Guild Wars 2

The interesting thing about all this is that it functions just like a commodities market in real life. You can learn from it. I have in the past bought gold Kruggerands and silver rounds from a dealer at a premium over spot price (ask price), as that is what you have to go with as a customer that goes through a middleman, and after waiting for a market shift resold them to another dealer at bid price. Even so, I managed to turn a profit due to a market shift. I managed to pay off a loan with the proceeds. The “tax” in this case was the cut the dealer takes as their premium to offset their own risk, and the 15% TP cut functions in much the same way as it cuts into your expected ROI and has to be accounted for. No matter if you’re the buyer or the seller, in the end the house always wins. The TP even as it stands now siphons gold out of the economy; money changes hands but the total transferred decreases with each transaction.

In the case of the TP, the power buyers and sellers are playing the part of a commodities dealer and get a better deal buying at bid and selling at ask, but even buying at ask and selling at bid can be profitable if you wait until there is a market shift upwards. It all comes down to research and buying when you expect a market shift, which is what specualtion is.

Nike.2631:

John Smith.4610:

Let’s simplify the debate a bit. We have a lot of variables in discussion, let’s say that TP Fees are still a factor, but controlling inflation and gold sinks are not.

Furthermore let’s assume that we all agree that 1c holds a virtual non-value on goods above X price. How do you determine a system that effectively chooses what an appropriate value is that doesn’t
1. Injure the market 2. Discourage use of the TP
Note: this value may be algorithmic, but you must PROVE the values are effective. You may assume you have perfect data of the TP if you need it to support a hypothesis.

The controlling factor could be Time.

The goal is (I think) to reach equilibrium price quickly – but without forcing an overshoot by having a minimum undercut larger than the distance from current lowest offer and actual equilibrium (hence the concerns around algorithmically generated undercuts for lower cost items).

What if the minimum undercut changes as a particular item stagnates? Specifically: the longer its been since an item has sold, the larger the undercut becomes – the item is clearly not at equilibrium and we want the system to speed things along.

Hmm. This is human look-up chart for a moment, rather than a formula, but bear with me.

Use the lowest that is true:
Item has not sold in the last 72 hours — 5% undercut required
If the item has sold in the last 72 hours — 3% undercut required
If the item has sold in the last 12 hours — 2% undercut required
If the item has sold in the last 2 hours — 1% undercut required
If the item has sold in the last 20 minutes — 1c undercut required

High volume items would be unaffected relative to our current model, but things that sit now require a sincere effort to get them back into motion.

If any changes were to be made, I’d vote for these.

John Smith.4610:

It’s unlikely that you’ve ever traded on a market that hasn’t sold an item in the last 2 hours.

Out of curiosity, is that 2 hours for there to be a “sell now to highest offer OR buy now to lower offer” or just counting the buy nows?

((shrug)) I’ve placed mid 50s rares as the lowest asking price and seen them turn a day or two later. Similar stories with some more boring low 70s exotics. I’m sure the addition of essence of luck has sped up the “back corners” lately, but my memory does reach (hazily) into the before time . By purely anecdotal experience there should be far more comparatively worthless dusty corners than there are grinning shields waiting on the rack for someone to come and love them.

Still, it seemed like an interesting approach. But it sounds like the scaling needs to be keyed to average velocity then.

John Smith.4610:

How do you determine a system that effectively chooses what an appropriate value is that doesn’t
1. Injure the market 2. Discourage use of the TP

Experimentally, you minimize price spreads or minimize transaction times with respect to the bid increment – the two should be very similar.

If all we have is archival data, you have to build some model of intertemporal preferences and fit it using the data.

There might be some work on optimal bid increments and things like efficiency and velocity as a function of increment, but that’s for a different set of experts.

As for worries about injuring the market – the optimal bid increment is not 0, even in an extremely efficient market; when you throw in a glaring inefficiency like the 5% listing fee, there’s every reason to believe the optimal bid increment for market efficiency is significantly far from 0.

I disapprove of this thread. The OP’s idea is totally arbitrary and would be a waste of time to implement. Better to spend that dev time working on the next Super Adventure Box.

The problem with the OP is that he sees the TP as a queue. “I listed the item first, so i deserve to sell first.”

But it’s not a queue. it’s a game

Let’s talk exclusively about high priced, slow moving items like legendaries, precursors etc, where a 1c undercut is not particularly costly for the under cutter, but the relisting tax is significant.

Ok, you have a precursor to sell. Theres another two on the market, and you judge their price to be fair. The obvious and optimal action is to undercut the lowest seller by 1 copper.

But wait! What if someone else undercuts you by one copper? How do you beat him? Maybe you should wait until he lists, and then undercut /him/ by one copper?

How long should you wait? What happens if someone buys one of the listed items whilst you are waiting? You missed out, in that case.

So you see, it’s a game, with risks and rewards. If you list first, you get the first chance on a sale, but you risk being undercut. If you wait and undercut, you get the sale, but you risk missing out on potential sellers whilst you are waiting.

It’s not an orderly queue where people line up to get their item sold. it’s a game where not only do you have to make choices about price, you also have to make choices about time, and order.

In a game, you can win and lose. If someone undercuts you, you lost the game. Just like losing a pvp match.

I’m seeing a lot of fancy math and reading a lot of fancy words but what I’m not seeing is a coherent justification as to why these people deserve to have their prices protected against undercutting. The reasons so far essentially add up to I WANT and IT’S NOT FAIR. That’s not a sound reason to reshape how the TP is set up and alter the economics of the game to give yourself special advantages.

You might as well get used to it. In real life people will undercut you and you won’t be able to go on a forum and post about how it should be different and how it should be set up so you won’t lose money (but they do) unless you are a lobbyist for a corporation trying to get a good deal for your company at the expense of others.

Essentially that’s the impression I’m getting. People lobbying for special deals for themselves.

In real life it doesn’t cost 5% to list an item. If I could re-list 1c lower for free every time I was undercut it would be more like real life. In real life I can have my bids automatically raised up to a preset limit each time I am outbid.

neonreaper.4805:

Maybe add an automatic system that allows you to carry a hidden minimum sell amount on your items, and will automatically update your sale if someone undercuts you. Maybe a checkbox that allows you to automatically undercut people up to 1% of your initial asking price.

Now that’s something that may work.

Instead setting a single price, you could set a max and a minimum. In the lists people would only see the maximums, but your minimum value would compete against the maximums set by people after you.

The system would try to sell your offer for the for the highest value over the minimum value you set, that is under a any maximum set after you. And if there’s any maximum under your minimum set after you, that would be sold first.

So if we have 3 players:

1: Sets 200 max and 110 min.
2: Sets 150 max, 125 min.
3: Sets120 max, 90 min.

Player 1 would sell first, for 119, because he set his offer first, and its minimum was under 3’s max., so he sells for 119, right under 3’s max of 120.

Player 3 sells after that for 120, because even though he set his offer later, his max was under 2’s minimum.

And finally, 2 sells for 150, because there’s no other max number under that in the list.

Yup. I think we are finally getting to something that could actually work to make things a bit more fair.

Drago Ivansen.5398:

Brother Grimm.5176:

(Am I having deja vu or has this poor equine been roughed up in here previously?)

Not unless you provide some links

I’d be happy to if the search feature worked, but it doesn’t and I KNOW I’ve seen at least 2 or 3 DOZEN other threads railing against 1c order bumps…..

…..we supposed to now categorize a new separate color system to satisfy one person?

You are aware that there are classifications of types of color blindness and it does not “effect” one person but a non-trivial percentage of the population, right? Your failure in this analogy proved the oppositions point.

