Rules for Non-Unusual Suggestions

[polar]

This thread serves as a list of up to date rules on suggesting prices for non-unusual items. When a rule is changed, the main post (this one) will be updated, and a changelog post will be added below. If you have questions, please join our discord to ask them.

 

This is designed to be a more comprehensive version of the general and suggestion rules - which basically introduce users to price suggestions. You can post any questions you have, including asking what a mod's rationale was for accepting / rejecting certain suggestions in this forum. Additionally, it is within your right to dispute both a suggestion or a rule, but be ready to provide sufficient reasoning.

 

Introduction

 

These are the basic rules that everyone should be accustomed to before suggesting.

• All of our prices are estimates. They are not definite and act only as a representation of the market at the time of the suggestion. 
• You must provide proof for all suggestions, not just opinion. See "Valid Proof" below.
• We are after the common trading points for items. People will always sell for above and below backpack.tf prices, but this doesn't mean we will make all-encompassing ranges to document every single sale value. We are only showing the most common trade values. The number of sales always trumps the time it takes for an item to sell.
• Prices we list are for unmodified items. For low value items (<5 keys), please do not use painted, parted items as proof. For any item, please do not use items with killstreaks as proof.
• Refreshing the price of non-unusuals is unnecessary. If the current price is 1.66 refined, please do not make a new suggestion at 1.66 refined.
• Rounding rules and guidelines:
  • Rule - If the item is over 1 Key, round to the nearest .05 Keys
  • Rule - If the item is over 10 Keys, round to the nearest 0.1 Keys
  • Rule - If the item is over 30 Keys, round to the nearest 1 Key
  • Guideline - If the item is over 100 Keys, please round to the nearest 5 Keys
  • Guideline - If the item is over 200 Keys, please round to the nearest 10 Keys
  • Guideline - If the item is over 1000 Keys, please round to the nearest 50 Keys
• Finally, please use common sense. If you have one sale at 9 refined but there are an abundance of sellers below that, consider whether the sole 9 refined sale is significant enough to be in the range. Use as many different sources as you can to make decisions rooted in common sense.

 

Valid Proof for Non-Currency Suggestions

 

Primary Proof

• Completed trades on a variety of sites, such as backpack.tf classifieds, bot.tf, and tf2outpost. More detail on how to find sales can be found here.
• Unsold listings that have been adequately bumped and open for a reasonable amount of time. Unsolds can be used to lower item values but not to raise them. Unsolds should not be used as sole primary proof if proof of sales exists.
• Buy orders may be used as primary proof, so long as they are proven to be successful.
• Marketplace.tf sales can be used as primary proof provided they are not significantly discrepant from the item market. If item sales are available, they should take precedent.

 

Secondary Proof

• Any unsolds on any site that have been bumped at least once. You should put as much proof as possible in your suggestion. However, given a lot of these sites see much less traffic than outpost or bp.tf, these should not be used as primary proof to define your range. 
• Steam community market. Moderators always take a look at SCM values of all marketable items and consider them in any suggestion. If an item is selling regularly at 1 key on the market and you have a single trade at 2 keys, more than likely your trade will end up an outlier if we wait for further sales. This doesn't mean that you should use SCM as the only proof to define your range, but particularly on rare items with few sales, mods always factor in SCM values to help identify outliers and common trade points. For unpriced items, we will consider SCM sales alone as proof only on items >10 keys in value unless you can demonstrate a good reason 
• Buy orders that are unsuccessful may be used as supportive proof that certain values may be on the low side, but cannot be used as sole proof to get an item updated. 

 

Valid Proof for Currency Suggestions

We recommend that you only try currency suggestions after you have gained some experience with suggesting, as they are significantly more challenging. Remember, we are after common trade points. To determine the common trade points for an item like the key, you will need to show a more substantial amount of proof. Have a look at prior accepted suggestions for an idea about what is reasonable. For key suggestions, please refer to our cooldown rules. For refined suggestions, you can find an in-depth description of how to approach them here.

