Magic: The Gathering Drive to Work Podcast - #1245: Test Decks

Episode Date: May 30, 2025

This podcast talks about a tool we use to playtest new mechanics.

I'm pulling on my driveway. We all know what that means. It's time to drive to work Okay, so today is all about test decks for which many you will say what our test decks Well, I will explain today. I'm getting into the to the nitty-gritty. So first off, what is a test deck? So There's a point in design, often early, usually in exploratory design, where we need to test a thing that we've never tested before, a brand new thing. Now, normally in design, the most common play testing we have is what we call limited play testing, where we're playing either sealed or draft. And that, you know, once you have the ability to do that, that is a very good way to test your things.

We need to test sealed, you know, limited formats anyway, and just using that as a means to have a chance to play the things. Now, there are a couple reasons why limited play tests won't work. One is you're testing actual formats like standard and you're building real decks and you wanna sort of test constructed. And for those, you have what we call a play test deck, which means you're building a deck

in the environment that you're playing in. So let's say you're playing in future future league, which is standard, but you know, a couple of years in the future, you still have, it's a 60 card deck, You're following four of rules. Like you're following all the rules to standard. And then you've access to any cards that are in standard, you know, within the, the time you're playing, you know, the future standard, these are the, these are the sets that are in it. That's what you can play from.

You can choose from those sets. But that is normal play testing, that's not what a test deck is. A test deck is, I'm testing something that doesn't exist, usually it's before there is any ability to play limited. For example, in exploratory design, we haven't made the file yet, there's no file to play, there's no limited to play. Also sometimes in vision or early set, you need

to try something out that is a mechanic. Like it's a mechanic that's more for constructed than it is for limited. So sometimes there's things that's harder to test than limited. Anyway, so before you have a limited file or even after you have a limited file, but you need to test something that limited won't demonstrate, we have to make test decks. Okay, what are test decks? So test decks traditionally are either 40 or 30 cards. Originally they were 40, because we're just trying to make them as easy to make as possible the smaller of a deck. I think at some point we realized that 30 cards was enough test to do a play test game. We still like I still tend to make them 40 just because limited

is 40 and just it's easier to match all the things you know from limited. As we do more 30 we get more used to 30. But anyway, a test deck is 30 or 40 cards. And some number of the cards in the deck are brand new cards, cards that you're testing. For example, let's say we're in exploratory design and I have a brand new mechanic. So let's say we are going to bakery world and we want to, so we do a brain, you know, normally in exploratory we'll do a brainstorming. I'm like, oh, what would you expect in bakery world? Well, there's probably cupcakes and there's breads and

What what are the verbs? What are the verbs, you know, and one of the verbs we get to is bake you bake in a bakery Okay, what if we made a baking mechanic? What would a baking mechanic be? Okay. Well, maybe it's like monstrous Except it has to be in exile for a turn. That you have to sort of let it bake. And the idea is that makes it a little bit cheaper, because you're losing utility for a turn. But in the end you get a better thing. So you have to bake it, you know.

And so you have, they represent baked good creatures or something, it's cupcake people and loaf people and stuff. So I'm going to try, I want to try out the bake mechanic. So how do I do that? Okay, what I need to do is I need to test it. So first and foremost, it means I need to make some bake cards. Well, there are no bake cards in existence, the mechanic doesn't exist. So I have to go make some. So real quickly, well, actually, let me finish this then I'll get into our database. So I need to make some big cards. So I sit down and go, okay, well,

what do baking cards wanna be? Now, normally we build a test deck, we tend to build two color decks. Sometimes we build one color. We really don't build three plus if we can help it. The problem is you just don't want the man to mess up your play test. And so normally play test decks are one or two colors leaning toward to the reason we like two over one is it just lets you express the mechanic you're trying out in a little more flavor.

So normally let's say I'm doing baking. I go, oh, what colors do I want to be baking in? Like, well, I want to try something. I want, you know, I want, baking is going to bake for one turn, but maybe I'll have one car that bakes for two turns. And I want it to get real big. So I want it to be real big. So I'm going to, green will be one of my two colors.