I’ve thought a bit more about Nike’s proposal (quoted below) and believe it could be a good starting point.

Use the lowest that is true:
Item has not sold in the last 72 hours — 5% undercut required
If the item has sold in the last 72 hours — 3% undercut required
If the item has sold in the last 12 hours — 2% undercut required
If the item has sold in the last 2 hours — 1% undercut required
If the item has sold in the last 20 minutes — 1c undercut required

Astral Projections.7320 brought up a question regarding why players deserve to have their price protected that I’d like to address as well. I don’t believe a player should be “protected” from undercutting either. I believe the market should be protected from items that are priced too high and Buy Orders that are priced too low. I’m also not sure of the motivation of Nike’s proposal as to whether or not it came from a place of wanting to protect a player’s place in line, but using his idea as a starting point, it could be tweaked in a manner that’s less about protection and more about getting items to move to their equilibrium pricing more quickly. What that means is pushing a price up or down to where the supply and demand of items is nearly equal at a given price point. If prices are pushed to their equilibrium prices, a 1c undercut wouldn’t be much of an issue as the velocity of trading, even in slow markets, would still allow those players selling item that were “first in line” to sell them in a reasonably quick time period.

My tweak to Nike’s idea (assuming perfect data that we as players don’t have access to):

• Determine the Running Average sale price of an item during a set time period (24 to 48 hrs)
  • Total value of items sold / Total number of items sold = Running Average price
    • This takes into account all items sold to Buy Orders
    • This takes into account all items bought from Sell Listings
    • This uses the ACTUAL items sold during the time period
    • If more items are sold to Buy Orders, the average price will be lower and give the impression that the player base “feels” the correct price should be closer to the current Buy Order price
    • If more items are bought from Sell Listings, the average price will be higher and give the impression that the player base “feels” the correct price should be closer to the current Sell Listing Price
• Use the Running Average and the current Sell Listing price to determine the gap between them
  • Current Sell Listing price – Running Average price = Price Gap
    • A larger gap would show that the player base “feels” that the current Sell Listing price is too high
    • A smaller gap would show that the player base “feels” the item is more appropriately priced at the current Sell Listing price
• Use a flat percentage (1%, 5%, 10%, 50%, TBD) of the gap to determine the increment at which the next item can be listed at below the current Sell Listing (rounded up to the nearest 1c with a minimum of 1c)
  • Price Gap * TBD% = Increment
    • The % is unknown at this time and would need to be determined by John Smith and his team
    • Could be tested offline using live data to determine a “best fit”
    • A lower % would bring the item price down more slowly, but more quickly than the current 1c model. Less protection against undercutting.
    • A higher % would bring the item price down more quickly, but may cause serious heartburn amongst players. More protection against undercutting.
• Determine the next lowest price an item can be listed at
  • Current Sell Price – Increment = Next Lowest Sell Listing Price

Continued….

What could this system do to the market (in a positive way)?

• It could move items to their equilibrium prices more quickly
• Trading volumes could increase based on more quickly reaching equilibrium pricing
• Sink a larger total amount of gold based on higher trading volumes, even though prices are lower
• It could offer players protection of their “place in line” on high value, high gap items
  • New sellers would have a tougher choice to make as to whether they want to make more profit, but get in line to do so, or make much less profit to jump ahead in line.
• It wouldn’t change the markets on high velocity, high volume, low gap items as these items are likely near equilibrium price now
• The longer the time period is for the calculated Running Average price, the less likely it could be manipulated.
• Increment changes aren’t based off of single items that are sold, but rather the group of items that were sold during the time period
• The minimum increment for all items will end up being 1c, but with prices at equilibrium values, a 1c undercut could be more tolerable as items will sell more quickly

What could this system do the the market (in a negative way)?

• It could sink less gold per item due to prices being lower
• It could reduce profits of the casual player that gets lucky with a drop
• It could make it more difficult for players to make money through trading as the price increment is now variable depending on the item they’re trading
• It could remove the possiblity of the casual player to make money via trading as the skill set required to profit are at a higher level, requiring more savvy investing and a better set of “tools” than those currently available online
• It could send the entire market into turmoil at the intial implementation and the time to settle down is unknown

I’m SURE there are other positives and negatives to this system, but I should probably get back to work as my company isn’t paying me to post on a video game forum, lol.

Sometimes items have such a large spread (look at precs and legendaries) because there’s too few of the items or because buyers refuse to pay more than X and sellers refuse to sell for less than Y. That wont change regardless of changes to listing methods.

Aidan Savage.2078:

Sometimes items have such a large spread (look at precs and legendaries) because there’s too few of the items or because buyers refuse to pay more than X and sellers refuse to sell for less than Y. That wont change regardless of changes to listing methods.

I’ll provide an example of how it would work for a precursor (specifically Dusk), but some assumptions will need to be made.

Assumptions:

• A total of 10 Dusks have been sold in the last 24 hours (John Smith gave data similar to this a few months back in a different thread)
• More items were sold to Buy Orders than were purchased from Sell Listings (6 from BO / 4 from SL)
• The Buy Order price was 866g03s97c (current highest BO at the time of this post)
• The Sell Listing price was 919g98s99c (current lowest SL at the time of this post)
• Buy Order and Sell Listing prices remain constant, to the prices above, at the time of a new listing
• A fixed % of 25% will be used for the example (keep in mind this % is used on the gap between the average price of the items that actually sold and the current Sell Listing, not the gap between Buy Orders and Sell Listings)

Example:

• Running Average price = Total sales / Total number sold (during the time period of 24h)
  • (6 * 8660397 + 4 * 9199899) / 10 = 887g61s98c
• Price Gap = Current Sell Listing – Running Average Price
  • 9199899 – 8876198 = 32g37s01c
• Increment = Price Gap * 25%
  • 323701 * .25 = 8g09s25c (this calculates to 0.8% of the current Sell Listing)
• Next lowest listing price = Current Sell Listing price – Increment
  • 9199899 – 80925 = 911g89s74c

In this example, the player base has determined that the price they’re willing to sell Dusk’s at is actually closer to the Buy Order price than the current Sell Listing price, because players actually sold Dusk’s that way.

With the calculation of the new increment, undercutting by 1c isn’t a viable option anymore. So, the new seller has to determine if they are willing to lose ~8g to undercut someone, or if they’re willing to wait in line behind the seller that is currently at ~920g.

As more items are sold in one direction or another, the increment will go up or down and the price should move closer to equilibrium.

As the price moves closer to equilibrium, the difference between Buy Orders and Sell Listings closes, the prices of items sold come closer together, the price gap decreases and the increment decreases (to a minimum of 1c).

It all depends on how Buyers and Sellers actually buy items and the prices they pay for them / sell them for.

I thought of another negative impact this system could have.

The increment system could be exploited to still place a Sell Listing at 1c below the current Sell Listing price (but it requires you to have at least 2 of the same item) by:

• Creating a listing at the lower increment price
• Creating Sell Listing(s) ABOVE the newly listed item, but 1c below the previous listing
• Remove the lowest listing

This “exploit” comes at the cost of 5% of the item listed to “get around” the increment.

I don’t see this as a viable option in most cases though as the cost to do so far outweighs the benefits.

Charismatic Harm.9683:

Aidan Savage.2078:

Sometimes items have such a large spread (look at precs and legendaries) because there’s too few of the items or because buyers refuse to pay more than X and sellers refuse to sell for less than Y. That wont change regardless of changes to listing methods.

I’ll provide an example of how it would work for a precursor (specifically Dusk), but some assumptions will need to be made.

