Jump to content

Unusual Suggestion "Auto-Pricer"


Keroro1454

Recommended Posts

Now, I'm not totally sure if this could work or if it would result in some wonky loops, but I'm curious what the reception to this kind of system would be.

Essentially, add a system that "automatically" prices unpriced unusuals by cross-referencing the effect averages with the actual taunts. The process is a tad more detailed:

 

Once one item is priced via manual submission (Or is simply a "updated" price), the "Auto-Pricer" can intervene for the rest of, for example, a new unusual. The system will then use appropriate multipliers against each item. For example, Hellish Inferno items are, on average, worth 1.194x the value of an Infernal Smoke variant. If one has the Infernal Smoke price, a guesstimate of the Hellish Inferno variant can be quickly generated, and so on and so forth.

 

Naturally, there are a few restrictions to prevent jumbled prices.

 

  • The initial item that the system draws upon must have been priced within the past 2 months, in other words the initial item cannot be one whose price is outdated.
  • Effect averages can only be drawn upon once a certain number of items with that effect have been priced (3-5 seems appropriate, otherwise new effects will take a while to be considered by the system)

 

Additionally, the system would introduce a new indicator to join the likes of Raised, Lowered, Refreshed, and Outdated. The indicator would be something simple, perhaps an orange gear. This indicator would show that the price for an item is a rough estimate, and thus not entirely accurate.

 

The system of an "Auto-Pricer" in this way would provide a number of benefits:

 

  • Lessen pressure on price suggestors to quickly price newer items
  • Significantly decrease the likelihood of sharking in a large number of instances*
  • Reduce confusion that often surrounds "unpriced" unusuals and assist in removing a degree of the stigma surrounding said items

*The issue of sharking is especially pertinent in this suggestion. Certainly, this system by no means can assist in preventing people from being pressured, conned, lied to, etc etc. It can, however, provide a significant deterrent to sharks. One issue in catching sharks is the problem of price, specifically, determining the value of the items sharked. Even assuming we operate under Outpost's commendable guidelines of 10%, sharks can avoid such a guideline by grabbing unpriced items, which technically have no value and excluding exceptional circumstances (Sunbeams Woodsman comes to mind) will not be "worth the time" to prosecute. This is especially evident with newer items- when updates drop and new items are added, newer players have a much higher likelihood of unboxing these new crates and cases. The items from these sources are, of course, unpriced.

 

I think this system would be a serious improvement to backpack.tf. I'd love to hear comments, criticisms, etc.

Link to comment
Share on other sites

Doesn't seem to be a particular point to this. People already take unusual prices too seriously, and arbitrarily setting prices is just not in the spirit of the site in my opinion.

Lessen pressure on price suggestors to quickly price newer items

New items have no baseline so there is no way to "auto price" them. Additionally they fluctuate so fast that even normal suggestions are usually inaccurate.

 

Significantly decrease the likelihood of sharking in a large number of instances*

Sharking happens primarily because people don't research prices. Autoprice won't change this.

 

Reduce confusion that often surrounds "unpriced" unusuals and assist in removing a degree of the stigma surrounding said items

Setting prices semi-arbitrarily adds to confusion, it doesn't remove it.
Link to comment
Share on other sites

  • Administrators

I think that this would provide a considerable drawback from as far as I understand the idea.

- If you're going to draw effect averages from updated hats, you're going to have averages that depend on which hats with those effects were last updated. Youre going to get very low results if someone updated all lowtier robos on them vs super high averages on those where the KE, TC, antlers etc were the last to be updated.

Aside from that, it will likely misrepresent a lot of values for more than just that reason. For instance, look at the huge discrepancy between similar valued effects CV, PS and SBTWC, just as an example, or the more extreme discrepancy between G.codex and A.codex. For a lot of effects, it will be very hard to assign a "similar tier effect". What are you going to compare effects like circling hearts or bonzo to?

Additionally, it will likely not work (properly) for new effects. The tier list on new effects will be hard to determine; your example of HI compared to IS is probably a good indication this will not work properly, as from my experience, HI is far more difficult to sell than smoke, and is therefore likely less valuable on average as opposed to more, realisticly speaking. Again, this is probably due to the types of taunts that are priced; if one obscenely overpriced item gets priced as a first, or if one person manages to sell an item at a very high price (or if only lowtier hats move around really fast), everything based on those will give poorly skewed values.

Personally, I would prefer to stick to using sales, especially on rare or new items. On very common (mostly firstgen) unusuals it may not be that bad of an idea, but for those, I feel like the influence it would have on the current list would be minimal, as most are already priced (or unavailable on the market). While the incentive is good, I feel like the result here will not be productive enough to be worth the effort (if not counterproductive, which I would personally fear).

Link to comment
Share on other sites

I think that this would provide a considerable drawback from as far as I understand the idea.