I can have a big two, make a giant, whatever, a giant challah man or something. Holla is a bread for those that do not know. Jews eat the bread, it's my go-to bread example, which everyone might not know. And then I also say, okay, along with green, I wanna show, maybe I'll put some red with it so I can do some faster and slower stuff and try some

different things. And maybe I want something that comes in the first turn and you can bake right away. So I'm going to try red and green. I want to make a red and green deck. So what I need to do is I need to make some baking cards. So with a new mechanic, you want anywhere from, usually we normally we want about six to ten. Sometimes we get as low as 4.

Sometimes we get above 10. The real question is, how often will the mechanic come up in your gameplay? If you're trying to do something, for example, bake is going to go on creatures. I mean, I might explore putting bake on an enchantment or something. But the idea is, it's going to go on a creature.

The creature is going to sit on the battlefield. And later on, when I have the mana or the tempo or whatever. But the idea is, it's gonna go on a creature, the creature's gonna sit on the battlefield, and later on when I have the mana or the tempo or whatever, I'm gonna bake it. And so I don't need tons of creatures because the chance of me getting to it, I just gotta play one of these, and odds are I'll get a chance to bake it at some point. So maybe I decide, okay, I'm gonna make eight,

I'm gonna make eight baking cards. Now, I don't have to make eight unique baking cards I could just make four baking cards And then double them and put them twice in the deck normally the rule of thumb that we tend to have is It's like I said, it's 30 to 40 cards We only tend to put commons and uncommons in our test decks every once in a blue moon You put in a rare but rares really can disrupt things

and you don't want to do that. And normally the general rule, it's not a hard and fast rule, is that you tend to have two copies of comments and usually only one copy of uncommons. If you're trying to test something, you really need two uncommons, you can have them. It's not that you can't break that rule,

but in general, we're just trying to get get once again. The idea of a test deck is we want to sample the mechanic We're playing we want to get a sense of oh, I'm trying out baking is baking fun So I I want to make an environment I want to make it sort of something that will give me an opportunity to see things I want to be generally balanced against other decks that I'm playing. That's why we're on some level We're kind of making a limited power deck I'm just using commons and uncommons and more commons and uncommons and I'm trying to make something It's kind of roughly like what you would see in limited. That's kind of power level

We've learned, you know now the one thing that goes on here that's not limited is we're just playing a higher concentration There's a good chance I'm limited. I'm not getting eight bake cards. That's a lot, but I'm trying out. Once again, early play testing is not about balance. It's about exposure. It's, hey, are these cards fun? I wanna see what they're like. And so just getting more exposure to the cards is important.

So the percentages of what it should, the Asfan that it shows up in a test deck is not necessarily normal. And that's something you keep in mind when you're playing with it, that it's gonna be a little, it's a little concentrated, but once again, you're sampling it. Concentrated is fine when you're sampling.

Okay, so I will make, let's say I make eight big cards. Now the rest of the cards in my deck, I don't have to make brand new cards. I can do whatever I want to, and I'm not playing standard or anything. I can go get cards from anywhere. Normally we tend to make cards that are a little newer only because they're more representative

of where magic is at. Things shift over time. So normally we're playing with newer things, but we don't have to. If the perfect card is 20 years old, that's fine. And if you need something that you can't find, like one of the goals of playing with the test deck is you really don't want a lot of complication

beyond the mechanic. And the reason for that is you sort of want to really let the mechanic shine. You want to focus on the mechanic. And if you have a lot of other mechanics going on It just can muddy the water and make it a little trickier to tell what's going on that once again That's not it. Nothing. It's not that you can't have other mechanics, but if you do you want to be pretty simple You don't want complex mechanics where you're trying to test a new thing And in general

a new thing. And in general, you tend to build them kind of like limited. Why I like 40 car just because when you build limited the default is you want 16 creature seven non creature 17 spell lands. That's roughly what you want. Oh, from a lands by the way, when we do two color lands, we're allowed to have four dual lens of whatever dual lens we want sometimes we'll put actual alpha dual lens in other times we'll put in shockland it's just something mostly the idea is you just want something that helps your mana the goal here is not to like not be able to play cards to join up the mana for it so we give ourselves good mana just so we can play the cards again we're not balancing anything we're just trying to test things and see what we want to see how they play. And so we would just want to make sure that