Assumptions:

• A total of 10 Dusks have been sold in the last 24 hours (John Smith gave data similar to this a few months back in a different thread)
• More items were sold to Buy Orders than were purchased from Sell Listings (6 from BO / 4 from SL)
• The Buy Order price was 866g03s97c (current highest BO at the time of this post)
• The Sell Listing price was 919g98s99c (current lowest SL at the time of this post)
• Buy Order and Sell Listing prices remain constant, to the prices above, at the time of a new listing
• A fixed % of 25% will be used for the example (keep in mind this % is used on the gap between the average price of the items that actually sold and the current Sell Listing, not the gap between Buy Orders and Sell Listings)

Example:

• Running Average price = Total sales / Total number sold (during the time period of 24h)
  • (6 * 8660397 + 4 * 9199899) / 10 = 887g61s98c
• Price Gap = Current Sell Listing – Running Average Price
  • 9199899 – 8876198 = 32g37s01c
• Increment = Price Gap * 25%
  • 323701 * .25 = 8g09s25c (this calculates to 0.8% of the current Sell Listing)
• Next lowest listing price = Current Sell Listing price – Increment
  • 9199899 – 80925 = 911g89s74c

In this example, the player base has determined that the price they’re willing to sell Dusk’s at is actually closer to the Buy Order price than the current Sell Listing price, because players actually sold Dusk’s that way.

With the calculation of the new increment, undercutting by 1c isn’t a viable option anymore. So, the new seller has to determine if they are willing to lose ~8g to undercut someone, or if they’re willing to wait in line behind the seller that is currently at ~920g.

As more items are sold in one direction or another, the increment will go up or down and the price should move closer to equilibrium.

As the price moves closer to equilibrium, the difference between Buy Orders and Sell Listings closes, the prices of items sold come closer together, the price gap decreases and the increment decreases (to a minimum of 1c).

It all depends on how Buyers and Sellers actually buy items and the prices they pay for them / sell them for.

+1. Excellent.

I don’t log in for a day, and I get lost in the current flow of the debate. Let me simplify this, since a lot of people have put up good arguments against this idea.

Q) Why should I have to do unnecessary maths in order to sell an item?

It’s far easier to think “1 Copper less” than it is to think “1% less”. Putting artificial limits just makes selling more complicated. When I’m out and about, I want to just effortlessly post items for sale, on my own terms.

There’s currently no reasons thus far that say the current system is broken, outside of a few players’ personal preferences that 1 Copper increments are bad. They have offered no evidence that a change is necessary in 7 pages of this debate.

Smooth Penguin.5294:

I don’t log in for a day, and I get lost in the current flow of the debate. Let me simplify this, since a lot of people have put up good arguments against this idea.

Q) Why should I have to do unnecessary maths in order to sell an item?

It’s far easier to think “1 Copper less” than it is to think “1% less”. Putting artificial limits just makes selling more complicated. When I’m out and about, I want to just effortlessly post items for sale, on my own terms.

There’s currently no reasons thus far that say the current system is broken, outside of a few players’ personal preferences that 1 Copper increments are bad. They have offered no evidence that a change is necessary in 7 pages of this debate.

To answer your question:
A) You don’t have to do the math, the system would do it for you.

As to why this discussion is currently ongoing:
Since you’ve been logged out for a day, you missed this post:

John Smith.4610:

Let’s simplify the debate a bit. We have a lot of variables in discussion, let’s say that TP Fees are still a factor, but controlling inflation and gold sinks are not.

Furthermore let’s assume that we all agree that 1c holds a virtual non-value on goods above X price. How do you determine a system that effectively chooses what an appropriate value is that doesn’t
1. Injure the market 2. Discourage use of the TP
Note: this value may be algorithmic, but you must PROVE the values are effective. You may assume you have perfect data of the TP if you need it to support a hypothesis.

After his post, the discussion has evolved.

Charismatic Harm.9683:

Smooth Penguin.5294:

I don’t log in for a day, and I get lost in the current flow of the debate. Let me simplify this, since a lot of people have put up good arguments against this idea.

Q) Why should I have to do unnecessary maths in order to sell an item?

It’s far easier to think “1 Copper less” than it is to think “1% less”. Putting artificial limits just makes selling more complicated. When I’m out and about, I want to just effortlessly post items for sale, on my own terms.

There’s currently no reasons thus far that say the current system is broken, outside of a few players’ personal preferences that 1 Copper increments are bad. They have offered no evidence that a change is necessary in 7 pages of this debate.

To answer your question:
A) You don’t have to do the math, the system would do it for you.

As to why this discussion is currently ongoing:
Since you’ve been logged out for a day, you missed this post:

John Smith.4610:

Let’s simplify the debate a bit. We have a lot of variables in discussion, let’s say that TP Fees are still a factor, but controlling inflation and gold sinks are not.

Furthermore let’s assume that we all agree that 1c holds a virtual non-value on goods above X price. How do you determine a system that effectively chooses what an appropriate value is that doesn’t
1. Injure the market 2. Discourage use of the TP
Note: this value may be algorithmic, but you must PROVE the values are effective. You may assume you have perfect data of the TP if you need it to support a hypothesis.

After his post, the discussion has evolved.

Ah thanks. I was lazy and probably missed that as I skimmed. So basically, John challenged the complainers to prove the 1% minimum is effective and doesn’t harm the market. So then, even if you come of with ways to automatically limit bid increments, I don’t think anyone has come up with any evidence that supports the effectiveness of this on the market as a whole.

Smooth Penguin.5294:

Ah thanks. I was lazy and probably missed that as I skimmed. So basically, John challenged the complainers to prove the 1% minimum is effective and doesn’t harm the market. So then, even if you come of with ways to automatically limit bid increments, I don’t think anyone has come up with any evidence that supports the effectiveness of this on the market as a whole.

The assumption I’ve made is that the “evidence” that supports the idea is evidence we, as players, can’t provide. What I mean by that is we don’t know if the current price is actually the equilibrium price.

My idea (above) also assumes that the equilibrium price for an item IS the best price and the algorithm used attempts to push the current price to that price quickly, yet reasonably.

Charismatic Harm.9683:

Smooth Penguin.5294:

Ah thanks. I was lazy and probably missed that as I skimmed. So basically, John challenged the complainers to prove the 1% minimum is effective and doesn’t harm the market. So then, even if you come of with ways to automatically limit bid increments, I don’t think anyone has come up with any evidence that supports the effectiveness of this on the market as a whole.

The assumption I’ve made is that the “evidence” that supports the idea is evidence we, as players, can’t provide. What I mean by that is we don’t know if the current price is actually the equilibrium price.

My idea (above) also assumes that the equilibrium price for an item IS the best price and the algorithm used attempts to push the current price to that price quickly, yet reasonably.

In regards to the assumed equilibrium prices. Even if we know the definite answer to what each price is for each item, the next question will always come up: Why are we forced to sell an item in the area of such a price? Yes it’s good for buyers, because then the item is sold towards a price to their liking. But that comes at the expense of sellers who want to sell items for prices “they” feel is valid.

Any change should benefit all players. It should not cater to the ones who want something for a good price, or the ones who feel a certain percentage bid is more justifiable on higher priced goods.

Smooth Penguin.5294:

In regards to the assumed equilibrium prices. Even if we know the definite answer to what each price is for each item, the next question will always come up: Why are we forced to sell an item in the area of such a price? Yes it’s good for buyers, because then the item is sold towards a price to their liking. But that comes at the expense of sellers who want to sell items for prices “they” feel is valid.

Any change should benefit all players. It should not cater to the ones who want something for a good price, or the ones who feel a certain percentage bid is more justifiable on higher priced goods.

