Let's discuss the 'minor changes' rule

[regi]

Hi. On 23 July 2015 I made what would be my last suggestion for three years. By that time, I had over 330 accepted suggestions, all of which required scrupulous scrounging through TF2Outpost and similar trading sites to find the proof required to modify the price. Because of the inherent difficulty in changing prices of many items, price modifications of all sorts were allowed. For instance, a change from .22 to .22-.33 , .44 to .33, a price difference of only one wep, and other various one scrap increases (one, two, three). Less than one month later, however, on 7 August 2015, polar introduced a new rule: a change in the mean of at least 5% is required for all non-unusual suggestions. When I returned a few days ago and made a few suggestions, I was surprised to see the opening and closing of ranges rejected, for at the end of the day, we're here to report on the market, are we not? Removing 11.22 from a 11.22-11.33 has no inherent downsides, especially when there's no reason for the low end to be there due to buyers. Adding 30.11 to an existing 30 price to make 30-30.11 is all the more inclusive, and thus more reliable, so what's to gain from rejecting it? In both of these cases, the suggestion had over a 70% approval rating, indicating--at least to me--that the community was not only fine with them, but supportive. Additionally, per polar, "a mod may choose to accept based on the votes". Additionally, I should note that exceptions to this rule apply, for instance, I was able to update a price by a weapon's worth, but not before providing my ramblings as to why it's necessary. With the introduction of bot.tf and the ease in updating prices, it seems like it's not only easier to push aside the amount of research it once took to make hundreds of accepted suggestions, but also harder to make meaningful changes when that site doesn't have enough data. I assume that both of my examples would have been accepted had bot.tf had 20 or 30 odd data points supporting the removal or addition of ranges, but TF2Outpost is barren and trade.tf hasn't listed closed trades for years so there's no way to back it up, besides backpack.tf classifieds.

 

Obviously I'm not going to walk into the forums parading around like I know everything after three years of absence, but this rule that was imposed during it is sincerely making the site take a step back in terms of reliability and accuracy. The very nature of backpack.tf is rooted into the idea of giving the thousands of TF2 traders insight into the market. I doubt anyone reading this hasn't suggested before, so let's take a moment to explore the site through the eyes of those uninformed into the inner workings. If someone were to come here and see an item marked as 30 ref, they'd probably assume the following: this item can be bought and sold at 30 ref and anything that deviates from this is an outlier. Because it's a flat price, we're saying that we're certain that this is what this sells at; this is what the market for this item looks like. On the other hand, if we set the price as 30-30.11, someone who stumbles onto the page won't think that the price is 30.05; rather, that this item sees success at both 30 and 30.11, and anything outside of that is an outlier. I'm not saying we should ignore items in desperate need of large updates because we've got small changes to make elsewhere, but rather we should treat both with the same amount of importance, because at the end of the day, we're here to give an accurate portrayal of what's going on in the trading world.

 

I respect the work that price mods here put in. They're obviously passionate about the site and want it to succeed, as do I. I discussed this with a few people before I posted this, and the argument that it is a waste of time updating items so slightly came up multiple times. I consider this a non-argument, or a non-issue. We shouldn't disregard accuracy under the guise of time management. Sometimes it's better to think of the bigger picture: we're not here to satisfy each other; we're here to provide a service to the thousands of TF2 players who come here for an accurate representation of the market of a particular item or items. If we can't provide this, even if the change is deemed minor by some, we might as well pack up shop. You might think this is blunt, but tip-toeing around it isn't going to help anyone, and if no one raises any signs of discontent, everyone will continue to think that nothing is wrong. I don't expect everyone--or even a majority--to agree with me on this. I just want to understand why this rule is here, why people continue to either enforce it strictly or loosely, and how the price moderators think that being inaccurate by a few scrap is accurate enough. I doubt people would support me moderating prices (and I'm not too interested, anyway), so I come to you as a long-time suggester inquiring as to why the site has an unnecessarily bureaucratic rule for this; if I'm able to get it removed, at least partially (items can have ranges added or removed by scrap's worth, but not refreshed (3.44-3.66 to 3.55)), then that's a bonus.

 

Thanks for hearing me out.

[Woifi The Viking]

Given the examples you provided I think you are missing an important point here, 1 weapon price changes are fine for really low tier items, lets say around 1 ref for example, but not for 30 ref items. A 1 weapon price change for a 1 ref item is around 5%. A 1 weapon change for a 30 ref item is 0.02%. 0.02% is so insignificant small, that's totally neglectable for any kind of data point.

