Jan 11

Card Popularity Data Mining – Kaladesh

For those interested in statistics, here’s a fun segment I plan to do quarterly for each main expansion release. The MTG game plugin uses a scripting system that interacts with an online database of card scripts. I had set up a quick statistical logging code that tracks when a card script is requested, recording the number of hits each card gets. Each card script only gets requested once per hosted game, meaning a specific card will only be tallied once per player, per game session, regardless of how many times it is used in-game.

Here are the top 30 most played Magic cards in OCTGN, starting at Kaladesh release update:

Forest — 16818
Plains — 16806
Swamp — 16124
Island — 16049
Mountain — 14684
Evolving Wilds — 3350
Sol Ring — 2946
Polluted Delta — 2680
Lightning Bolt — 2625
Windswept Heath — 2522
Flooded Strand — 2469
Wooded Foothills — 2276
Bloodstained Mire — 2108
Path to Exile — 1934
Command Tower — 1703
Clue — 1638
Cavern of Souls — 1629
Scalding Tarn — 1618
Stomping Ground — 1558
Misty Rainforest — 1537
Gideon, Ally of Zendikar — 1527
Prairie Stream — 1507
Birds of Paradise — 1484
Temple Garden — 1474
Overgrown Tomb — 1470
Verdant Catacombs — 1444
Shambling Vent — 1430
Hallowed Fountain — 1420
Brainstorm — 1393
Zombie — 1350

Feb 19

400k anonymized OCTGN stats now available

It’s been a long time since I extracted some stats, so now that the second stimhack league is over, I have finally done another one. This time the number of games is more than 400.000 which is bringing my spreadsheet editor to its knees while preparing it for anonymizing.

You can find the latest export raw data here

I’ve also gone ahead and separated the Second Stimhack league (SHL2) stats as well. This is about 4000 very competitive games for the period of November 2014 – January 2015

Apr 02

210k Data Set updated

Ooops! Just noticed an error in the way I anonymized the data set caused some rows to break, ending with game names in the wrong place etc. Fixed it now and you can find it on the same place.

Apr 01

210k Anonymized Data Set available

Now that March is finished, and we reached  a nice round number of over 200k matches, I decided to pull out the next data set from the OCTGN, totalling around 212k matches. This should reach until the early availability of Double Time.

You can fetch the .csv from my google drive.

Please share any results you compile.

 

Mar 09

Analyze your OCTGN games

Sample Analysis

A sample game, analysed. Click for larger img.

The people behind Acoo.net have released a very interesting little web application that takes your OCTGN Android:Netrunner chatlog and analyze it to provide a game overview. It’s kinda neat as it provides you with statistics of how many credits you made, how many ICE you rezzed etc.

Check it out here

To use it, simple go to your chatlog after the game ends, press Ctrl+A, Ctrl + C, to copy all text to the clipboard, then go to the site, paste it in the box and press the button to analyze it. If you’ve played more than one game in the same session, make sure the text you put in contains the logs only from a single game.

I don’t know how the magic happens behind the code but it’s some impressive feat of text parsing, especially since my code tends to be standarized but often provides customized text as well. Not to mention all the times players take back actions (including credits and clicks and so on) so it’d be interesting to know how they manage those situations.

PS: The analyzer is still in beta, so expect some bugs and improvements in the future.

Nov 22

Stat-tracking for League and Tournaments

I see various people running their own little leagues and tournaments but usually nobody bothers to inform me about them, or make sure their stats can be correctly tracked. Usually this is not a problem, but if there is any confusion, and you want me to inform you of the stats of how a particular match ended, that is not easily achievable, as non-league stats are hidden by default and I need a full export to reach them, which can take some time.

A better solution would be, as a league organizer, to make sure you contact me to tag your league matches within the code of ANR in OCTGN. This way I can both retrieve a matches stats easier, and I can even provide a full statistics spreadsheet for your tournament or league after it’s finished.

All  you need to do is contact me and let me know two things:

  • Name of the League or Tournament (e.g. “Accelerated Beta League“)
  • Unique game tag that all your participants will be using in their game title. (e.g. “[BGG-L02]“). The tag should always be surrounded by [square brackets] to distinguish it in the title.
    Using this tag in the title will automatically mark a game as a league match, but even if you forget, you can use the game menu to record the match for the league.

So send me those details before you’re about to begin and I’ll make sure you matches are tagged properly in the statistics. If you don’t however, please do not ask me for individual game stats, as that’s far too much effort on my behalf.

Sep 12

100k anonymized matches now available

I’ve just anonymized the latest statistics export for the ANR games played on OCTGN and made it available publicly. If that kind of thing is your area of expertise, feel free to perform some interesting number crunching 😉

Aug 13

C&C Runners are Dominating

Oh dear, the stats don’t look good. After reaching a 50% average balance, the runners are now running roughshod over the hapless corporations. AlexFrog has some interesting points to make about this sudden change and I’m inclined to agree. Whereas runners got great new toys all over the board, all corporations other than HB got very little they can consistently use.

Let’s hope this unbalanced spike flattens out a little once corps figure out how to adjust.

Jul 04

~72k anonymized games dataset

I recently got a new export out from the slaghund and I’ve prepared an anonymized version for people to crunch. This includes all the game stats but player names have been anonymized but still consistent from game to game.

Unfortunately these stats will contain some dirty stats, such as people doing test games against themselves and so on, but these should be statistically insignificant. However if you see something like a corporation flatlining, you can safely delete the entry.

You can download the export here.

Also worth mentioning is that the Alpha BGG League has ended a few days ago. shobalk and his troupe of magical stat-crunching elves are already preparing the league’s results and if you want, you can also get your hands on them.

Jun 07

New A:NR analysis posted

99th statsHollis has just released the second part of his analysis of the balance on the 99th percentile of A:NR OCTGN users. He actually crunched the Glicko-2 rating of the players in the 40k dataset and set about seeing how the runner-corporation balance sits in matches between high-rated players.

Definitely worth taking a look.

As an aside, I’ve now also created a Statistics page, where I compile links to such interesting articles and also provide the raw data dump for people who want to do further analysis.