The only prices that a seller aren’t allowed to post a new item at are the ones between the current minimum increment and the current Sell Listing.

A seller is able to post at any price between the current highest Buy Order and minimum sell increment, or at the current Sell Listing and any price above that.

Keep in mind that if the average running value of the sales is at or above current Sell Listing, the minimum bid increment would be 1c.

The calculated increment is done in such a way that shows an undercut increment of 1c, at this point in time for increments above 1c, would create a Sell Listing that is too far away from the equilibrium price because actual sales of items, during the time period used to calculate the increment, have determined that Buyers and Sellers (the player base) have deemed it so.

Without actual transactions, you have neither Buyers nor Sellers. You have offers and listings.

To address your question: Why are we forced to sell an item in the area of such a price?

You state that it’s good for buyers and you are correct. The prices of items should be lower than they are now, but by how much remains in question.

As for sellers, it makes the 5% listing fee a little smaller because of lower prices and shows you a price at which the item your listing is more likely to sell, providing less risk, although accompanied by less reward. If the seller wants more reward and more risk, they can very easily list at the current Sell Listing price.

You also forget there are many cases where items with a large spread (buys @ 40s and sells at 70-80s for example) still sees purchases and sales at those prices. Even I’m guilty of that at times. I’d rather sell an item, usually an exotic weapon, for the buy price which can be 1-2g lower than the current sell listing because I’d rather not wait. Conversely, some people would rather buy the item at it’s 2-3g price instead of placing a buy order at 1g1s1c.
While such cases wouldnt happen in a “perfect” market, or very rarely, they do happen often enough that they’re the driving force between prices rising and falling on items with a large spread.

For large spreads, I’ll use a hypothetical situation.

Player A buys Dusk for the listed price of 799 Gold 99 Silver 99 Copper.
Player B buys Dusk for the listed price of 800 Gold.
Player C buys Dusk when someone fills their Buy Order for 500 Gold.

If the market is moving at this rate, wouldn’t a 1% implemented minimum increment hurt the market? Remember, the seller is part of this market. The person who sold his Dusk to the highest Buy Order hurt himself in exchange for a quick sale.

Precursors and legendaries are two items types that I believe should be left outside the consideration of an incremental listing like what’s being suggested. Particularly due to how they’re obtained, as well as their value. Some legendary weapons I can create for less gold than what they cost on the TP, but due to other factors such as time, or other currency being lacked, I might elect to pay the TP price for the legendary and write off that gold difference as the value I assign to the items that require karma, dungeon tokens, or other limited materials.

Charismatic Harm.9683:

Aidan Savage.2078:

Sometimes items have such a large spread (look at precs and legendaries) because there’s too few of the items or because buyers refuse to pay more than X and sellers refuse to sell for less than Y. That wont change regardless of changes to listing methods.

I’ll provide an example of how it would work for a precursor (specifically Dusk), but some assumptions will need to be made.

Assumptions:

• A total of 10 Dusks have been sold in the last 24 hours (John Smith gave data similar to this a few months back in a different thread)
• More items were sold to Buy Orders than were purchased from Sell Listings (6 from BO / 4 from SL)
• The Buy Order price was 866g03s97c (current highest BO at the time of this post)
• The Sell Listing price was 919g98s99c (current lowest SL at the time of this post)
• Buy Order and Sell Listing prices remain constant, to the prices above, at the time of a new listing
• A fixed % of 25% will be used for the example (keep in mind this % is used on the gap between the average price of the items that actually sold and the current Sell Listing, not the gap between Buy Orders and Sell Listings)

Example:

• Running Average price = Total sales / Total number sold (during the time period of 24h)
  • (6 * 8660397 + 4 * 9199899) / 10 = 887g61s98c
• Price Gap = Current Sell Listing – Running Average Price
  • 9199899 – 8876198 = 32g37s01c
• Increment = Price Gap * 25%
  • 323701 * .25 = 8g09s25c (this calculates to 0.8% of the current Sell Listing)
• Next lowest listing price = Current Sell Listing price – Increment
  • 9199899 – 80925 = 911g89s74c

In this example, the player base has determined that the price they’re willing to sell Dusk’s at is actually closer to the Buy Order price than the current Sell Listing price, because players actually sold Dusk’s that way.

With the calculation of the new increment, undercutting by 1c isn’t a viable option anymore. So, the new seller has to determine if they are willing to lose ~8g to undercut someone, or if they’re willing to wait in line behind the seller that is currently at ~920g.

As more items are sold in one direction or another, the increment will go up or down and the price should move closer to equilibrium.

As the price moves closer to equilibrium, the difference between Buy Orders and Sell Listings closes, the prices of items sold come closer together, the price gap decreases and the increment decreases (to a minimum of 1c).

It all depends on how Buyers and Sellers actually buy items and the prices they pay for them / sell them for.

Bravo, more food for thought +1

Smooth Penguin.5294:

For large spreads, I’ll use a hypothetical situation.

Player A buys Dusk for the listed price of 799 Gold 99 Silver 99 Copper.
Player B buys Dusk for the listed price of 800 Gold.
Player C buys Dusk when someone fills their Buy Order for 500 Gold.

If the market is moving at this rate, wouldn’t a 1% implemented minimum increment hurt the market? Remember, the seller is part of this market. The person who sold his Dusk to the highest Buy Order hurt himself in exchange for a quick sale.

You are correct, the seller is part of the market. So are all of the other sellers, the 3 buyers you listed above….and all of the other buyers. While the one seller sold his item at a much lower price than the other 2 sellers, he still sold his item, so his item counts. Player C deemed the value of his item at 500g, while Players A and B valued theirs at ~800g. All of them sold, so all of their values are viable.

Using your hypothetical sales, the calculated increment would come to approximately 25g according to the system I’ve proposed, or about 3% of the current Sell Listing price.

Since most of the comments in this thread have been focused on “how much is enough / too much to undercut by”, my initial explanation of the new system focused on the Sell Listing side. There is no reason whatsoever that this minimum increment shouldn’t be placed on the Buy Order side as well. It would even more efficiently move the Sell Listings and Buy Orders toward an equilibrium price.

If that is the case and there remains a Sell Listing at 800g and a Buy Order at 500g, the next Sell Listing would be set to 775g and the next Buy Order would be set to 525g.

Side note:
I’ve been having a conversation with some co-workers and we’ve been discussing the basics of economics. Here’s what it boils down to:

• Buyers want to acquire an infinitely large amount of “things” for an infinitely small amount of money. Buyers will be displeased at any price higher than that.
• Sellers want to sell an infinitely small number of “things” for an infinitely large amount of money. Sellers will be displeased at any price lower than that.
• The equilibrium price of a “thing” is the point where both Buyers and Sellers are equally displeased with the price.

Charismatic Harm.9683:

Smooth Penguin.5294:

For large spreads, I’ll use a hypothetical situation.

Player A buys Dusk for the listed price of 799 Gold 99 Silver 99 Copper.
Player B buys Dusk for the listed price of 800 Gold.
Player C buys Dusk when someone fills their Buy Order for 500 Gold.

If the market is moving at this rate, wouldn’t a 1% implemented minimum increment hurt the market? Remember, the seller is part of this market. The person who sold his Dusk to the highest Buy Order hurt himself in exchange for a quick sale.

You are correct, the seller is part of the market. So are all of the other sellers, the 3 buyers you listed above….and all of the other buyers. While the one seller sold his item at a much lower price than the other 2 sellers, he still sold his item, so his item counts. Player C deemed the value of his item at 500g, while Players A and B valued theirs at ~800g. All of them sold, so all of their values are viable.