when when we get the things we're able to play them. Okay, so I make my new cards, I pick the cards to the rest of my deck, maybe I make some other cards, maybe for example, maybe I think it's cool to have an enchantment that bakes a creature, you know, it comes into play and it removes a creature and bakes it and brings it back stronger or you know maybe I want some baked cards that aren't on creatures but some you know I can try things I can sample things or maybe I'm trying to do something simple and I just the simple thing I want doesn't exist that happens more than you realize and so maybe like I just need a three drop that does this. Okay, I'll just make it. Okay, so where do we make these cards? Let me let me get into that. So we have a database that we use. Obviously, we have to print magic cards. So to do that,

we have to have a file that not only can we R&D use, but that editing can use the people who lay out the cards, the printing people. So we have a database and the cards in the database stay in the database for the whole run and that allows us to assign art and do all sorts of things to them. Our database is called Drake. This is the third database we've had. The previous database was called Multiverse and the one before that was a FileMaker Pro but it didn't have a name. When I first got to the company, our database, there was a guy named Bill that ran the database,

and it just was a FileMaker Pro. And like, if you had an issue or a problem, you'd call Bill on the phone and go, Bill, can you change this? I need to do this. But that was, we were much, much smaller. We were making a lot more cards than we were back then. And then eventually we made an official database called Multiverse and then we updated it later

to Drake. And the idea is, so the database does a bunch of things. One is you can make card files and not just card files for official sets, you can make cards for anything. So each one of the designers has what we call a sandbox. And what that means is it's a place for us to make cards. That if I want to play test cards, I need somewhere to make the cards.

And so I have my sandbox and my sandbox are my cards. The one quirk about the sandbox is the way it works is once you make cards, we have what's called a deck builder. And what that means is, let's say we plan on play sealed, or we want to do draft, or we want to do draft, or we want to make playtest decks, or test decks.

We have the ability to print cards out of the database. We used to print them on stickers. In the early days, when I first got there, we would print them on cardboard and cut them up into card size. Then we got the technology that we could print them on stickers and then put the stickers on cards. And now with the technology, we just have blank cards and we print directly on the card. It's a magic back with a blank front. We get these from the printers. And the idea is that it makes it look as much like a magic card as you could imagine. It's in color. There are frames. There

are... you can print in black and if you just need down and dirty, you can print in black and white. If you just need down and dirty, you can print it black and white. Sometimes for test decks, sometimes we just do black and white. But the colors look pretty. And if you have art, if art exists in the file, whether it's sketch art or final art, that will also show up in the card. So one of the nice things, especially for the playtefters,

is as you get more of the cards being closer to sort of real you get more and more the cards looking real. Now I'm talking about super early so a little less real for sure we have no art. I mean the cards we're just making stuff out of our database. So anyway one of the quirks of it is that it will print you can put cards and say here are the 40 cards, 30 cards, 60 cards, 100 cards. Here are the cards I want to print. We also can use the same process to print sealed decks or print decks for drafting.

We can do that as well. But anyway, the one little quirk of it is, if you use a name of a card that already exists, it won't pull your play test card, it'll pull the real name card, card that already exists. It won't pull your playtest card It'll pull the real name card because that prioritizes So what you know what happening is I need counter spell but some random person in their sandbox made a card called counter spell I don't want that so a priority it prioritizes real cards over what I will call fake cards or unofficial cards So if you accidentally name your playtest card with a name of a real card,

it won't print out your play desk card, it'll print the real card. And that happens a lot. There's a lot of magic cards. So I tend to give my cards silly names just so that won't happen. Like names that I know will never actually call the card. Because sometimes when you use real names

and it pops up, you're like, ah, damn it. Okay, so what happens then is you put the card, you put your cards into the database, you make your deck in deck builder, and you print your deck. Okay, oh, and one of the things before the play test is you find out how many people are making new decks. If not a lot of people are making new decks, you might make more than one copy of your deck. That way more people could play it. In exploratory, exploratory has somewhere between four and six people. Sometimes we'll

pull other people in for play tests, especially if we have an odd number of people. But let me get to the actual play test. Okay, now you've made your test deck and it's time to play. So there's a couple things about play. So in exploratory when we play with test decks, there are a couple things we try to do. One, you try to play as many decks as possible. The goal in the play test is not just to see your deck, it's to see everybody's deck. And so what happens is, normally you're playing in somebody, it's not odd in this type of play test to just keep playing the same person and you just keep changing decks. So partner doesn't change your opponent doesn't change but your the decks you play change and Ideally, it's nice if you can play with every deck

