Home | About | Tournament Winning Decklists | Forums

Know The meta


#1

Not sure if he’s on here, but @necro (other name here?) has been busy grinding numbers

http://www.knowthemeta.com/

With Film Critic the decline of Jinteki was kinda expected, but this is far more than I expected.

@necro ISNT the same guy


#2

Wow this is a great tool. Thanks to those behind it.

Where is it mining data from? Acoo or NRDB?

nvm just found it in the Info section


#3

Both + Stimhack winning decklists

http://www.knowthemeta.com/Info.html


#4

I was wondering why Anarchs sky-rocketed at The Underway, then I checked the set.

Faust (and Peddler)

The two red cards made a monumentally bigger impact than the 6 blue ones


#5

This is really cool! Thanks Necro!


#6

Wow, this is awesome. Really cool to dive into.

It’s interesting that (faction) diversity has been declining ever since Order & Chaos—when it was really good. It’s easy to complain about a particular faction being underpowered at the moment but props to FFG because it’s incredibly difficult to get close to the balance that was going on during that period, with every faction seeing a pretty fair share of play.


#7

Don’t give credit to me, but to @ necro (@necro isn’t the one)

@Stiv , you know his name?


#8

The release of Data and Destany coincides with a big jump in kit decks. I’d have thought that you’d see that with seamus’ success with kit in UK nationals, but you don’t see a bump for almost six months later. What’s that about?


#9

Hey dudes,

I’m the guy who created www.knowthemeta.com :smile: Thx for the interest and the good words. If anyone has ideas or feedback, please share.

Just a clarification: I’m mining deck, tournament and ranking data from Acoo and Stimhack. (I’m also trying to prevent multiple entries, sometimes tournament winner decks are on both sites) I’m using NetrunnerDB to get basic card information (name, type, subtype, etc.).

Be sure to use the data pack drilldown, there is so much more data there. Once you selected a data pack, check out “Decks” and select an identity to get the deck similarity matrix and card statistics.


#10

Also you can check out my article on the SanSan cycle using the mentioned data here: http://nedrunner.nl/reminiscing-the-sansan-cycle/


#11

I have a new article on the NAPD Most Wanted list here:
http://www.knowthemeta.com/Blog/NAPD-most-wanted/


#12

Hey you! Looking for more statistical data for whatever reasons? I have more statistical data on whatever card! For example:
http://www.knowthemeta.com/Cards/Clot/

Go crazy with ze clicks. I’m pretty sure that you can figure it out. Just go to the Card list or to any deck related table, list and you will find the card links.

Your feedback is still loved, cherished and thought of fondly while I am in a bar and drinking. Often at other less significant times as well.


#13

Alright :wink:


#14

Thanks; this is interesting, and very disheartening in many ways.

The most relevant card from O&C Corp side is CVS… a neutral. The next most relevant, at a whopping 8% is Traffic Accident, which we all know is only played in NBN.


#15

It would be nice if the usage percentages for non-neutral cards had in faction vs out of faction in addition to a total percentage. It’s neat to know that Account Siphon is being used in 32% of decks, but how much of that is in faction vs out of faction? Or maybe have the “Average card usage in identity” statistics per faction in addition to per identity.


#16

Thx for the idea, I will consider it. :slight_smile:

Guys, if you would like to see stats based on Crunchums’ idea, please like his comment.

http://www.knowthemeta.com/Cards/Account%20Siphon/
Well, you can already see the the top 5 identities with Account Siphon and their overall percentage. You can figure something out of that, right? I can extend the list to top 8 or top 10 easily, if that helps.


#17

I would love to see ICE breakdowns, such as most-used ICE in tournament/top decks. Right now we just have top 20 cards of all types.


#18

@ProfNecro

Great site. No, scratch that… really great site. Couple of points of feedback:

  1. On the Deck Drilldown page, when you select a Deck and click “See decklist”, this link doesn’t work on a tablet and nothing comes up. Tried on Android and iOS. Not sure why.

  2. Again on the Deck Drilldown page, it would be awesome if we could look at the stats below for a subsection of decks on that ID. Eg, see the card distribution on all Foodcoats ignoring HB:FA. Ideally you would be able to draw a box on the Deck Similarity matrix and the stats to be calculated on-the-fly only from the decks underneath your box. I’m not sure if you are using Elasticsearch for your data layer (check it out; it’s very good) - it can do these on-the-fly aggregations.


#19

@dodgepong Nice idea, ICE and ICE breakers are important, maybe I will implement it in the near future.

@iReidBooks Thx for spotting the tablet problem, I will check it out.

As for advanced filtering and complicated queries: What you can see on the site are pre-calculated, cached, static content. Statistics are not calculated on the fly as you visit the site. The reasons are to make it faster for the visitor and less stressful for the backend and easier, cheaper to host. Currently 1500+ query results are stored. Introducing more filters would increase this number exponentially. When a new idea comes up I have estimate how useful would that be for the average user and how many users would actually use it. Also replacing the data layer means a lot of work for me :neutral_face:

I believe that filtering for cardpool and identity is generally a good way to go. New cardpool usually shakes up the meta and results changes in deck composition. Different IDs usually mean different archetypes as well, it’s rare that you can build the same deck concept with different IDs. There are always niche situations of course.


#20

@ProfNecro

Yes, I realize the results are pre-calculated. With Elasticsearch (or similar technologies) though you don’t need to do that. You simply load the decks into your data layer, and have your HTML pages demand the aggregations (as POST requests typically). The data layer then generates stats on the fly, which for data sets of the size you are dealing with should be instant. The technologies themselves can then worry about caching, etc, and as a visualisation designer you focus on serving up the best data science.

The problem you have at the moment is that IDs with multiple archetypes, eg, Kate, NEH, ETF etc have data for all their archetypes rolled up on one page. Its not really relevant that 29% of all NEH decks are running News Team - its much more relevant to know what % of NEH kill decks are running it.

I agree though that this would take a bunch of migration effort if you are doing everything differently at the moment.