Using your hypothetical sales, the calculated increment would come to approximately 25g according to the system I’ve proposed, or about 3% of the current Sell Listing price.

Since most of the comments in this thread have been focused on “how much is enough / too much to undercut by”, my initial explanation of the new system focused on the Sell Listing side. There is no reason whatsoever that this minimum increment shouldn’t be placed on the Buy Order side as well. It would even more efficiently move the Sell Listings and Buy Orders toward an equilibrium price.

If that is the case and there remains a Sell Listing at 800g and a Buy Order at 500g, the next Sell Listing would be set to 775g and the next Buy Order would be set to 525g.

Side note:
I’ve been having a conversation with some co-workers and we’ve been discussing the basics of economics. Here’s what it boils down to:

• Buyers want to acquire an infinitely large amount of “things” for an infinitely small amount of money. Buyers will be displeased at any price higher than that.
• Sellers want to sell an infinitely small number of “things” for an infinitely large amount of money. Sellers will be displeased at any price lower than that.
• The equilibrium price of a “thing” is the point where both Buyers and Sellers are equally displeased with the price.

So then we go back to how your system harms sellers. Your idea effectively forced me to lose ~21.25 Gold where Dusk was sold at:

800 – 120 (listing fee + tax) = 680
775 – 116.25 (listing fee + tax) = 658.75 <—- forced minimum undercut

The person who sold his at 500 took a huge hit, but that was his own choice. I took a hit as well, but that was not by choice. Why should I be forced to lose money, when theoretically both 800 and 799.99.99 price ranges would sell?

Smooth Penguin.5294:

Why should I be forced to lose money, when theoretically both 800 and 799.99.99 price ranges would sell?

Because John Smith asked us to explore systems where 1c is a negligible cost for skipping ahead in the FIFO que?

Its sort of the ground rules for the current discussion .

Nike.2631:

Smooth Penguin.5294:

Why should I be forced to lose money, when theoretically both 800 and 799.99.99 price ranges would sell?

Because John Smith asked us to explore systems where 1c is a negligible cost for skipping ahead in the FIFO que?

Its sort of the ground rules for the current discussion .

And you still could list at 800g, you would only be 2nd in line. But as you stated that 800g listings would theoretically sell, this shouldnt be an issue.

Smooth Penguin.5294:

So then we go back to how your system harms sellers. Your idea effectively forced me to lose ~21.25 Gold where Dusk was sold at:

800 – 120 (listing fee + tax) = 680
775 – 116.25 (listing fee + tax) = 658.75 <—- forced minimum undercut

The person who sold his at 500 took a huge hit, but that was his own choice. I took a hit as well, but that was not by choice. Why should I be forced to lose money, when theoretically both 800 and 799.99.99 price ranges would sell?

The system I proposed didn’t “force” you to do anything other than to make a more difficult decision. The system makes you choose between: “Do I take a significantly lower amount of money for my product to jump ahead of the current line?” or “Do I wait in line behind the other guy to get a larger amount of money?”

The answer to those questions lies in your understanding of the market. If you feel the market moves fast enough for your item to sell at the higher amount, why not get in line? If you feel it doesn’t and you won’t see a return on your investment in a time period suited to you, then take less for it and jump the line.

I believe many of your contradictions to the idea I’ve proposed come from a place of “I make money trading on the Trading Post and I don’t want my profits impacted.” That’s a perfectly fine and legitimate opinion to have. I make my money on the Trading Post too, but I am willing to sacrifice some of those potential earnings if the system, and the game, becomes more fair to everyone. Fair is a subjective term though. What’s fair to one player is not necessarily fair to another. I guess it’s not about making a system that’s equally fair, because that would be extremely difficult. It’s about making a system that’s equally unfair.

If this system were to be implemented, the Trading Post will not become instantly unprofitable. It would just be more difficult to make profits. Savvy traders should see that as a good thing….and a challenge. If it becomes more difficult for the average trader to make money, there’s less competition. Above average traders should now have an even greater advantage. Traders with extremely high level trading skills and tools will still look at the Trading Post as a candy store.

A FIFO system would only work if item quality degraded over time, or the items have such a high movement speed with a very minute spread that even 1c lower is profit loss (some crafting materials).

The problem with forced undercuts is that IF I’m selling an item that has a really high value, if I’m forced to lose 10’s of gold JUST to list it, I may as well never sell the item because I’m going to be forced to wait until stock of the item sells so that I can list at the price I want. All because some system made the decision that losing that money is a nonvalue to protect some scrub that listed on the TP.

Why should someone be afforded such protections when participating in a market specifically opens you up to such risks? Undercutting, buyouts, and market crashing are all risks you are open to, and Anet should not protect you from them. Asking for such protections is basically asking Anet to hand-hold you while selling your items.

Aidan Savage.2078:

A FIFO system would only work if item quality degraded over time, or the items have such a high movement speed with a very minute spread that even 1c lower is profit loss (some crafting materials).

The problem with forced undercuts is that IF I’m selling an item that has a really high value, if I’m forced to lose 10’s of gold JUST to list it, I may as well never sell the item because I’m going to be forced to wait until stock of the item sells so that I can list at the price I want. All because some system made the decision that losing that money is a nonvalue to protect some scrub that listed on the TP.

Why should someone be afforded such protections when participating in a market specifically opens you up to such risks? Undercutting, buyouts, and market crashing are all risks you are open to, and Anet should not protect you from them. Asking for such protections is basically asking Anet to hand-hold you while selling your items.

This system doesn’t force you to list at the undercut price. You can still list your item at the current Sell Listing price. IF you WANT to undercut, the system attempts to push the price toward equilibrium based on actual sales over a set period of time by calculating the increment at which you can undercut by.

You asked why someone should be afforded such protection. Ask yourself why you should be able to jump ahead of the line for essentially the same price. The same goes for Buy Orders. Outbidding by 1c is rampant on the Buy Order side as well. This isn’t just about sellers. It’s about creating a more efficient market that’s equally unfair to all.

Have you ever been undercut? Were you ever upset about it? Did you understand that it’s just the way the game is played? Did you find that when you were undercut, someone undercut them, and then someone undercut them, then again….and again? Now you’re like 5th or 6th in line. With the system I’ve proposed, there’s still the option for other sellers to undercut you, but what’s the likelihood they will take less money just to jump the line? Maybe one or two of them will, but the other 4….they’ll get in line behind you. That means your item sells sooner….and for more money.

Regarding equilibrium price and assuming this is something desirable, what if the listing fee were simply removed for re-listing items? E.g. if you have a Dusk listed for 800g and someone lists it for 799g99s99c, you could click a button to re-list your Dusk for 799g99s98c without having to remove it and pay the additional listing fee. This might quickly drive prices to the aforementioned equilibrium. It would also reduce the significance of undercutting as people would effectively be forced to list their item at the lowest price at which they are really willing to sell it. It would also probably be simpler to implement something like this than calculating and setting a minimum price increment. It also would not prevent anyone from asking or bidding at exactly the price they want. Not sure if it would result in a bunch of extra TP server processing though as people flip back and forth undercutting their way to equilibrium. But it would be simple enough to implement a timer to prevent re-listing too frequently, e.g. once per minute.

Free re-listing might also get some of the lingering junk off the TP as people could re-list for free rather than leave it there in hopes that something they listed in a falling market will eventually get sell.

There are probably a bunch of holes in this idea, for which I will blame not enough coffee yet this morning…

Charismatic Harm.9683:

If it becomes more difficult for the average trader to make money, there’s less competition. Above average traders should now have an even greater advantage. Traders with extremely high level trading skills and tools will still look at the Trading Post as a candy store.

