That’s a mouthful ain’t it?!
Let me caveat this whole article by saying, this is my best guess at their parameters. I am by no means saying these are the parameters that GW uses. However, they’re fairly close.
Firstly, let’s take a look at the latest Metawatch table:
Credit: Warhammer Cummunity
Luckily, with our new system of obtaining data, we can now attempt to replicate what GW draws down for its data. To do so, we had to use snippets of information that we’ve gleaned from both this, and past articles.
- Firstly, we know that they are obtaining their data from Best Coast Pairings (BCP), and now Stats and Ladders (based on the most recent article).
- We know that the data covered the previous 60 days. Therefore using the date in the bottom of their table as a guide we worked this back to the 4th January, give or take a day based on the date given.
What did that mean? Well we pulled the same data, and deleted the additional we get now from Ecksen to match GW’s smaller sample size 😉
I then worked throught this database and began excluding certain parameters I thought GW would. With each change I compared the stats to their data and then made the next change.
It was a long process….
So what did I include and exclude?
- We excluded Ecksen as GW does not draw data from this site.
- We then made the assumption that GW would also count draws as 0.5 of a win and 0.5 of a draw.
- Excluded matches where the opponent had no army list or faction included.
- Included mirror matches!!!!
- Included team events
- Included double events!!!!
- Included all points sized games.
- Excluded online events using TTS
The result based on the above, is the closest I can get to match their data. By closest, I mean all bar 2 of the factions are 0-1% different. Apart from Lumineth Realm-Lords who are 3% different and Ironjawz who are 2% different.
If we change any of those factors above we get at least two more factions that are more 2% or more different to GW.
Again, this isn’t saying this is their parameters, just the closes that I can get the database to match their stats. Would GW rally include mirror matches? Possibly, as it reflects the popularity of the faction within the customer base. Would GW include doubles events where armies are made up of two factions? I doubt it, but as I say, excluding these takes me further away from their data.
On to the table!
We’ve tried to show the two stat bars side by side. Metawatch in red (they seem to like red) and ours in yellow.
Of these factions, I have managed to have 11 match and a further 13 within 1%, which could be due to rounding somewhere.
Whats going on with Lumineth and Ironjawz? I have no idea. But whether I exclude smaller points tournaments, excluded mirror matches, exclude doubles tournaments or team tournaments, or combinations of these, the variance for these two factions never gets any closer and in doing so makes others worse.
My only other guess, is that perhaps GWs data includes some other source, perhaps events held at their Warhammer World venue, that isn’t included on BCP or Stats and Ladders.
Any, way I’m done looking at this for now. But, I’ll probably have another crack next week and also share the database when I do.
