In a long reddit comment, game dev is talking about keywords in Hearthstone.

Yesterday, Peter Whalen, who is a game designer at Blizzard, working on Hearthstone, has posted a long comment on reddit thread. The thread was about Y'Shaarj not getting the new "Recruit" keyword, which was one of the more common questions since Recruit was announced. After all, Y'Shaarj's effect basically Recruits a minion. Some players found it confusing when Peter has confirmed on Twitter that Y'Shaarj will NOT get a new keyword. However, since short Twitter messages are not enough to communicate more complex things, he explained it in-depth on reddit. 

Since the post is very insightful and full of great points detailing the keyword design process, it's hard to give a quick summary. But as a quick tl;dr: there are both upsides and downsides of creating new keywords. In case of Recruit, changing some of the old cards to feature it wouldn't make things easier to understand, nor more flavorful.

Twitter isn’t always the best place for a nuanced discussion, so let’s talk for a minute about keywords. Keywords can be great but they can also terrible – that balance is a bit delicate. So what are the upside to keywords?

1) Keywords can make learning cards easier. Once you learn a keyword, it can often boil down a complex idea into something simple and make it easier for you to understand a new card.

2) Keywords are flavorful. They tell a story – Discover, Recruit, Taunt; these all help you get a better feeling for the world.

3) Keywords can make for prettier card text. They condense text which makes it fit better. This is something of a double-edged sword because it means they hide complexity, so that’s definitely something to watch out for.

4) Keywords allow us to key off of them. So we could do something like “Whenever you Recruit this, give it +5/+5”.

But keywords have downsides too:

1) Keywords are a barrier to entry. You have to learn all the keywords to play. Sometimes you can mouse over the cards, but if you’re looking on a website that might not be an option for you.

2) Keywords sound like “rules-speak”. They sound technical rather than like plain English. One of the things Hearthstone does well is that it has accessible card text; most of the cards can be understood without any special jargon.

So let’s take a look at Recruit. In Kobolds and Catacombs we wanted to the opportunity to do some more complex Recruit cards. It’s a fun, skill-testing mechanic that creates interesting deckbuilding opportunities. You’ve seen a couple of the simpler ones that aren’t helped that much by having a keyword; there are others that it helps a lot. Recruit as a whole presents interesting deckbuilding choices – with Gather Your Party, you ask yourself how many minions do you want to include? Which ones? The randomness is also very controlled; instead of having huge variance, since you pay for the advantage of getting the full minion, the swing is much smaller. The gap between the best result and the worst result can be pretty small depending on how you built your deck. Gather Your Party also isn’t the only version of Recruit; most of the Recruit cards (like Guild Recruiter) have parameters around them that create further constraints.

This isn’t to say that Recruit isn’t going to pull out huge monsters from your deck. It is! That’s a lot of the fun! It’ll create big moments, but it’ll do it at a relatively high mana cost and with a variance that you can control.

So that’s why we wanted to make Recruit – create fun moments and interesting deckbuilding decisions. Why keyword it? We wanted something to flavorfully tie to adventurers. We wanted to make some more interesting and complex Recruit cards. And we wanted to make more of them than normal. That made it a pretty natural fit for a keyword. The one thing we didn’t want to do was “Recruit-matters”; cards that key off of Recruiting. Recruit already creates interesting deckbuilding choices; we didn’t want to restrict you to only play the special “Recruit-me!” cards (though we certainly made some minions that are cool to recruit). We wanted you to have more freedom to build your Recruit decks with the minions you wanted.

So why not keyword Y’shaarj, Finja, and Patches? These are cards players already know, so changing the wording doesn’t help someone better understand them. You don’t understand how Finja works any less because we released Gather Your Party, even if they have different wording. We don’t gain anything on the flavor side and, again, probably lose out a little bit. We didn’t want to create Recruit-matters cards, so we don’t lose out on any potentially cool synergy. And this isn’t a mechanic we want to hit super heavily in lots of future sets, so there isn’t much advantage in creating an eternal keyword. So there’s downside to changing these cards, not much upside, and we’d incur the cost of another keyword that someone new would have to learn forever. So we decided not to keyword the old cards (and likely won’t keyword new cards unless we made it an important part of a set again).

There’s been some talk about consistency. Consistency is great when it leads to understanding. You see a card, you learn it, now you can understand another card. That’s a great learning experience and a powerful tool for getting a handle on something complicated. When consistency hinders comprehension rather than aiding it, it isn’t worth it. This feels like one of those cases. By changing the old cards, we make it harder to understand them for a small aesthetic gain. But in changing future cards, we meaningfully increase the complexity burden to start playing and, especially if we use Recruit sparingly, might actually increase the comprehension complexity for an invested player who’s trying to learn a new card.

So overall, on the Kobolds and Catacombs cards, we gain clarity and flavor by keywording Recruit. On past cards, we lose clarity and flavor by keywording Recruit. On future cards, we generally lose flavor and, especially once Recruit is out of standard and many players have forgotten about it or new players haven’t learned it, lose clarity and increase complexity by keywording Recruit. Like most issues, this isn’t entirely one-sided, but overall we felt it was better to keyword Recruit here and not keyword it either in the past or future.