Doesn’t this classify as injuring the market? I could be wrong with my interpretation of the market, but, in general, people are risk adverse. A complicated market increases risk. To avoid the extra risk, more buy orders will be filled by sellers and sell orders by buyers. The equilibrium price is less likely to be reached as there will be less middle ground.

I think, theoretically, the concept holds value in determining and pushing equilibrium pricing for slow, high price markets. However, other markets have more buyers and sellers and would be more likely to see a greater disparity between buy and sell prices, rather than the intended push to one price point.

William Bradley Knight.2609:

Regarding equilibrium price and assuming this is something desirable, what if the listing fee were simply removed for re-listing items? E.g. if you have a Dusk listed for 800g and someone lists it for 799g99s99c, you could click a button to re-list your Dusk for 799g99s98c without having to remove it and pay the additional listing fee. This might quickly drive prices to the aforementioned equilibrium. It would also reduce the significance of undercutting as people would effectively be forced to list their item at the lowest price at which they are really willing to sell it. It would also probably be simpler to implement something like this than calculating and setting a minimum price increment. It also would not prevent anyone from asking or bidding at exactly the price they want. Not sure if it would result in a bunch of extra TP server processing though as people flip back and forth undercutting their way to equilibrium. But it would be simple enough to implement a timer to prevent re-listing too frequently, e.g. once per minute.

Free re-listing might also get some of the lingering junk off the TP as people could re-list for free rather than leave it there in hopes that something they listed in a falling market will eventually get sell.

There are probably a bunch of holes in this idea, for which I will blame not enough coffee yet this morning…

This could possibly work too if:

• The 5% listing fee remained in place for the item when it’s originally listed
• If an item is removed from the Trading Post (cancelled listing, not re-listed) and the seller wants to re-list the item again, the 5% listing fee would be required.
• A “free re-list” would only be granted if the new price is lower than the current price they’ve listed their item at.
• A “time between re-lists” was implemented (as you suggested).

I like the idea of “free re-lists” as it would still keep players from using the Trading Post as storage. It would also not penalize a seller twice, three times, or more, for lowering their price to get their item to sell.

A QoL change to the Buy Order side would be a nice compliment to this as well. Allow players with existing Buy Orders to increase the price on it, if they have enough money to cover increase. Currently, the player has to cancel their Buy Order, pick up their money (if they don’t have enough to place the new one in full), then submit a new Buy Order.

Overall….I like it. It’s simple and could be effective at moving prices to equilibrium as well.

brittitude.1983:

Charismatic Harm.9683:

If it becomes more difficult for the average trader to make money, there’s less competition. Above average traders should now have an even greater advantage. Traders with extremely high level trading skills and tools will still look at the Trading Post as a candy store.

Doesn’t this classify as injuring the market?

I don’t believe this injures the market. It just makes it more difficult for players to make money by using the market inefficiencies to their advantage. The system would be attempting to correct those inefficiencies by moving prices to the equilibrium point.

The average player with little analysis skill that makes money using the Trading Post does so through “flipping”. “Flipping” is the easiest type of market to get into, thus has the most competition. “Flipping” is simple math. Profit = 0.85 * Sell Price – Buy Price. “Flipping” is also the reason that many players say “It’s easy to make a metric tonne of money off the Trading Post.” Savvy traders know that you can make a little bit of money a lot of times by flipping, but you won’t get super rich doing it. There comes a point where you just can’t invest enough money to make a size-able return for the amount of work you have to put in. Players that are making lots of money aren’t doing it by flipping. That’s why I said that players with a higher level of skill will still see the Trading Post as a candy store. They’re already not flipping, so new markets will likely open up to them, like crafting.

Reposting my suggestion because I honestly think it’d work reasonably well and be simpler to implement and transition to than anything else suggested so far:

Listings (bids and asks) on the market should only be possible in the following price increments:

1s and below -> increments of 1c
10s and below -> increments of 10c
1g and below -> increments of 1s
10g and below -> increments of 10s

And so forth.

The granularity can be adjusted to say 0.1% if 1% is not enough (i.e. starting increment becomes 1c at 10s and below).

Kaon.7192:

Reposting my suggestion because I honestly think it’d work reasonably well and be simpler to implement and transition to than anything else suggested so far:

Listings (bids and asks) on the market should only be possible in the following price increments:

1s and below -> increments of 1c
10s and below -> increments of 10c
1g and below -> increments of 1s
10g and below -> increments of 10s

Actually, I believe that William Bradley Knight’s suggestion is the most elegant and easiest to implement. Your price increments are no more or less arbitrary than any other static increment, such as 1c (as it is now) or 1% (as has been suggest).

Using your 10g and below increment as an example:

• A 10g Sell Listing has a 10s Listing Fee, or 1% of the Sell Listing
• A 1g00s01c Sell Listing has a 10s Listing Fee, or 9.9999999% of the Sell Listing

If what you’re wanting is a 1% undercut increment, it would be better served using a static 1% rather than “buckets” as you’ve suggested.

Volume based is still interesting, but the ability to “shift” your asking price downwards without de-listing/re-list the item has potential.

It would cause a lot of 5% fees to not exit the system, but it could drive things towards equilibrium much faster in some markets.

Basically we let the ‘early posters’ and ‘undercutters’ beat each other to death for the buyers’ amusement. I’m so ok with that .

Nike.2631:

Volume based is still interesting, but the ability to “shift” your asking price downwards without de-listing/re-list the item has potential.

It would cause a lot of 5% fees to not exit the system, but it could drive things towards equilibrium much faster in some markets.

Basically we let the ‘early posters’ and ‘undercutters’ beat each other to death for the buyers’ amusement. I’m so ok with that .

I agree. There would be a lot less gold sunk through 5% fees at higher prices and from those re-listing currently, but would it be an overall increase or decrease of gold sunk through the Trading Post based on a higher trading volume, meaning more items sold, at lower prices, thereby sinking gold at the actual sale removing an additional 10% from the economy?

It’s interesting to think about. That’s for sure.

I’m all for a more healthy economy where more Buyers are getting the items they want and more Sellers are selling items. I think the faster an item moves to its equilibrium price, the more healthy that item is.

Charismatic Harm.9683:

Actually, I believe that William Bradley Knight’s suggestion is the most elegant and easiest to implement. Your price increments are no more or less arbitrary than any other static increment, such as 1c (as it is now) or 1% (as has been suggest).

Using your 10g and below increment as an example:

• A 10g Sell Listing has a 10s Listing Fee, or 1% of the Sell Listing
• A 1g00s01c Sell Listing has a 10s Listing Fee, or 9.9999999% of the Sell Listing

If what you’re wanting is a 1% undercut increment, it would be better served using a static 1% rather than “buckets” as you’ve suggested.

Yes I agree that William Bradley Knight’s solution will be an improvement in market price agility over the status quo, but I don’t think it fully addresses a different fundamental issue discussed in this thread. Drago Ivansen has done the math and explained it much better than I could so I’ll just link his post and quote his conclusion and expand on it.

https://forum-en.gw2archive.eu/forum/game/bltc/Min-1-price-difference/page/5#post3873034

The change in willingness to buy is equal to a constant benefit divided by the product of the costs given all of the above assumptions. This proves as cost increases to a very large number the difference in willingness to buy will approach zero.

The problem with the status quo is that any later listing can bypass the FIFO queue at a relative cost that approaches 0 as the price of an item increases.