 

Making Suggestions That Are Meaningful and Relevant

 

We ask suggesters to avoid making suggestions changing the values on items by less than 5% unless there is a compelling reason to do so. Sometimes the compelling reason is to significantly alter the range while keeping the mean the same. Suggestions changing the mean value by less than 5% are left to the moderator's discretion whether they should be accepted or not. We want to encourage meaningful suggestions that are helpful to the community, not suggestions that promote clutter. When considering whether to suggest a small change, try to use some common sense and determine whether or not the change really represents a useful or meaningful one to people casually using the site. Is 15 refined vs 15.11 refined really a price change that matters to the vast majority of users? In practice, the answer is no - they are basically the same, and thus the suggestion is not really needed.

 

Automated SCM pricing

 

By default, an item will use an average of the Steam Market sales when it lacks an accepted price. These rules are in place because when items are sold en masse on the SCM, the automated pricing is simply significantly more able to keep the displayed price up to date than we are by means of suggestions. There are some situations in which manual pricing may be beneficial, such as if an item is rarely sold thus the average of the previous X sales is not updated quickly to current value. As such, you can make suggestions on items within this category but only if you can demonstrate a solid rationale for doing so.

 

Using Suggestions as a Soapbox

 

Please avoid extraneous commentary in suggestions. Please post only the relevant proof and a brief conclusion to justify your range. Extraneous commentary encourages spam, and on key suggestions this makes it hard to distinguish posts that are actually relevant to the suggestion. For additional commentary, please use the forums.

 

It is also worth reading the suggestion etiquette guide if you care about your own and the sites reputation.

[polar]

Rules for Decorated Weapon Suggestions

 

(1) We will only accept suggestions for skins greater than 10 keys or ~25 USD on SCM in value. Quality of the item will not matter.

 

Rationale: A lot more sales happen on SCM than anywhere else for the really low tier skins. And the SCM price will automatically be updated and updated much more quickly than manual updates for suggestions. Let's leave these <10 key items alone. 

 

(2) SCM sales can only be used for unusual skins, and even there, preference will be given to item sales (SCM sales will be considered the outlier if an item sale value and an SCM sale value are discrepant).

 

 

EXCEPTIONS: Items already priced with accepted suggestions can be updated with new suggestions.

[Foamy the Fearsome]

On a suggestion earlier I saw some confusion regarding "rare items". I assigned it to myself so I wouldnt be typing everything for nothing, but for some reason the suggester was able to close it despite that.

 

http://backpack.tf/suggestion/57de6ec9c4404560424e2471

 

leaving it here as a reference.

 

 

Leaving common items aside, there are rare items in different categories. You can't treat all rare items according to 1 set of rules; look at them as individual exceptions to the rules set for regular items, depending on the situation. For most 'rare' items, that means that single buyers/sellers have more power over the market, as the market is limited and thus the amount of potential evidence is limited.

 

For some items however, this will backfire. Thats where it starts to matter how rare something is.

 

 

A few cases below. Listing them here because the comment section people seems to mix these up:

 

 

1) A rare item unlikely to ever sell, that shares its stat page with a much more common item:

 

http://backpack.tf/suggestion/5761a99086674730a54b9596

 

The main reason this did not pass was because there was an invisible difference between prewrapped (and actually tradable ones) and normal untradable ones. The only ones that could ever possibly sell were the wrapped ones, which are far more rare than the non-wrapped ones. Aside from that, it makes no sense to raise such an item with a buyer, as the chances they'll ever be exchanged are super-limited. 

 

 

2) A situation where an item as mentioned above actually had an exchange:

 

http://backpack.tf/suggestion/5753669ec440457a6c3b2fb8

 

Despite being unlikely to sell, this item sold, which normally means you have a go for your suggestion. However, there was another problem at hand, a criterium that still isnt satisfied: This rare 'glitched' version's sale would influence all non-glitched ones as well, meaning that a sale on a super-rare item would influence a super common item. That's why it was closed

 

 

3) A situation where a rare item thats normally unlikely to exchange can actually be updated:

 

http://backpack.tf/suggestion/57ac4b1c83de6b631572c15f

 

This suggestion actually had a sale. Since theres a difference between craftable and non-craftable that is properly separated on the individual item's stat pages, this wouldnt influence regular hetmans. Theres only a dozen or two of these in existence, and since the only added value they could possibly have is a collector's value, theyre unlikely to exchange. Nearly all exchanges youll see are collectors buying one.