 

- If you're going to draw effect averages from updated hats, you're going to have averages that depend on which hats with those effects were last updated. Youre going to get very low results if someone updated all lowtier robos on them vs super high averages on those where the KE, TC, antlers etc were the last to be updated.

 

Aside from that, it will likely misrepresent a lot of values for more than just that reason. For instance, look at the huge discrepancy between similar valued effects CV, PS and SBTWC, just as an example, or the more extreme discrepancy between G.codex and A.codex. For a lot of effects, it will be very hard to assign a "similar tier effect". What are you going to compare effects like circling hearts or bonzo to?

 

 

Ideally the averages would sort themselves out, but I see what you mean. The idea stemmed largely from my viewing of unusual taunts, which are a more isolated market. Looking at this in terms of the much more expansive cosmetic unusual market, I think you're right in that it would struggle to properly compare effect averages in such a way that multipliers could be accurately determined. Frankly, the issue more than anything else is effective "human error"- logically there's no reason for PS to rank so much higher above SBTWC, but human preference dictates a wide difference. I suppose in an "ideal" world where preference, theme etc doesn't occur the idea could occur, but given that's not the world we live in, I should have taken that into account.  :P

 

 

Doesn't seem to be a particular point to this. People already take unusual prices too seriously, and arbitrarily setting prices is just not in the spirit of the site in my opinion.

New items have no baseline so there is no way to "auto price" them. Additionally they fluctuate so fast that even normal suggestions are usually inaccurate.

 

Sharking happens primarily because people don't research prices. Autoprice won't change this.

 

Setting prices semi-arbitrarily adds to confusion, it doesn't remove it.

 

You keep using the word arbitrary, but I think that calculating these is inherently not "arbitrary", whereas people dictating without basis prices for unpriced items would be far more arbitrary. Unless you're referring to the flaw in effect multiplier calculations, in which case as I said above I wholeheartedly agree there is a serious, effectively un-fixable in the suggestion's current state, flaw.

 

On the topic of sharks, the preventive measure that this implements is a deterrent. It's not meant to actively educate people on prices in these instances (Although it certainly could in a small handful of cases), since sharking largely occurs to people unaware of the trading community and thus backpack.tf. The preventive measure is that this system would ideally allow for easier prosecution against people sharking "unpriced" items and thus would serve as a deterrent should it indeed prove to allow for faster and more effective prosecution on these matters.

 

 

If any mod would like to lock this I believe that would be agreeable, the suggestion has been examined thoroughly enough and the flaws present have been made clear. Thanks.

Link to comment
Share on other sites

Going to have to agree with everyone else here, it's impossible to really formulate a ratio for one effect to another, and the market wouldn't be able to set a price on its own. Also, even though you said there would be a 'rough estimate' price on every item, people would still get confused as they do with items priced purely by SCM. I think you know how many threads there are with people who bought a killstreak kunai or something for like 100 keys only to find out its actual price later on  :P . I don't really see any issue with the current system, so I doubt this is necessary

Link to comment
Share on other sites

Why are unusuals unpriced?

 

Reason 1: Item is a newly released item and prices are too volatile. Prices on these kinds of items can change even in a matter of 1 week. As Foamy suggests, this means auto-pricing off other effects of the same hat is impossible because it would be dependent on when the suggestion was last accepted. The other main issue with this is that pricing is entirely dependent on how many of an item are on the market. All the new taunt effects are roughly equal in value. The 10-15% variation you are seeing is not an indication of actual value difference (imo) but rather a reflection of how many of each effect is actually on the market at a given time. This would also be really difficult to factor in with autopricing. 

 

Reason 2: Item is very rare. These just cannot be priced with any kind of autopricing system because extremely rare items can fetch "outlier" type values from collectors. Impossible to predict on these. A 1 of 1 arcana hat can fetch values 2-3x over the burning effect on the same hat. A 1 of 5 arcana hat might be in the same range as energies. 

 

There are definitely a lot of situations were prices are "predictable" where your strategy might be applicable, but in all these situations suggestions have probably been made already.

  • Like 1
Link to comment
Share on other sites

I agree with polar for the most part.

 

This idea valid for certain scenarios. However it is very hard to guess the price of a newly released effect or a hat if there is no other information, and even if there is, it is probably wrong.

 

It could, however, help with items that are not rare, but vastly outdated. We have been using machine learning to predict unusual prices for over 3 years now. See  http://forums.backpack.tf/index.php?/topic/51055-overpricedunderpriced-unusual-price-analysis-hats-taunts/ for example for an analysis on existing hats (that chart has not been updated in ages). And before that there was this: http://forums.backpack.tf/index.php?/topic/17120-every-single-unpricedoutdated-unusual-updated-2015-12-30/

 

There are still a ton of hats that fall into this category, and I think there are plenty of things that can be done about them. 

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...