The change suggested by William Bradley Knight would simply turn the Last List First Fill system we have right now into a Last Relist First Fill system, that still benefits TP bots/notification users at the expense of regular buyers and sellers. Again, it’s admittedly better than the status quo where even relisting is prohibitively costly and slows the convergence to equilibrium price, but ideally I’d like to see a solution that addresses the other fundamental problem as well.

Kaon.7192:

The change suggested by William Bradley Knight would simply turn the Last List First Fill system we have right now into a Last Relist First Fill system, that still benefits TP bots/notification users at the expense of regular buyers and sellers. Again, it’s admittedly better than the status quo where even relisting is prohibitively costly and slows the convergence to equilibrium price, but ideally I’d like to see a solution that addresses the other fundamental problem as well.

This is true, but the point that you’re missing is:

• Once prices reach equilibrium, the benefits of undercutting no longer exist because Buyers and Sellers have reached a mutually beneficial price for an item. It’s only a Last Relist / First Fill state until the price reaches equilibrium.

Currently, it’s exceedingly beneficial for new sellers to undercut, especially on high valued items, because the new seller KNOWS it will cost the old seller an additional 5% to pull his listing and undercut him in return. This new seller gets to jump ahead in line for virtually the same price. The only risk this new seller has is being undercut by another new seller.

In either William Bradley Knight’s, or my model, undercutting is discouraged by pushing the price to equilibrium. His model does it in a manner that allows endless undercutting by everyone “for free”. My model associates an increased cost to do so.

The equilibrium point is the point where the Supply and Demand curves cross. It’s the point where the quantity of items buyers want equals the quantity of items sellers are willing to sell.

If there is a sudden increase in Demand for an item, the price will push up to accommodate for it. If there is a sudden increase in Supply of an item, the price will push down.

William’s model actually does this better than mine by making the Sell Listing side of the equation act more like the Buy Order side. Currently, there is no real penalty for outbidding someone with a higher Buy Order, outside of a little bit more money, but there Buyer is already willing to pay that, so it’s a moot point. Currently, on the Sell Listing side, there is the penalty of an additional 5% of the Sell Listing to lower your price. His model removes that penalty and is the reason his model works better to push the price to equilibrium faster than mine.

Charismatic Harm.9683:

brittitude.1983:

Charismatic Harm.9683:

If it becomes more difficult for the average trader to make money, there’s less competition. Above average traders should now have an even greater advantage. Traders with extremely high level trading skills and tools will still look at the Trading Post as a candy store.

Doesn’t this classify as injuring the market?

I don’t believe this injures the market. It just makes it more difficult for players to make money by using the market inefficiencies to their advantage. The system would be attempting to correct those inefficiencies by moving prices to the equilibrium point.

This does injure the market. You dont make it difficult for players to make money using inefficiencies, you also make it difficult for them to make money just selling what they’re looting or crafting. If you want to push towards a minimal spread of buy/sell prices, you need to increase demand and supply of the item. Artificially closing the gap via forced relisting prices is not “reaching equilibrium,” it’s attempting to induce a pattern of behavior that particular market clearly has no intention of doing. True equilibrium is the market not doing wild fluctuations in prices in a short order of time, and then persisting at those prices.
The perfect example of this is Radiant Dust during the LA ls week. It was extremely stable at 1-2s each, and the introduction of the recipe caused a massive spike in demand such that the price rose to over 20s each, and then induced a nearly instantaneous dump of supply that suppressed the price back down to 2-3s. All in about 10 minutes. That’s a shift in equilibrium due to the increased demand and the subsequent surplus of supply.
Many items on the market are already in equilibrium because the prices that buy orders and sell orders are at do not fluctuate much. The fact that the state of equilibrium also includes a spread that is often 50% of the highest buy order is irrelevant. Not to mention that outside of external influences, in which case it’s new recipes anet adds, those prices will not shift because the supply and demand are consistently equal. The only way to bring those prices together by natural market movement is to increase the demand and supply of the item.

The average player with little analysis skill that makes money using the Trading Post does so through “flipping”. “Flipping” is the easiest type of market to get into, thus has the most competition. “Flipping” is simple math. Profit = 0.85 * Sell Price – Buy Price. “Flipping” is also the reason that many players say “It’s easy to make a metric tonne of money off the Trading Post.” Savvy traders know that you can make a little bit of money a lot of times by flipping, but you won’t get super rich doing it. There comes a point where you just can’t invest enough money to make a size-able return for the amount of work you have to put in. Players that are making lots of money aren’t doing it by flipping. That’s why I said that players with a higher level of skill will still see the Trading Post as a candy store. They’re already not flipping, so new markets will likely open up to them, like crafting.

Flipping can make a lot of money if you’re able to move large quantity on a regular basis. If I’m making 2s every time I flip 1x of Item A, and item A sells on a consistent and regular basis, I can make 200g off Item A if I move enough of it quickly enough that someone else does not impede on my attempt. The low profit margins on some items make them unusable for large-scale flipping because if something goes wrong, you’re stuck at a no profit value or a loss at worst.
Safer money can be found on the tradepost by supplying the needs of crafters, whether those that are leveling to 400 or to 500, or those crafting ascended gear.

Rather than some system of minimum undercut wouldn’t a system that supports a bid/offer range solve the problem, and make trading a lot more interesting? If every market order came with a min and max price that adjusted to all future offers, then your 5% investment would be protected as a seller who got there first, and as a buyer you would be spared continuously re-posting your offers.

This way you could stake your claim on any range from 0%-n% and anyone who attempts to undercut you has the same choice and doesn’t get in front for ‘free’ with a 1c undercut.

Elric Of Melnibone.4781:

Rather than some system of minimum undercut wouldn’t a system that supports a bid/offer range solve the problem, and make trading a lot more interesting? If every market order came with a min and max price that adjusted to all future offers, then your 5% investment would be protected as a seller who got there first, and as a buyer you would be spared continuously re-posting your offers.

This way you could stake your claim on any range from 0%-n% and anyone who attempts to undercut you has the same choice and doesn’t get in front for ‘free’ with a 1c undercut.

Making it “more interesting” is not a solution. Nor is there any reason to “protect” someone’s listing just because they’re the current lowest listing. If you’re being undercut regularly, that’s the market’s way of saying “cough up a lower price if you want to sell, this price aint working for you or others.” Nor does it address the other factors of a market.

Charismatic Harm.9683:

Your price increments are no more or less arbitrary than any other static increment, such as 1c (as it is now) or 1% (as has been suggest).

Using your 10g and below increment as an example:

• A 10g Sell Listing has a 10s Listing Fee, or 1% of the Sell Listing
• A 1g00s01c Sell Listing has a 10s Listing Fee, or 9.9999999% of the Sell Listing

I’m not sure if I follow your thinking here, but let me give it a shot:

You seem to be seeing the mandatory increments in this system as a form of “Pay to Undercut”, correct? I suppose that is a valid way of looking at it, so I will also refer to it as such for the rest of this post. But I wouldn’t agree that it is no less arbitrary than a 1c increment across all price points.

While it’s not FULLY dynamic as some of the other suggestions (I’ll get to why a fully dynamic system isn’t feasible in a moment), the system I proposed is definitely less arbitrary than a static 1c “cost to undercut” across all price points.

In the current system, the relative cost to undercut does not have a lower bound, this leads to the fundamental problem I discussed in my previous post (any later listing can bypass the FIFO queue at a relative cost that approaches 0 as the price of an item increases). But in the system I suggested, the lower bound will be equal to the price increment granularity (which can be adjusted), i.e. the lowest possible cost to undercut will be 1% of the current price with the price granularity set to 1%, and 0.1% if it’s set to 0.1%.

I believe this will adequately address the problem of cost to undercut approaching 0% as price increases.