If not, it's nice at least if you could play with every deck or against every deck So at least the cards you need to see have been in the game that you are playing And a lot of that just depends how many decks were made Let's say it's a really busy day and a bunch of people made decks and a few people made more than one deck. Sometimes you just can't see them all. In those days, just try to play against things that you don't see. But if some days it's like, oh, only two people made it.

Okay. Everybody can play with both decks and see what they are. It's not odd, by the way, when you're making decks, some people will make multiple decks. Some people will take their mechanic and go, I'm going to make a black, red and a white blue version of this mechanic. So try out, you know, one's more control-ish and one's more aggro, whatever, whatever you're trying to try out. So the idea is you are playing with the cards.

First and foremost, in this type of play test, the goal is to have a good play test and test the cards. If you figure out in the middle of your first play test that a card is wrong, it is too cheap, it is too expensive, the timing doesn't work, like whatever, it's causing some problem. You there on the spot, not wait for the next game, right then, right there in the middle of the play test, change the card. You got a pen, you can write on the play test card. You will just physically change the card right away. And the idea is, look, once we figure out something's wrong, there's no need to further test that we got it. It's wrong. Change it. For example, I'm

not the best at card power level. So quite often I'll be playing somebody and someone goes, Oh, that's you made that too too cheap. And I'll add a manna to it or Oh, that's too expensive. I'll make it one cheaper or you know or anyway all sorts of changes can happen but the idea is as soon as you play with it you can change it and you can sense what's going on and then you can adapt it to what you're doing. It's not uncommon sometimes when you're playing with your deck and you go oh wow this doesn't work at all but wait a minute and then sometimes you will change the entire mechanic.

Let's say, you know, I'm playing bread and I'm like, oh, wow. You know, let's say for example, I tried a bunch of breads of summer one, summer two, summer three, and I play going, three takes forever. Okay, make them all one. You know, and the next time we play, they'll just all be one.

Or even mid game, I change them so they're all one. You know, whatever happens if something isn't good. So once again, let me remind you of the buckets that we use in Exploratory Design. What I like to call the good, the bad and the ugly. Good bucket means we like this mechanic, vision should look at it's good mechanic. Bad means it's the bad mechanic,

vision could save their time, don't worry about looking this, it sucks. And ugly means there's something about it we like, it's not in its right form. Oftentimes, if things go in the ugly bucket in your exploratory, you will further play around with the mechanic to try to get it to the good or bad buckets. Usually what happens is you'll try a couple times and then either it gets to the good, you're like, ah, let's move on to something else. Not that's not inherently a bad thing. Maybe someone else will crack it later.

But normally what will happen is vision design will start by looking at all the good stuff in the good bucket. And if they get desperate, they'll look at they'll look in the ugly bucket and see maybe there's something there with potential that they can fix. But anyway, so a lot of what you're doing is you're sort of getting a sense of where things lie. You'll play with your mechanic, you'll play with other people's mechanics. Now normally the way it'll work is we will play for 45 minutes of the hour.

Sometimes if games are taking longer, we might play the full hour. If we are able to finish 15 minutes, we'll have a conversation in the room. If we end up playing to time, then we will we have a thread where we can talk, the team can talk and we'll post a thread and say, Oh, give us the feedback of your games. Also, if somebody is play testing, and they're not part of the team, we often will get them to write their feedback, write down their feedback and give it to us because they're not there later on for nuance of feedback. So the idea essentially is okay, so we then have a little meeting about what do we feel

about this? So the most important thing that you're evaluating is what I call the fun center. Does it have a chewy fun center? Is there something inherent about this mechanic that is joyful? That is fun Is it just play? Well, is it super flavorful? Is it just do something splashy? Is there something about it that just makes you go? Wow, this is there's something about it. That's cool There's a bunch of ways to be cool because really what you're trying to find is is there something here that's like