Also if people think everything besides 30 ref is an outlier they do not understand the most important basic rule of backpack.tf, all prices are estimates. As I said you should probably view all prices with a 5% deviation, the suggested price being the mean.

[JithraRufure]

32 minutes ago, regi said:

I discussed this with a few people before I posted this, and the argument that it is a waste of time updating items so slightly came up multiple times.

 

Actually the main argument against this in Discord was that its far less important to update something slightly rather than to update something by a far larger amount, meaning that the time could be better used elsewhere. Already, a ton of suggestions which change the pricing of an item by a high amount take a long time before they get accepted. Its not that its a waste of time to update those items, its that the time can be better used elsewhere, its called maximizing the resources you have available to you for maximum welfare. Maybe if some system could be used to reduce time required on suggestions, such as what someone said which would be to lift bot.tf item stats for some items and just auto create suggestions, then it would be feasible or viable. 

[AD⚡HD]

Being able to price something 30-30.11 sounds like the worst thing you could do, on items above 10 ref you should have to round to the nearest reclaimed metal

[AdamWTS]

37 minutes ago, AD⚡HD said:

Being able to price something 30-30.11 sounds like the worst thing you could do, on items above 10 ref you should have to round to the nearest reclaimed metal

If there's enough proof to establish a price in increments of scrap metal, why not do it? The whole point of price suggesting is to provide prices that are accurate as possible. This is also why several semi-rare stranges and hats are priced in increments of a 0.05 of a key.

[Mengh.]

58 minutes ago, AdamWTS said:

If there's enough proof to establish a price in increments of scrap metal, why not do it? The whole point of price suggesting is to provide prices that are accurate as possible. This is also why several semi-rare stranges and hats are priced in increments of a 0.05 of a key.

I think the reasons why any suggestion is made is either to better the accuracy of the current Backpack.tf pricing of the item in being able to provide a price at which the item would be valued at, or to better its precision as to suggest a change with the spread of the value of the item's sales.

 

Making an accurate price suggestion would be something like taking a unique all-class hat from 17 refined to exactly 5 refined due to an occurrence like the hype around the Keyless Multi-Class crate, where the sale price is off some ways what it use to be. As for a price suggestion to make the documented value of an item be more precise, you'd essentially be tweaking the price just a little, though the price may be accurate in some occurrences, as there are sales that do show a slight fall or rise, maybe including a range that includes the currently accepted price as with what some of the newer non-unusual price suggestions have been doing. With some leniency, the price of the item you're tweaking is still correct with sales that occur in that price region, but there is a more common selling point that could be included that is close to the currently accepted price.

 

It's understandable to want to make a price as precise as possible e.g 30 refined to 30 - 30.11 refined, however, as Woifi said, the prices documented by Backpack.tf is merely an estimate. Using  Backpack.tf Classifieds and Bot.tf is currently the main ways to price non-unusual items, however, there is a hindrance in that those two places aren't the only two places to document sales for items, and it could be that sales at 30 refined are more commonplace, just not observed. Being too precise, when you don't have access to every single TF2 trade upon Steam to see every sale of an item, does mean that you jump the gun a little in determining a price that could be more wrong than what is currently accepted, so I believe that there's the leniency of 5% as to say that these prices are rough estimates at this moment of time for the time being. The sales price of keys on 1 day could be 30 refined, another with the majority of keys being sold at 29.77 refined (e.g a seller flooding the market to get refined), the next could be 30.11 refined, simply down to just the randomness of people listing items. 30 refined is still accurate as the point of market equilibrium, but I don't think there is a need to be incredibly precise with community prices as it is simply not possible with the uncertainty of trading, unless there is a substantial difference as given by the 5% rule.

[regi]

18 hours ago, Woifi The Viking said:

A 1 weapon change for a 30 ref item is 0.02%. 0.02% is so insignificant small, that's totally neglectable for any kind of data point.

 

 But people don't use backpack.tf as a data repository, they use it to get an idea of the price. If we're able to be more accurate, not in terms of percentage, but in relation to the big box that displays the price, I fail to see any downsides. I'm not saying people treat this site like the trading bible, but that's no excuse for poor estimation. An approximate price of 30.05 allows for a sense of uncertainty. 30 ref flat means that the suggester and accepting moderator, along with all voters, agree that 30 is the price to trade this at. It has been determined that the sales at the other prices have been outweighed by the volume of sales and purchases at the accepted price. The stock market and share prices are exact. If I were trading shares I certainly wouldn't want to be told that "eh, $3 per share is close enough. You might be able to buy for less and sell for more, but we deem these few cents to be insignificant". The price difference may be indiscernible to those who have hundreds of accepted suggestions, but people who are here to sell an item they just unboxed don't care about the site's internal workings. They've googled "tf2 trading" and stumbled here. They've typed in the hat's name and they've landed on the page. They see no warning that these are estimates or "suggestions". The price is 4 ref, but people with bot.tf access and similar know that it sells at 4.11 too. They don't. Bam, you've basically scammed someone out of a scrap's worth because we were too high and mighty to bother to update it.

