ESO Xbox Sales History has too few data points. It’s only based off of so many sales, and therefore a poor representation of the market.
every critic ever
Author: ArchAmber
Date: April 9, 2019
Okay, to begin, Xbox Sales History is not a price guide. It’s a sales database. I can not emphasize this too much. Your number one resource for checking accuracy on pricing is major city traders in-game. Second is probably knowledgeable traders willing to lend their expertise. And somewhere fairly far down that list is pricing apps like ours – a tool that’s best for a quick ballpark.
Now that aside, let’s discuss real quick why ESO Xbox Sales History often has very few data points, even on some fairly high volume items. This is mainly due to what I refer to as either the Monthly Reset or the Data Dump. That’s right, we purge old data on the regular.
It may seem almost counter-intuitive at first, but the goal of ESO Xbox Sales History is to be a somewhat accurate snapshot of sales across the past 30 days in guild stores. That data gets skewed if it’s weighed down by information that was gathered one month, two months, three months ago. The ESO market is extremely fluid and shifts daily. When you have too much data, you can not show those shifts.
To illustrate, let’s look at an example: Zircon Platings.
As most know, platings in general have been steadily declining in value over the past few months. Right now zircon platings are comfortably in the 29-32k range on trader. But only a few months ago, folks were still able to push as high as 40k on zircon. The last row in the table shows where Zircon Plating would be sitting if we didn’t dump data. As you can see, the median (33k) is actually outside of the majority of sales (29.2-32.6k). Not only that, but the sales range gets much larger, both when looking at the majority of sales and the overall minimum/maximum. This data is an inaccurate representation of the current market, and it’s also a confusing representation showing sales data in a huge range from 22k to 43.5k per.
Without the reset, we would have several hundred data points on Zircon Platings, but that amount of data actually starts to hinder our ability to be accurate to the true and current market. So instead we have a few dozen data points that give us a better look at value.
This effect is not just limited to crafting materials, as seen in the major fluctuations in Spell Strategist jewelry value:
While medians have stayed somewhat consistent, sets operate slightly different in that max/min values are huge indicators of market power. As time has progressed, Spell Strategist has started to stabilize a bit, resulting in a much more narrow range of value. But this could not be indicated without dumping old data to really see how the market has self-adjusted.
How the data dump is done:
Now when we say “data dump,” we aren’t literally purging down to base zero. At the start of every month, data is instead consolidated into a handful of data points. The number is dependent upon how the data is broken down in each category. Categories that have only a min/max/median keep only their median value. Categories that have a min/lower/median/upper/max keep their lower, median, and upper values at average stack size.
For example:
When we understand how old data is consolidated, then we can see that the few data points left over actually represent a lot of data. Only three sales may be the consolidation of tens to hundreds of sales, but now our data is flexible and can show shifts in the market more accurately.