Because really what you're trying to find is, is there something here that's like, our goal is to make things that really excite the players. And so we have to figure out like, is there anything here that could, no, we do have what we call workhorse mechanics. What a workhorse mechanic is, is, hey, it helps limited or constructed work, but it's, it's not exciting, you know, but it gets the job done. There's a lot of decisions, like cycling is a good example of sort of a workhorse mechanic. There's nothing inherently exciting. I mean, maybe the first time you saw it is exciting because

you've never seen it before. But you know, once you've seen it like, okay, it is a mechanic that does good work. It's valuable. It helps get more expensive cars into play. It can get more niche cards into play. There's a lot of cool things you can do with cycling. It's a neat mechanic, but it's, it's not especially, you know,

the eighth, ninth, 20th time we've used it. No one's, everyone knows what it is. It's not that it is something novel, but it gets the job done. So sometimes we make mechanics that are workhorse mechanics. We need workhorse mechanics. That is fine, but you want to balance any, you want to sort of get a sense of what is the, what's the optimal thing for a mechanic. Now note, if a mechanic does something that hasn't been done before, it has a little

bit of splash and novelty value, even something that would eventually be a workhorse mechanic, if it's the first time you've ever done it, oh well there's some splash there. A good example is a mass we did in War of the Spark. We had never really done that before. So the first time we ever did a mass, it was new, it was novel, we hadn't done that before. So in War of the Spark, there was some novelty to it. But we've since learned it does a really good job of having sort of a large group representing

an army in a way that plays well that doesn't have infinite tokens and so we use a mass now and a mass now is more of a workhorse mechanic. Players know what it is. Not that they dislike it. I'm not saying workhorse mechanics aren't fun. They play well and you can make cool new designs with them. So it's not as if you can't make a workhorse mechanic fun in the individual designs. But the idea is the core mechanic itself is not inherently going to excite people, mostly because it's either it's a known quantity or it's just messing in space. It's not inherently exciting. So you want to figure out whether or not your mechanic is there some fun, like is there something about, did you enjoy playing it?

Was it fun to play? Were there interesting decisions to make? Was there cool interactions? Were there neat synergies? Whether there's something about it that just made you go. Oh, this is a little bit different in some way Or it really played so nicely with other things or whatever Whatever the cool thing is you just want to make sure it's there if it's there That's how it gets into the good bucket, right? It's like, oh, okay, we did it. It plays well. The execution's nice. The ugly bucket is, yes, I see. I see the kernel of fun. But either we're not quite getting there, there's something in the way, or there's

some logistical thing, or memory issues, or there's a bunch of things that can get in the way. And so one of the things when we're identifying, we're talking about whether or not we think a mechanic has potential. get in the way. And so one of the things when we're identifying, we're talking about whether or not we think mechanic has potential. That's super important. And then what are the problems? And the big issue with the problems is, hey can we solve those? Like if the mechanic plays fun but, okay can we address the but? Can we, you know, it plays us but there's memory issues. Can we find a way to do with less memory issues?'s fun but really logistically complicated can we lessen the logistics is there a way to use a frame you know like we have a lot of tools

available to us to try to solve problems and the idea essentially is you only want to solve problems and something that shows promise if something doesn't show promise and and by the way we play test a lot of mechanics a lot of mechanics and most of them Do not advance to the good bucket And even the stuff that gets in the good bucket doesn't necessarily get into the file There's so you know that the the needs of the set will really dictate What things it uses and there will there could be very very good mechanics that just don't make sense for the set So they won't get used not a knock against the mechanics

Mechanics have to fit the need of what the set wants Okay, so we will do it will go around one by one and we'll give feedback on what we thought of the mechanics Normally the way we do it is we'll go Mechanic by mechanic or deck by deck and say okay who all played the baking deck? Okay what do you guys think of baking? And you know someone may say oh the flavor is pretty good and someone else might say well the biggest problem is you know losing a creature for a turn has real tempo utility issues and we have to figure out how to offset that. And then we can start talking about like, okay, but is it cool?

The fact that it has some real cost to it mean we can give you something more for less. Can we make it look exciting enough that the player goes, okay, I'm more than willing to have my creature go away for a turn. It comes back so much bigger. Can we do that? And the question is, is the things we need there? And that's when we walk through sort of the promise of the mechanic and the cool parts of the mechanic and the synergies and that.