 

And before anyone says "it's only a scrap", think about those who can't or don't want to spend money on pixels. Think about those who see an item and go "yeah, I like that. Perhaps I'll try to trade for it". Just because we are able to decide that .11 is too minor of a change because we're sitting on backpacks worth hundreds (I'm not, but the majority of price mods are) doesn't mean we can tell others not to bother. Players are dropped ten items per week. That's .55. .11 is a lot more than .02% for some.

 

18 hours ago, shaunleewenjie said:

Actually the main argument against this in Discord was that its far less important to update something slightly rather than to update something by a far larger amount, meaning that the time could be better used elsewhere. Already, a ton of suggestions which change the pricing of an item by a high amount take a long time before they get accepted. Its not that its a waste of time to update those items, its that the time can be better used elsewhere, its called maximizing the resources you have available to you for maximum welfare. Maybe if some system could be used to reduce time required on suggestions, such as what someone said which would be to lift bot.tf item stats for some items and just auto create suggestions, then it would be feasible or viable. 

 

I can appreciate this. I understand that the mod team here is small, but perhaps allowing for more roles would be beneficial. Take note of the acceptance rate of many people with hundreds of accepted suggestions. I have a 95% acceptance rate. That means that out of every 100 suggestions, five may be rejected. It would be easier on the price mods if the staff gave some people the ability to make suggestions, have them up for a day or so, and if no issues are raised (any comments whatsoever) and the votes are above 80%, then they get automatically accepted. Obviously this opens the site up to possible manipulation, which is why you don't hand it out as a reward for hitting a certain amount of accepted suggestions. Have people apply for it. And besides, everything that happens here can be reverted with one simple suggestion. I'm certainly not saying we should rush to our feet and implement this, but it's a nice idea to think about. Wikipedia employs a similar method of administration and that hasn't fallen to its knees yet.

 

16 hours ago, AD⚡HD said:

Being able to price something 30-30.11 sounds like the worst thing you could do, on items above 10 ref you should have to round to the nearest reclaimed metal

 

You've got to be joking, right?

 

14 hours ago, Mengh facepalms. said:

It's understandable to want to make a price as precise as possible e.g 30 refined to 30 - 30.11 refined, however, as Woifi said, the prices documented by Backpack.tf is merely an estimate. Using  Backpack.tf Classifieds and Bot.tf is currently the main ways to price non-unusual items, however, there is a hindrance in that those two places aren't the only two places to document sales for items, and it could be that sales at 30 refined are more commonplace, just not observed. Being too precise, when you don't have access to every single TF2 trade upon Steam to see every sale of an item, does mean that you jump the gun a little in determining a price that could be more wrong than what is currently accepted, so I believe that there's the leniency of 5% as to say that these prices are rough estimates at this moment of time for the time being. The sales price of keys on 1 day could be 30 refined, another with the majority of keys being sold at 29.77 refined (e.g a seller flooding the market to get refined), the next could be 30.11 refined, simply down to just the randomness of people listing items. 30 refined is still accurate as the point of market equilibrium, but I don't think there is a need to be incredibly precise with community prices as it is simply not possible with the uncertainty of trading, unless there is a substantial difference as given by the 5% rule.

 

It's not only understandable but desirable, in my opinion. Hiding under the guise of estimation helps no body, especially the people who come here for advice. In relation to the problem of bot.tf and our dependence on it, I mentioned that in the original post: "...it seems like it's not only easier to push aside the amount of research it once took to make hundreds of accepted suggestions, but also harder to make meaningful changes when that site doesn't have enough data." We're here to price items based on what we have on offer. If our statistics show that 30 and 30.11 work, there is literally no reason to not change the price. If I manage to sell a 2 ref item at 5 ref, I'm not going to rush over here and go "well I managed to sell this for 5 ref, so the price should be 5 ref", and I doubt anyone else will. It's an outlier and should be disregarded.

 

I'm also confused by your change of heart one line into your reply: "I think the reasons why any suggestion is made is either to better the accuracy of the current Backpack.tf pricing of the item in being able to provide a price at which the item would be valued at, or to better its precision as to suggest a change with the spread of the value of the item's sales." This is exactly right and is what I'm arguing for. Substituting accuracy for time management is no way to run a system of price determination.

