Why do I run this site? Because visualising Arkham Horror data is a fun way to start a conversation.
I really hope you enjoy this blog and take all its data-driven conclusions in the manner they are intended: to start a conversation. Every time I post, I love seeing people ask more questions.
I do also see the same comments over and over regarding problems with the data I use. Of course there are problems with the data! So, before you comment on whatever the data problems in my latest post are, here are the stock answers to the most common criticisms:
“This data’s incomplete, and the criteria for collecting it is too vague.”
Absolutely, 100% correct. All data is incomplete. I’m here to have fun with the data that is available to us, and to spark conversations.
“Arkhamdb.com does not represent the real playing community”
Correct! Of all the AHLCG players across the world, some know that ArkhamDB.com exists. Of those, some will visit it. Of those, some will actually use it to build decks. Many of those decks won’t be public. So yes, 100% correct, Arkhamdb is an incomplete picture.
“Older cards are more common simply because they’re older.”
Correct! To mitigate this issue, I generally use Days per Deck when comparing cards from different campaigns. This measure is a ratio of number of decks and days since card was released.
“But days per deck doesn’t account for how many users on Arkhamdb.com own a card.”
Once again: this observation is 100% correct. If a card is old, but few people own it, it won’t be used much.
My metrics are not perfect. I run this blog to start conversations and provide a new perspective on card use. Data is not truth. Otherwise, let’s just acknowledge the imperfections, and have fun talking about the topic instead.
“Loads of the decks on Arkhamdb.com aren’t polished or really used, so they aren’t representative.”
Really? Yes, there are some bonkers decks on Arkhamdb (exhibit A: a 185 card, 188XP deck). But does that account for a large proportion of decks on the site? I don’t know. My hunch is that the majority of decks on ArkhamDB are intended to be playable and that the number of meme-decks is small enough not to break the numbers.
Also – how do we know that experimental decks aren’t a feature of real-life play too?
“Most decks only use Core+Dunwich, so the results are skewed.”
Again – that feels like a good hunch. But is it true? As of June 2020, the majority of decks have cards from many campaigns. Decks with just cards from 1 or 2 campaign sets are pretty rare on the site:
“This isn’t what good data analysis is.””
Let’s talk about online comments. After a couple months of doing this blog last year, I got bored of criticisms of things I already knew. The final straw was being told on Reddit about what data analysis is, and how to to do it. On one level, this is kinda hilarious (you know, coz it’s not like I’ve been writing books and teaching this for 10 yrs or anything). But on another level, I found it so frustrating I stopped blogging.
However, the curiosity of what the data contains couldn’t keep me away. Thanks to the support of others, I decided to start the blog again. So: enjoy the content, talk about the topics, but please leave criticisms about the data out of it, if they’re already addressed in this post. What I do want to hear is the questions or comments you have about the data.
4 CommentsAdd Yours →
What an arrogant post. A fabulous way to get people to stay and read more, tell them that you’re not interested in criticism, and when you do respond, do so with sarcasm. And nobody is going to give a toss about your credentials, I’m afraid.
Thanks for the feedback! I wrote the first version of this post kinda late one night. I’ve edited it to take out all the sarcasm. I hope you continue to enjoy the site.
From a data scientist and Arkham fan, please keep up the blog. It’s great!