And we also walk through where are the problems and do we see things we like or dislike? And then we sort of, you know, everybody will talk about the mechanic and then we'll sort of say, how do people feel about it? And once again, I should also stress, it's not, I keep talking about the mechanic and then we'll sort of say, how do people feel about it? And once again, I should also stress, it's not, I keep talking about the buckets. Fundamentally in the end, they go in buckets, but that's as much conceptualized. It's not like what bucket you put it in. It's not quite the way we talk. But inherent in there is understanding whether something shows promise and we want to do

more. Then if the mechanic has issues, we will sometimes spend a little bit of time, depending on how many things we have to talk through, we sometimes spend a little bit of time just brainstorming like, oh, what might we want to do with the mechanic? We like it, but it has this problem and we'll sort of walk through, okay, well, what can we do that will help solve this problem? You know, what can we do that will, you know, is there something else we can try that keeps what we like about it but lessens the problem of it? And we'll walk through that. And sometimes when we do that, you know, the person who made it or sometimes even a different person, someone made it,

we'll go, okay, I want to try that. You know you know we sort of say here's what I'm going to do just so we're sort of aware who's working on what normally by the way even in a brainstorming session sometimes if we have certain areas people will call dibs on things just so we know what different people are doing sometimes we don't sometimes people surprise us and it is possible for two people to bring a similar thing it doesn't actually happen all that much, just because when you're playing in a space, usually there's a bunch of different things.

And different designers will get excited by different things and go in different directions. And then once we're done talking it through, we will take notes from the meeting. Somebody in the meeting will take notes and they will post it to our team channel. The thing that's important there is let's say we flash forward to vision and vision is looking through our notes because they want to figure out you know what mechanics to play with and we'll often explore

for example give them a list of what we consider to be the good mechanics but they might want to go back and look at our play test and go oh what what what was their issues at the time let's's see what the notes are. Um, and so just having the notes is valuable so that we can go back and that we can look at it. Um, and then after we're done, we shred the test decks, then we get rid of them. We have a little shredder, um, cause it's, you know, it's private information. We don't want people seeing them. And so we will, we will shred the deck. There's little, um, like, uh, sort of trash cans that you can put things in, although they have

a little slit. You can't take them out once you put them in. And then those will get destroyed. I, for those who have ever seen a picture of my desk, tend to throw them on my desk and go, I'll get rid of them later. And then my desk has like 8,000 test decks on it. So test decks are most often used in exploratory design. Like I said, before we have a limited file,

sometimes they're used in vision or early set design just because you have a mechanic you're interested in and you want to get a sense of it. And it is something that might not really show up in limited. There are definitely mechanics that either they're high as fan or they're just the nature that either they're high as fan or they're just the nature of what you need to make them work is something that limited has trouble with.

So normally once you have a file, you're play testing most things in some form of a limited form. But sometimes you need to test something you can't test it there. So sometimes you'll make test decks to do that. Once you start getting to later in the process, you're actually making actual play test deck with real cards and matching the format and stuff.

You don't tend to use test decks late in the process. They tend to be used earlier. Every once in a while in set design, like later set design, they need to swap out a mechanic. They're testing a new mechanic and sometimes they'll use them there. So anyway, guys, that is me going deep on test decks what they are how you build them It is definitely I built a lot of test decks and they're built. Oh somebody didn't explain it doesn't take that long You got to build four to six cards usually because your mechanic and you double the cards a lot of times You can make a test deck in half an hour if you need to that's going a little fast you really have to know what you're doing and once you make your cards you have to input them

and then you got to make the thing and then you got to print them so half hours about the quickest you can do it sometimes takes a little longer than that. Normally if you have a play test coming up you'll like you know when your play test is coming up and try to give yourself enough time to do that part of the homework is making the test decks. Anyway guys that is probably everything you ever wanted to know about test decks, since most of you probably didn't know what they were before we started. But I'm a thousand plus in, so I'm getting some crunchy topics.

Anyway guys, I'm now at work, so we all know what that means. It means the end of my drive to work. So instead of talking magic, it's time for me to be making magic. I'll talk to you all next time. Bye bye.