 

 

4) This suggestion, which should fall under a separate category of the situation above:

 

http://backpack.tf/suggestion/57de6ec9c4404560424e2471

 

There are only two of these in existence. A generic buy order on one of these is as useful as making a buy order on something like the sun, or - more realisticly - on an artwork. In the case of the artwork, it makes no sense for me to make ads in newspapers or billboards stating "I want to buy 'the scream' for 1000 money units. Must be the original, not accepting duplicates". If I want this piece of art, I can go to the owner of the only one of them in existence and offer him 1000 money units. The owner will either accept or decline. If he accepts, then 1000 money units is a reasonable price. If he declines, then 1000 money units is not an accurate price.

 

 

This analogy may sound stupid and far fetched, but think of it this way: how is it different from this scenario?

 

"I want to buy 'the dapper disguise' for 1 money unit. Must be tradable, not accepting untradable copies". A generic buy order on the classifieds makes no sense similar to the billboard/newspaper ads in the other situation. If I want this item, I can go to the owner of the dapper disguise (or in this case, one of the two owners, which still really isnt that much effort) and offer him 1 money unit. There are two relevant scenarios: one in which one of the two agrees and takes the money unit, and one where none of them agrees. In the first case, 1 money unit is a fair price for this. In the second case, 1 money unit is not an accurate price.

 

 

The problem with this suggestion [still referring here: http://backpack.tf/suggestion/57de6ec9c4404560424e2471] is that the concept of a 'buy order' makes no sense here. Buy orders are meant to make owners of rather common items come to you instead of the other way round, so that you dont have to ask all 10000 owners of an item if you can buy them. The number is simply artificial - whether its too high or too low isnt the issue here - and holds no actual value. There is nothing here indicating that 1 key is a good price for this item, and the situation here doesnt allow this to be seen as 'better than current'

 

[Teeny Tiny Cat]

Classified snapshots in suggestions:

 

Now that the site automatically provides a snapshot of the classifieds at the time the suggestion was made, it is no longer required to include screenshots or histories in your suggestion body. It is still advisable to check other areas of the market and include proof from outpost, steam community market, etc, if it exists. It is also recommended that you include a rationale for your suggested price in your suggestion. Showing that you put thought into your work will only help you with the votes, and it is more likely that a mod will accept your suggestion, especially in close calls.

 

Please note that snapshots will also allow you to more easily track sales in previous suggestions. Have a look at snapshots in previous suggestions, especially those made in the last 2 weeks, to help you find sales.

[Teeny Tiny Cat]

Sales in non-unusual suggestions

 

Recently we have been trying to emphasise sales more in non-unusual suggestions. From now on, we will be doing this much more religiously. You can find sales easily on common items in a number of ways:

 

- screenshot classifieds and check back in a day or two

- check classifieds in cache

- bookmark outpost trades

- check recent suggestion snapshots

- check histories of current sellers, and refresh their backpacks - they may have sold some items but the site hasn't updated yet

- check compare links for bots - if they're selling one, they may have sold more

- check compare links for buying bots - if you're dropping a price based on an unsold, but bots are managing to buy regularly, that's a more common trading point 

- marketplace.tf (unless the item market is different)

- scm (unless the item market is different)

- bot.tf stats available here if you have donated 1 key or more

 

From now on, if you make a suggestion using just sellers/buyers and mods consider that you could have easily found sales, your suggestion may be closed without giving you time to close it yourself. This is to encourage people to check markets properly and make sure they have the most evidence possible before suggesting. If you absolutely cannot find anything, you can still suggest with sellers/buyers if the change is meaningful. I am not putting a #% on meaningful. It's context dependant and I'd like suggesters to start using some common sense and making arguments for their chosen ranges.

