Sorry for the slow response - I discussed this with the other moderators and then the suggestion got lost amidst the workload :(

While there is truth in that, it does not mean that a pure sale automatically overrules all else: even pure sales can be outliers.

The idea behind the desired usage of pure sales is that we attempt to reflect pure values as much as possible, and what better way to do this than to use a sale consisting of keys only? That's the best way to ensure something can actually sell for the listed amount of pure.

However, in the situation here, an item that would be worth 420 pure had to add almost 150 in other items for a ~370 key hat. While not expressed directly in keys, this does cast a lot of doubt on whether 420 is a good pure approximation for this item. Even in the assumption that 370 would be low for the hat it sold for (which could theoretically also be true; it is not in this case, but for the sake of explanation lets assume it could), the sale there had a buyout of 400 pure, which means that in terms of pure values for the items involved, the aggregate is considered to be 400 (max). Valuing all other hats there at buyer's value (which would yield the most generous value) would still put the pure estimate for the medgun in that sale well below 300.

With that in mind, we now have two sales, one at 420, and one that - if used here - would be around 220-230. In such situations, where the sales are so far apart, it is generally best to wait for a third sale to determine the outlier. In this particular case, the average of the two sales is actually surprisingly close to its current value, which would mean that on average, current is not a bad reflection of its selling potential. As such, I believe that leaving this outdated with the data we have here is the best course of action.

As an additional explanation on the usability of complex sales:

You are absolutely right in saying that in most situations, sales like the one for the team captain here are not favorable for use in a suggestion, because it is usually tricky to determine the value of smaller items when a bunch of them sold for a larger one; the smaller the item in comparison to the 'large' one it sold for, the more this holds (e.g. for hats like the spiky viking or the noir in this trade, their value is so minimal when compared to that of the TC, that this sale cannot possibly provide accurate estimates for any of those tier items involved). In such trades, you sometimes see that despite one hat selling for multiple others, that on the other side of the trade, one item makes up the vast majority of the value. In this case, the medgun is that item (it makes up approximately 75% of the full value of the hats with its current value in mind, as opposed to the smaller chunks of ~4%, ~8%, ~10% and ~14% the other hats would be). In such cases, the sale may still not be optimal, but it would not be completely unusable for those.

As a general rule of thumb: if a hat makes up less than 50% of a trade, you generally don't have to evaluate the sale, but if one makes up more than 50% of the trade, it is usually at least helpful to see how it would compare to the other sales: if hat X sold with 49 in adds for a 100 key item, and its other sales are around 55, then its likely that 55 as a value for hat X is probably alright. it's generally not necessary to include the bulky sale in such a case, but it does support ~55 as a value. If in that same scenario, the only other sale on hat X is for 150 keys, then the bulky sale is a strong indicator that 150 is going to be too high. In neither cases its a good idea to make a range with the bulky sale, but in the latter case, its implied value does put the value of the other sale in perspective. This does not mean that the bulky sale could not be an outlier; if there are many sales around 150, then that value is likely still right despite the more complex data point implying otherwise.

These are generally reviewed on a case-by-case basis depending on the exact situation (how many other sales, how large a percentage of the full trade does a hat cover etc.).

I hope this helps you understand the rationale behind our decisions. This is probably quite a read; if you have any further questions (regarding either this suggestion, or possible future ones), feel free to message me or one of the other price mods on Discord.