a couple of "smart" AIs if we wish to handling them in real time strategy simulation game. Current Capitalism AIs are like newbies who can only look up a few lines of text-book rules to run their corporations.
-------------------------------------------------------------------
Twitch channel : twitch.tv/ancientbuilder
Youtube channel : www.youtube.com/user/countingtls
-------------------------------------------------------------------
Post Reply