 

I keep seeing "buyers counter" on non-unu suggestions when buyers/sellers are 1 scrap below suggested price. To clarify this; buyers do not counter if they are successfully buying. Those are sales.

[Woifi The Viking]

On 6/29/2016 at 11:15 PM, polar said:

Rules for Decorated Weapon Suggestions

 

(1) We will only accept suggestions for skins greater than 10 keys or ~25 USD on SCM in value. Quality of the item will not matter.

 

Rationale: A lot more sales happen on SCM than anywhere else for the really low tier skins. And the SCM price will automatically be updated and updated much more quickly than manual updates for suggestions. Let's leave these <10 key items alone. Sale values and lowest seller values are usually close enough on SCM for these low tier items.

 

(2) SCM sales can only be used for unusual skins, and even there, preference will be given to item sales (SCM sales will be considered the outlier if an item sale value and an SCM sale value are discrepant).

 

 

EXCEPTIONS: Items already priced with accepted suggestions can be updated with new suggestions.

A quick reminder that those rules also count for the new war paints and their skins.

[Teeny Tiny Cat]

For the sake of clarity, here is the definition of a random craft hat. This is in place to price the lowest value craft hats that people use as currency for crafting, not those people buy as cosmetics.

[Foamy the Fearsome]

Over the past few months we have seen a drastic decrease in effort put in regular cosmetic suggestions - an explosive increase in quantity at the expense of quality as it seems. The moderators that handle these suggestions most are displeased, but accepted them if they were correct simply to keep up with the supply of suggestions. Due to an increasing number of complaints about this, I want to combat this lack of effort in an attempt to restore the quality standards to the bare minimum we've set a few years back.

An example suggestion of things we do not want to see, but that seem to be increasingly more common lately:
https://backpack.tf/suggestion/614ec415ba96c15da1584cbb

Having sites like Bot.tf, and now Gladiator.tf already eliminates all the time and effort that used to render common cosmetic item suggestions so tedious in the past. All evidence gathering can be done through acquiring and sharing one link; it only takes a few extra seconds to accompany that link with a CLEAR indication of which sales you use, as well as a summary of the sales written in the suggestion (along with a rationale for outliers, if applicable), so that people that wish to check the suggestion need not investigate everything themselves to figure out where the range is coming from. 
I've highlighted 'clear' here to stress that a 7 in front of a link (which blends in with the link at first glance, hurting its readability considerably) is not an appropriate substitute for "taking sales from the past 7 days".

This is why we require you to have all the information regarding the range visible in writing on your suggestion in addition to the link to the Gladiator.tf stats page that you receive your info from so that anyone that reads your suggestion can see where the range is coming from. This includes how many sales happened at which values, and an indication of which time frame you have taken your sales from (a proper one, so as mentioned before, not a random 7 somewhere that new or inexperienced users will not understand). 

The post below contains a template/explanation of what we generally expect of suggestions to ensure they are clear for everyone. Of course, the current stats page will come from Gladiator.tf and not from Bot.tf. This post is also pinned in #Suggestion-discussion. Note that snapshots from Gladiator.tf are used now as opposed to screenshots.

 

[ᴾᴾᴹ Shuffle Spy]

Recently, we've been deliberating over how non-Unusual suggestions should be handled due to feedback we've received, primarily regarding the large volume of these suggestions that gets made, how small the price changes usually are, and how long these suggestions remain up / how frequently they have to be remade due to changing markets. The moderators that handle most of these suggestions agree that it's best something is changed, in part due to simply not being able to keep up with the sheer volume. One of the things we often see is items that keep bouncing back and forth between two values very close to each other, only just above the minimum change requirements, which serves little purpose but costs a lot of effort from both sides. This is something we want to change.

 

It's a difficult problem to tackle since many solutions to individual problems will result in other issues: forcing faster accepts will make many markets susceptible to bot/price manipulation, implementing simple cooldowns will cause issues with fast-shifting items, increasing price change thresholds will have limited effects on some items whereas it will make other items unable to be updated, and making those thresholds flexible to account for that will make the suggestions unnecessarily complex, just to name a few.

