Wednesday, June 19, 2013

TMC Followup: Burn Jita

Burn Jita: An Industrial Perspective

This article took longer to write than I originally expected.  I should have anticipated that a factual article is going to take longer to gather together and present.  Also, continued trouble with Google Drive processing power keeps putting a damper on my progress.

Why Freighters?

The simple answer is freighters have been at the front of my mind recently.  With the JF Project, my own BPOs, and my historical charting work, freighters have proven to be the most directly useful and recognizable to the widest player base.

Also, by combining the data into freighters as a group, rather than a specific ship, I can accomplish a whole list of things:
  1. Avoid graph overload.  Having to reference 4 charts is worse than referencing 1 chart
  2. Provide a higher level to talk at, rather than get tangled in the minutia of each item's behavior
  3. Not share "secret sauce" to the global audience that reads TMC while still making useful content
I did want to also show T2 mirrored the post-event dumps, but it ended up on the cutting room floor.  I instead used the semi-interview with Warr Atini (head of Miniluv), staying on-topic.  I was afraid of running off the rails with 5 pages of analysis for each sub group and speculation therein.


Pre vs Post Event

This was another topic that fell out pretty fluidly.  The TL;DR, which matches my intuition, is that as a producer, pre-event is nearly useless (except for volume spikes, if you're wanting to cash out), but the post-event is an enormous opportunity to save a lot of money.  I don't usually stock raw materials, but come the next big gank-event, I make efforts to buy up some percentage of materials to suit me for 30 days.

The interesting thing I learned was the speculator side of the equation.  I've never had success participating in the margin/speculating trading, and I'm starting to see why.  Burn Jita was a margin trader's paradise, peaking prices on gank fittings, but that price bubble evaporated once the event was complete.

Missing Part of the Picture

The big missing slice of my article is other hub analysis.  I did start fiddling with feeds to try and generate that data, but found the results too messy and hard to display in one or two graphs.  I also skipped T2 production as a topic because it added another page without adding a lot of value.

I expect to take both the T2 and "everywhere but Jita" topics into their own articles.  The not-Jita article will probably be next, since I still would like 2 weeks for the T2 market to shake out. 

The To-Do List looks something like this:
  1. Write Decryptor Tutorial with new decryptors
    • Working on refreshed 2.0 tools
  2. Draft Mining Report artcile for TMC
  3. Draft "Anywhere But Jita" article
  4. Write T2 Post-Odyssey analysis
    • Working on new charts to group data intelligently 

Data Sources

One of the themes I'd really like to drive home here is "this data can be mined".  These graphs were generated purely using EVE-marketdata's history API.  To increase quality, I had to download their private data pusher to get the history graphs to reflect reality.  

As I wrote before, it took writing some custom code to process it, but I now have a moderately easy to use tool, except when Drive is under heavy load.  In an effort for transparency, so I don't get accused of making it up, I hope to publish all my data sources for review.
DISCLAIMER: history feeds tend to break 9a-5p M-F US TZ.  This is a limit of Drive, please be patient.

None of this data was collected with any special privilege.  The gank fittings came from goon's killboard, the price data from EVE-Marketdata.com.  A theme I would like to push among my articles is "the data is already out there", though you may need to process it.

Error Bars

There's one glaring assumption I made that still makes me uneasy: summing sales volume where prices were averaged.  On every graph that shows volume, I just summed together the volumes to get a general picture.  I feel reasonably okay on the freighters (everything is similar order-of-magnitude), but had a lot of trouble justifying it on the ship fits and mineral report.  I played around with a few different weight modifiers, but never found a combination that I felt was truly honest.