[AD⚡HD]

You put more effort into the wall of text than I did into my summer homework projects.

 

 

bp.tf is a guide not the law, prices should be general and clean. 

 

Anyways item pricing is kinda pointless cause bots just base their prices off other bots not bp price

[regi]

1 hour ago, AD⚡HD said:

bp.tf is a guide not the law, prices should be general and clean. 

 

Anyways item pricing is kinda pointless cause bots just base their prices off other bots not bp price

 

With all due respect, did you read what I wrote? I am not at all saying bp.tf is the 'be-all-end-all' of pricing, but rather that accuracy shouldn't be left behind in the mist of generality.

 

Also, we're here to price items for people, not bots.

[AD⚡HD]

1 hour ago, regi said:

 

With all due respect, did you read what I wrote?

 

Not at all

[regi]

12 hours ago, AD⚡HD said:

Not at all

 

Then please, either read it or don't respond.

[regi]

Right, I just had a pretty insightful conversation with the site's resident fish. We discussed the following topics:

1. The nature and understanding of ranges
2. The idea of flat prices, and their reliability
3. What this site really is, and the differences between estimation and accuracy
4. The obscurity of the help page explaining how the site works, and how idiot-proofing may prove detrimental

I'll try to explain our misunderstandings and the ways that I think that perhaps the responses here have been misinterpreted by both sides at some points in time.

 

In regards to point one, and the purpose of ranges, we discussed whether or not having a range allows for more diversity in pricing. I'm not sure what I think of this anymore, as I can see arguments from both sides. Fish and a couple of other people I asked both said that if we say 3 ref flat, it's easier to deviate from the price, and thus there's more chance of sales outside of the accepted backpack.tf price. If there's a range, they said that people are more inclined to think that traders are unwilling to go above or below the prices in the range, sticking instead to the median or something around it. Now, I can respect this opinion, but I can also see arguments opposing it. For instance, having a range denotes that we're not entirely sure of the price, and that it being flat says that we're quite certain that this is it. From this discussion, I'm content with the prospect of saying that ranges should not be added onto existing flat prices. For instance, a 30 ref should not go to 30-30.11, but can go to 30.11 directly (assuming there's evidence to remove 30) without interfering with this rule as it's more accurate. I'd also say that perhaps we should add a clause; that prices over 3 months old can have minor ranges or updates made without contradicting the rules. In this case, it's better to have a day old range than a year old flat.

(edit for clarification: this is not to say that these people oppose ranges all together, rather that there is a time and place for them. Apologies for any misconception this might have caused)

 

Point two is heavily rooted in point one, and as such, I don't see much reason to discuss it further, besides that they are interpreted differently by everyone. We should write a guide and link it somewhere officially that denotes how to read flats and ranges, and how reliable each really are.

 

Point three is very much a restatement of the discussion we've already been having. I argue that estimation should not be used as cop-out for accuracy, but many, including Fish and those who have voiced an opinion above, consider an estimation to be good enough. I'd say that we should try to be as accurate as possible; to the nth degree, if you will.

 

The final point is relating to how not everyone knows how this site works and that we can't assume everyone reads this handy page before trying to buy or sell. I think the best way to get out of this would be to make a one time pop-up for everyone's first log in, saying something like "Hi! Welcome to the site. Please note that all prices on this site are estimations and may not be accurate or up to date. Please only use the suggested price as a suggestion, not as something set in stone". However, this is neither here nor there, but I thought it'd be good to bring it up.

 

In conclusion, I still believe that this 5% rule has some shortcomings and there was, at least in my vocal minority's opinion, some shortsightedness when it was created. I've tried adding polar to see if they'd be willing to chime in, but my friend request was rejected. Here's what I propose:

• Flat prices above 10 ref cannot be updated to have a range if the range is less than a reclaimed's worth (10-10.22 is a no, but 10-10.33 is okay).
• Key prices must see a change of at least .05 of a key, no matter the price. However, when pricing more expensive items (20-30+ keys), try to be within .5 of a key (20.5-21 keys, instead of 20.45-21.05 keys).
• Prices can be changed by a scrap's worth at any time, as long as there is sufficient proof to indicate that the current flat price is outdated and doesn't work.
• 3 months must pass before prices above 10 ref can have a range of under a reclaimed added to them.

Pinging those who may be interested: @FishtheFish @Woifi The Viking @Mengh facepalms.