Presently, we see two things that seem to cause the biggest problems. Firstly, many suggestions in which the data used is exclusively from the past 24 hours (or even past 7 days on less common items) show significant deviation between the value of that day (or week) and the values from the days/weeks before that. These often leads to changes that rarely induce a long-lived market shift. As such, we feel broadening the timeframe of acceptable data will put this in perspective and will help avoid many of these ultimately insignificant suggestions. Secondly, the reliance on present-day buyers and sellers is something that has been an issue for years in all suggestions, as buyers/sellers tend to fluctuate a lot, especially now that most buyers/sellers are bots that overcut and undercut each other multiple times a day. The moment a suggestion is made is simply one snapshot in time that currently holds far too much power in determining outliers.

 

So, what will change? We want to start by trialing these changes with non-Unusual items valued under 1 Key to gain community feedback. This does not apply to weapon skins. In addition to lengthening the timeframe in which we consider data to be usable to a month, we wish to broaden ranges to encapsulate more relevant data and be less dependent on the current state of bots. Some examples of how suggestion ranges will be handled differently are:

• https://backpack.tf/suggestion/646279cd2919e2c97001b84d - This was accepted at 1.66 - 1.77 but only used sales within the prior week. Had the prior month been used, a high-end of 2 would've been supported with a higher volume of sales. This price would've been valid for nearly 5 months, and the suggestion made 3 months later to raise each end of the range by 1 scrap would've been unnecessary.
• https://backpack.tf/suggestion/64f3872e34d2205a4e097479 - This was accepted at 2.22 - 2.33 but excluded 14/24 sales from the prior week. There were only a handful of sellers below 2.55 shown in the snapshot, so if some of the higher sales were used to make a range like 2.22 - 2.55 or 2.22 - 2.66, the next two suggestions would've been unnecessary because sales at both end were still frequently happening.
  • https://backpack.tf/suggestion/650655daa8526d82520f600b - This was made 2 weeks later
  • https://backpack.tf/suggestion/6597a8b7e143adb890014417
• https://backpack.tf/suggestion/657ab4d876c68178ac00a976 - This was accepted at 2.22 - 2.33, but there were many sales at 2.22 and 2.66 within the prior month. The suggestion accepted a month prior at 2.22 - 2.66 - https://backpack.tf/suggestion/6567fbbd0ac9afee63034877 - would've still been valid at this point.
• https://backpack.tf/suggestion/663d8039855b29bc7d09a9f1 - This was accepted at 13.22 - 13.33, but a high-end of 13.88 would've been acceptable with the volume of sales to support it.
• https://backpack.tf/suggestion/65431a9e451f53dc690685b1 - This was accepted at 3.11 - 3.22, but a range of 2.77 - 3.22 would've been acceptable looking through the sales of the prior month. Several suggestions accepted afterward would've been unnecessary.

 

With these changes, we will get values that should remain relevant for longer. As our current rules state, use judgment in determining an optimal range. Just because there's a wide array of sale values doesn't mean they should all be used. Our stance regarding price refreshers still stands - suggestions changing the current mean value of an item by <5% should generally be avoided unless a strong argument can be made for why the new range is necessary. As this trial period begins, a refresher widening a range may be acceptable to serve as a starting point, but these cases will fall to our discretion.

[ᴾᴾᴹ Shuffle Spy]

On 5/14/2015 at 3:16 PM, polar said:

• Please round to the nearest .05 keys if the item is over a key. For items over 10 keys, please round to the nearest .1 key. For example, a sale at 5.86 keys should be rounded to 5.85 keys while a sale at 12.24 should be round to 12.2. For items over 30 keys, please round to the nearest 1 key. For example, a sale at 69.6 keys should be rounded to 70 keys while a sale at 44.4 should be rounded to 44.

For expensive non-Unusual items, we're amending our rounding rules to align with the rules in place for Unusuals at higher values. For items over 100 Keys, please round to the nearest 5 Keys; for items over 200 Keys, please round to the nearest 10 Keys; and for items over 1000 Keys, please round to the nearest 50 Keys.