Charismatic Harm.9683:

If what you’re wanting is a 1% undercut increment, it would be better served using a static 1% rather than “buckets” as you’ve suggested.

From a programmer’s perspective, this “static” 1% you speak of can actually be prohibitively complex when you start considering the implementation details of such a system.

Firstly there is no such thing as “static” percentages. Percentages are by definition relative, and thereby need to be applied to a value in order to yield a dynamic result.

Do we apply 1% to the current bid/ask prices in order to determine the next incremental price between the spread?

If so, how do we handle orders placed outside the bid/ask spread?

Rounding makes applying 1% a lossy process, and thus not reliably reversible, so we can’t apply the same algorithm to reliably find the previous price points outside of the spread. The fact that orders can be removed at any time complicates this issue even further.

So then do we allow any listings at any price points outside of the spread? But this would in turn open up opportunity for exploits like placing an order inside the spread at the 1% increment, placing another order at 1c increment over the previous bid/ask price, and cancelling the existing order to yield a new ask/bid with only a 1c increment over the previous price.

There’s a whole other can of worms that I’d rather not open when you start considering how to transition prices from our free-for-all current system to the new dynamic 1% system, which is important to because I doubt the population will accept just having all their orders ripped off and everybody having to relist everything from scratch under the new system…

Even if all the implementation details can be sorted out, there is also the performance impacts of adding multiple floating point operations to every single trading post transaction, along with the validation logic required. Trading post performance is just barely good enough for most people as it is, but with such added complexity this might no longer be the case.

Hopefully I’ve made my case as to why my suggestion would be more realistic to implement.

Charismatic Harm.9683:

Kaon.7192:

The change suggested by William Bradley Knight would simply turn the Last List First Fill system we have right now into a Last Relist First Fill system, that still benefits TP bots/notification users at the expense of regular buyers and sellers. Again, it’s admittedly better than the status quo where even relisting is prohibitively costly and slows the convergence to equilibrium price, but ideally I’d like to see a solution that addresses the other fundamental problem as well.

This is true, but the point that you’re missing is:

• Once prices reach equilibrium, the benefits of undercutting no longer exist because Buyers and Sellers have reached a mutually beneficial price for an item. It’s only a Last Relist / First Fill state until the price reaches equilibrium.

Currently, it’s exceedingly beneficial for new sellers to undercut, especially on high valued items, because the new seller KNOWS it will cost the old seller an additional 5% to pull his listing and undercut him in return. This new seller gets to jump ahead in line for virtually the same price. The only risk this new seller has is being undercut by another new seller.

In either William Bradley Knight’s, or my model, undercutting is discouraged by pushing the price to equilibrium. His model does it in a manner that allows endless undercutting by everyone “for free”. My model associates an increased cost to do so.

The equilibrium point is the point where the Supply and Demand curves cross. It’s the point where the quantity of items buyers want equals the quantity of items sellers are willing to sell.

If there is a sudden increase in Demand for an item, the price will push up to accommodate for it. If there is a sudden increase in Supply of an item, the price will push down.

William’s model actually does this better than mine by making the Sell Listing side of the equation act more like the Buy Order side. Currently, there is no real penalty for outbidding someone with a higher Buy Order, outside of a little bit more money, but there Buyer is already willing to pay that, so it’s a moot point. Currently, on the Sell Listing side, there is the penalty of an additional 5% of the Sell Listing to lower your price. His model removes that penalty and is the reason his model works better to push the price to equilibrium faster than mine.

I see your point and agree that at equilibrium, the consideration I raised is moot. However, having unlimited chances to undercut is still no guarantee that prices will reach equilibrium at a reasonable pace, especially when there is no incentive to undercut more than 1c, because the only factor that affects Fill Time is Listing Time, or Relisting Time in the new system.

Price agility in the current system is unnecessarily slowed down by this lack of a minimal increment, which reduces the probability that an item is near equilibrium at a given time.

And when prices are not near equilibrium, the consideration I raised remains a real issue.

As such I feel minimal increments are still necessary even without a relisting fee.

Aidan Savage.2078:

This does injure the market. You dont make it difficult for players to make money using inefficiencies, you also make it difficult for them to make money just selling what they’re looting or crafting.

Moving prices to the equilibrium point does not injure the market. It injures your ability to make money by identifying inefficiencies in the market of items with prices that aren’t currently at equilibrium.

If you want to push towards a minimal spread of buy/sell prices, you need to increase demand and supply of the item. Artificially closing the gap via forced relisting prices is not “reaching equilibrium,” it’s attempting to induce a pattern of behavior that particular market clearly has no intention of doing. True equilibrium is the market not doing wild fluctuations in prices in a short order of time, and then persisting at those prices.

Supply and Demand of an item will adjust to equal levels once the price of the item in question reaches its equilibrium price. I will agree that artificially closing the gap isn’t the best way to reach that equilibrium point. Up until the point in this thread where I suggested my idea, the units used to move a price to its equilibrium point were completely arbitrary (1c vs. 1%). After that, William Bradley Knight introduced an idea that does a MUCH better job of pushing a price to equilibrium by letting the Buyers and Sellers do it. I endorse his proposal over mine as I believe it to be more efficient and much more elegant.

The perfect example of this is Radiant Dust during the LA ls week. It was extremely stable at 1-2s each, and the introduction of the recipe caused a massive spike in demand such that the price rose to over 20s each, and then induced a nearly instantaneous dump of supply that suppressed the price back down to 2-3s. All in about 10 minutes. That’s a shift in equilibrium due to the increased demand and the subsequent surplus of supply.

This demonstrates the idea of Supply and Demand perfectly. As Demand spiked, so did the price. Once Supply caught up and surpassed the Demand, the prices crashed. I’m not suggesting that the equilibrium price of any given item is a constant and stagnant number. I’m suggesting that the equilibrium price will shift up or down depending on Supply and Demand for that item. Radiant Dust is the perfect example of this.

Many items on the market are already in equilibrium because the prices that buy orders and sell orders are at do not fluctuate much. The fact that the state of equilibrium also includes a spread that is often 50% of the highest buy order is irrelevant. Not to mention that outside of external influences, in which case it’s new recipes anet adds, those prices will not shift because the supply and demand are consistently equal. The only way to bring those prices together by natural market movement is to increase the demand and supply of the item.

If an item is already at an equilibrium price, William’s idea to remove fees to re-list an item at a lower price would not impact these markets whatsoever. If Demand or Supply for these markets in equilibrium currently were to change….a new equilibrium price point would be found.

Flipping can make a lot of money if you’re able to move large quantity on a regular basis. If I’m making 2s every time I flip 1x of Item A, and item A sells on a consistent and regular basis, I can make 200g off Item A if I move enough of it quickly enough that someone else does not impede on my attempt. The low profit margins on some items make them unusable for large-scale flipping because if something goes wrong, you’re stuck at a no profit value or a loss at worst.
Safer money can be found on the tradepost by supplying the needs of crafters, whether those that are leveling to 400 or to 500, or those crafting ascended gear.

This basically proves my point. While a great deal of money could be made flipping, it depends on how much competition you have and as you’ve stated, there are much safer ways to make money on the Trading Post, such as supplying the needs of crafters. I’m sure Wanze utilizes even more ways to generate safer money, such as bags and salvaging (he’s stated as such before). Finding those safer methods of making money take a higher skill set though.

I still believe that moving item prices to equilibrium creates the most ideally healthy market. My idea of artificially doing so isn’t great and I don’t endorse it any longer, nor would I endorse an idea that places any arbitrary increment….especially since William Bradley Knight has proposed a much more elegant solution.