Recently, I asked for article ideas based on process. One of the questions was below:
@samstod Standing development challenges. The big questions you guys are currently trying to resolve.— Jeff Cunningham (@WJC83) May 23, 2015
I think that is a great topic, especially as we get to the end of what was an incredibly successful run of block structure formatting and embark on a new one. In today's article, I want to talk about some of Magic R&D's big, open questions in regards to design and development, and give you some idea of the things we do to work on them.
A lot of our standing issues come from the fact that we got really good at making blocks in the three-set + core set world and made a major change to our process, which we need to iterate on. Fortunately, Iteration is one of the things that Magic R&D does best. We probably won't get everything right on our first time, but I am confident that we can get there over time.
We announced the change in rotation well before it had actually affected us internally. That is because we wanted to make sure people knew how long their cards would be in Standard before we started selling you Khans of Tarkir. We work pretty far ahead in the future, but not that far. We use real world Standard to balance the set a year later in our FFL, so we are just now experiencing the first true three-block Standard as Khans of Tarkir rotated out for Tears.
We spent a lot of time theory crafting how to keep Standard interesting under the two block model, but now we need to improve it. The change removed our easy use of the core set (or the spring large or small set) as a rudder to impact next year's Standard. For example, we put Steel Overseer is in M11 and both the check lands and Farseek in M13 to make sure those blocks would have the tools to do what they needed to do, but something would rotate out after a year. We also could put some pretty mild hate cards in the core set to act as safety valves if the previous block's synergies were too strong, and risked keeping the new block from being exciting.
Things are more complicated now. Imagine a rotation like this.
We need to seed cards into sets A-B, to go with set E- F, but the development and design times between those sets is further away than C-D was to E or F, which means we are going to have less accurate shots. We run the risk of making a seed card that is too strong, or having to put too much stuff in E and F because the seed cards missed. If we put hate cards in C and D to deal with A and B, they have to also not risk the themes in E-F or G-H. And because we theoretically want to some overlap between A and B's themes and E and Fs themes, that is difficult.
Like many of the other problems on this list, we spent a good amount of time theorycrafting what was going to have to happen to make the two-block model work, and have some ideas of what to do to solve these problems, but it will be a while before we know if they work in the real world. I am confident that we can keep real world Standard diverse and fun, but there is always room to improve.
As we leave the world where core sets exist, a big question that has come up is: How do we manage the complexity of the game? We don't want new players to have no good starting point, but we also know that some complexity is good for getting players hooked. It's getting the balance that is hard, and finding that balance is one of Development's top goals over the next few years.
Keeping the game from becoming too complex isn't just for new player acquisition, though. There is a barrier for all players, but some have a higher pain point. I'm sure it would be possible to create a set where every card is written in the smallest font possible, and many of our players would love it. But, I think most would not. We crossed this line before and that's what inspired New World Order, which has been great for the game. We aren't in a spot where we need to do a New-New World Order, but we certainly need to be diligent about managing where we are going.
To give you an idea of why NWO was needed, Time Spiral had a common word count of 26.55. Future Sight went even further, with a common word count of 29.87. While there were a lot more problems there with the number of mechanics, the common word count as a bar for just how complex a set was a reasonable method. M10 and New World Order pushed things down a lot, with M10 only having 10.35 words. Words do us a lot of good, but we certainly didn't need 26.55 to make a good Draft environment—Innistrad came in at 17.3. The extra nine words per card on average added a lot of complexity, and made the game much harder to learn. I don't think Time Spiral was, on average, 53% more fun than Innistrad. At the same time, M10 was less fun than Innistrad, but probably not 73% less fun, so we can't just try to get the most efficient word count. We need to find the sweet spot where the games are fun, but also most people can grok what is going on.
Traditionally, core sets had common word counts in the 10–14 range, but they crept up a bit over time, then M15 had a big leap up to 18. Our regular expansions were hovering in the range of 17–18, but Khans came in over 19, and Dragons almost hits 21. Some of this may just be a local phenomenon—morph, on which the block was based, is a very wordy mechanic. But some of it is also that many of the set's designs in the last few years have become more ambitions. Ambitious is good, but that ambition has real world costs—some of that is just how wordy the sets are.
I don't expect our large sets to go down to 15 words or so to even out the whole year with the core set. But we are trying to figure out what we need to do to keep every set as accessible to new players as possible, while still providing the game play that our enfranchised players want from our sets.
Red's color pie
This is a big one. While red does see its share of top-tier Constructed decks, they almost always do the same thing. Attack with small creature and burn the face. Or attack with medium sized creatures and burn the face. Part of the problem is that a lot of red's power is in burning the face, and not in other things. Over time, we have decreased what red can do as a color, removing cards like Stone Rain, Wheel of Fortune, and the like…but not added enough to make up for it. Making prowess evergreen helps. It gives us a keyword on creatures that scales better than haste or first strike. Menace also does a lot, as it gives an evasion ability that is much more fun than intimidate.
We have been experimenting with things like red card draw on Act on Impulse; Outpost Siege; and Chandra, Pyromaster. This has gone well, and something you can see more of in the future. It has helped make more red midrange decks possible, but red is still pretty one-dimensional as a color.
A big problem is that red's removal is less powerful than black's…unless it hits a player. And we can't have it all hit players, so we end up in a big balancing act of trying to get the right cards for the aggressive decks, but often find that the removal that doesn't hit players just never ends up seeing play. At the same time, we have also had quite a bit of toughness creep on creatures as a whole, which has led to us creating cards like Roast to deal with them, but that isn't super satisfying.
We have some plans of things to try out, and to try rebalancing red away from being just "the color of weenies and burn" in the same way we have transitioned other colors over time. If you told me 20 years ago that black could exist without Dark Ritual, I would've thought you were insane, but all of the black cards needed to be balanced around being able to be Ritualed out on turn one. We aren't going to remove burn, but hopefully find more things for red to do so that we don't need to put all of its power in those cards.
Hexproof, as a keyword, has a lot of problems. One of the challenges with making new evergreen keywords, which hexproof was in M12, is that it leads us to overusing them in the first year or two. Once we have an easy way of quantifying it and using it in our toolbox, the barrier to putting it on cards becomes lower. And it's more likely to get used when a card is right on the line but needs a little push to get it to where it needs to be. If you look at Innistrad block, it is the largest offender, with cards like Invisible Stalker; Geist of Saint Traft; and Sigarda, Host of Herons all being cards that were definitely a little more powerful than they needed to be.
The thing about hexproof is that it is really good when used right. I think that hexproof is great on cards like Ranger's Guile, and I think that it's great when used on creatures that are kind of big and doofy, and this is the thing that pushes them over the top—but only if your opponent has no good way of dealing with them. It is much less fun on efficient evasive creature.
Within R&D we talk about killing hexproof now and then, and maybe going back to shroud, but I do think that hexproof does a lot of good—we just need to be more diligent about what we put it on. We've also been experimenting more with conditional hexproof—something that doesn't work well with shroud. Hexproof while untapped, for example, does let decks deal with the creature, but gives players a way to interact with it.
I hate to say it, but I am sure that prowess and menace, and maybe scry, may be overused in the coming year or so…until we get a better handle on just how strong those abilities are in different environments, and we see how they actually affect real-world tournaments. We try to get the numbers of these effects about right, but we are inevitably going to be a bit high or low most of the time, and can see how they play in the real world, and adjust them. Also, once we get more experience making cards with the abilities, we can often find better points for how to make them fun.
That's it for this week. Next week, I'll be back with my Magic Origins preview card!
Until next time,