January 27, 2010

Apologies to everybody trying to get to my Google App Engine site. That shout-out has run me right out of bandwidth. I’ve been linked to from before without being overwhelmed by it, so this has taken me by surprise a bit.

Normal traffic for the site is well within the bandwidth limits that Google sets, so I’m sure I’ll be baaack soon. Just bookmark the link and try again when the storm has passed…


15 Responses to “slashdotted!”

  1. heath Says:

    We have a question over at (Warlock site)…How is your query handling dual specs? Does your breakdown by spec pick up both specs, or just the spec that the character last logged out on?


  2. zardoz Says:

    Just the “active spec” – the spec that the character logged out on. I’m assuming that, given the sample size, everybody’s spec switches would cancel out so that the distribution of active specs more-or-less represents the real popularity of each spec.

    But that is just an assumption. I’ll adjust the text over at my Google site to make the methodology clearer on that point.

    The talents page for each class – which you can find from the menu system – has a much more detailed analysis of spec popularity which does fully account for dual specs.

  3. tankadin Says:

    I have some concerns about the data in your protection paladin “stats” page.

    The first is level 80 “mainhand speed” being 1.2 for the minimum. There are no weapons in the game other than wands and thrown items and both not usable by paladins.

    The second is the offhand dps. Protection paladins should have a shield equipped which would be zero dps. They also cannot use off-hand weapons, so I seriously doubt some of the “maximum” numbers.

    Do you think you could do a post some time with links to some of these “anomalies” in your data?

    Which leads me to my next question… Do you think you should throw out the outliers to get a better picture of the true averages?

    And speaking of averages, would the “mean” be useful to anyone in addition to the average? And/or possibly some sort of a “weighted average”?

    I am not a statistician so I’m just speculating on what might be useful (or even possible).

  4. zardoz Says:

    I’ll have a look at where those numbers you cite are coming from and get an answer for you.

    But the stats pages give a pretty high level picture and any “average” will cast a very wide net – too wide, sometimes. I’ve encountered toons in the game that appear to be set up for tanking, but in fact aren’t carrying a shield. My queries are generic and will pick up all sorts of oddities like that. Showing median and standard deviation is on my to-do list, and that may help a bit.

    The problem with filtering outliers, or weighting stuff, is that it introduces assumptions into the data. Obviously, some assumptions, such as warrior and paladin tanks should have shields, are pretty safe, but many other assumptions can be a lot more debatable. I’m reluctant to go too far down that road.

  5. Aaron Says:

    Been toying around with grabbing and parsing the armory XML. Was wondering if I could ask you a few questions about your method and grab rates and whatnot.

  6. zardoz Says:

    No problemo; I’ll be in touch.

  7. Zinerva Says:

    This is very cool. Are you maintaining a history of server population over time?

  8. zardoz Says:

    No unfortunately. I build separate databases for each scan. And I only do a large random sample; I don’t collect every character that there is. So there are some questions like that which I can’t answer.

  9. coldbear Says:

    Really like your approach to analyzing wow trends. Recently Ghostcrawler has been a bit snarky when it comes players posting links to statistics and trying to support their arguments. Granted he may have good reason to do so – but I’d love to see you analyze and post the following kinds of information:

    * Win/loss ratio analysis for WG on low-pop unbalanced realms.
    * Dps boss analysis in ICC per spec
    * Feral druid effectiveness in arena pvp (ok, fine – all specs)



  10. zardoz Says:

    My contacts with Blizz have been pretty friendly but stats are a dangerous (and somewhat unguided) weapon, no doubt about that.

    I’ll have a look at all those suggestions for new reports. I can’t promise any action in the short term, but I’ll have some free time later on which I want to use to update the site. I had the impression that arena stuff was pretty well covered by other sites, but I haven’t looked at them for a while.

  11. Baleygr Says:

    Hi Zardoz,

    this is what I would love to see:
    * horde/alliance population ratio per battlegroup
    * faction win/loss ratio in BGs per merged battlegroup
    * faction win/loss ratio in BGs per realm or per battlegroup (not merged, i expect here the values for horde and alliance need not add up to 1, see below)

    A bunch of friends are considering to play again, and we are looking for realm where:
    -Horde is not too much underrepresented, so we won’t loose every time in open pvp zones and the risk of our faction on the realm suddenly dying is minimal. I can check that over at Okoloth’s site, he has realm reports -thanks for linking
    -Horde is considerably underrepresented in the battlegroup, waiting in a BG queue when you only have some minutes to play is not nice. I found XML for the battlegroups, but not for the realm reports, so adding up the data by hand would be quite time-consuming
    -Realmtype is PvE, one of our players insists. That leads to the biggest problem: merged battlegroups. There were pve only battlegroups, but they got merged with pvp battlegroups. The teams themselves are not mixed. Now teams from the PvE-battlegroup play against teams from the PvP-battlegroup, and loose quite often, I presume. The player I mentioned is a little thin-skinned when it comes to loosing when he doesn’t have a chance, and gets frustrated quickly (allready stopped playing once because of constantly beeing ganked while leveling)

    The information I need to choose the right realm is really hard to find.
    I also checked warcraftrealms, but their stats seem to be inaccurate.

    So it would be terrific if you could help 😉


  12. Teni Says:

    OMG I bow to your greatness. I was just talking to guildmates about whether they knew of any sites that data mined the Armory for info; specifically, I wanted to know what the overall trend was for Disc priests now that active ICC raiders have enough emblems to purchase 4 pieces of tier gear. The set bonuses for Disc really sucks for t10 but I wanted to know if this was something generally agreed upon through the community… and yours will do it!

    That said, can you please do a sweep of priest builds again soon, once you’re able to? I would appreciate it very much!

  13. zardoz Says:

    I’ll be doing another scan soon. Can’t promise any new report types for a little while but, I’m flat out elsewhere.

    I do have enough data to look at BGs and battlegroups. Let me see what I can come up with.

  14. coldbear Says:

    @ Baleygr

    I’m pretty sure most of the numbers you’re looking for are already available on

    …which, however, do not list WG.

    Which is my main issue of contention, currently.

  15. coldbear Says:

    I’d be willing to do a bunch of legwork to get the “Number of times defended” for both Horde and Alliance on all US realms – if you were willing to take the activity & number of characters per realm data from Armoury and and interpolate(?) the data for some reasonable conclusions.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: