one million words's picture
By: one million words, Pete Jahn
Oct 15 2007 5:00pm
Login or register to post comments

How to Playtest

Let’s assume you have a solid idea for a new deck, and you want to win some packs with it in a PE or the eight man queues. Going straight from the drawing board – okay, the deck design screen – to the PE is rarely successful. You need to playtest first. You need to playtest a lot – and in the right way. That way, by the time you start paying tickets to win packs, you will know not only how to play your deck, but how to beat the opposition. 
The Team
You can build decks and playtest alone, but it is hard. A team helps.  A team lets you share the work, and provides valuable sounding boards for your ideas, as well as contributing their own. If at all possible, find some other people to team up with, and work together. Even one other person helps, since playing someone else is far easier than playing yourself. That is especially true in MTGO, when the one account rule is in force. 
A team is useful to share the work – including building decks. It is tempting to have each team member specialize in a particular deck or archetype, usually the one they like and know best, but be wary of that approach. In the beginning, you want everyone to be playing every kind of deck. Even if you never intend to play control decks. for example, in a tournament, if you have playtested them a lot, you will understand how they work. You need that understanding before you can learn how to beat them. You will need to find the chinks in their armor and attack those chinks – and you will know how vulnerable each chink really is only if you have played the deck. 
The other advantage of a team – or better yet a good clan – is that they may be able to loan you cards. It is tough enough to get the cards together to build one tier one deck. Getting enough to build decks to playtest against is even tougher. However, often your team can loan cards back and forth so each tester can play complete decks.
Getting cards for online testing is tough – far tougher than paper. You see, paper players cheat. They use proxies. Sometimes the proxies just represent certain cards – sometimes they represent whole decks. Here are two sets of proxies I used a while back when working on a series of articles on the best Extended decks ever.   I had to create thirty two decks – and I played all of those decks against each other. Since I do not own, for example, twenty copies of Force of Will, I used proxies. The proxy on the left has four different decks: if I wanted to play Turboland, for example, I would use the green names, and the shown card would be Time Warp.   For the others, I used a lettered and numbered grid, and I would write the cards being played onto the grid.
Anyway, you cannot play proxies on MTGO. That’s just the way it is. That makes having big collections, having friends with big collections and having a clan with cards to lend all the more important. However, nothing beats simply owning the cards. 
My suggestion: Simply go to and order everything you need to get complete playsets of every card in the format in which you intend to compete. It’s easy.
No, I can’t afford it, either – but I’m working on it.
(Editor's Note:  Of course, you should order from, the customer service is top notch, and the prices are very competitive, but you can off set the costs of MTGO if you pick up a gig, maybe writing for
I really wish I could get a digital Sharpie, and turn those spare commons into proxies.
Where to Play
Of course, having a clan or playtest team available 24-7 would be perfect, but that is almost as unlikely as having a playset of everything. Many times, you end up having to look for opponents. Fortunately, you have many options.
The Casual Play: New Players Room
Okay, just stay away from this one. The only players in the new player’s room are 1) new players and 2) total losers who cannot win anywhere else.   You will learn absolutely nothing from playtesting here. If you play you blue black control deck against a new player with a total of one hundred cards in his/her collection, you will win. Great – you will never play anything like that deck in a tournament. Alternatively, if you play someone with a complete netdeck in this room, the odds are that he will be so bad that, if anything, you will absorb misinformation (e.g. there’s no need to play around tricks - like Mystical Teachings for Slaughter Pact in response to a Griffin Guide - because these opponents don’t know how to perform them.)
Casual Play
This is a step up from the new players room, but it is still pretty much useless for testing. The players in this room generally have a mix of partly completed net decks, interesting ideas and random (from a tournament perspective, at least) garbage. This is a great place to play if you have a cute deck idea, or want to play a bad deck for whatever reason, but there is no point to bringing tournament caliber decks into this room. Once again, you don’t learn anything. Imagine if the Green Bay Packers wanted to work on their running game – would they learn anything in a scrimmage against the North Hickfield middle school JV team? 
Anything Goes
This is better, but it has, in my experience, two main drawbacks. First of all, it has too few players. I have often wasted a lot of time waiting for a game here. Secondly, the decks in this room are all over the board. Once in a while you can find serious and relevant competition, but far too often you just find weird stuff. 
Tournament Practice
This is better – and probably the best option for those times when you can’t meet up with a teammate or playtest partner. The tournament practice room has the highest proportion of tournament-caliber decks. The room also has plenty of players and matches, so you don’t have to wait long. Best of all, it provides sideboard games, so you can test maindeck and sideboard. The only downside is that the opposing decks are somewhat random. For example, if you are testing a new-standard red green aggro deck, and want to play it against blue black Coalition control, you may play against a half dozen aggro decks before you find someone with a blue black deck. Advertising what you want can help, but many people don’t read those notes, and others stay away, assuming you have built something random that beats UB, but folds to anything else.
Eight Man Queues
In the eight man queues you will get a consistently high level of competition. You will also play sideboarded games. However, for the purposes of playtesting, this may not be the best option. First, you have little control over what decks your opponents are playing. Second, if you lose, you are out, and that can make playtesting here really expensive.   In that respect, PEs are a bit better – you at least get to play more rounds, and if and when top eight playoffs vanish, you can probably get to play every match, regardless of your results to that point. 

The other disadvantage of playing a half-tuned deck in these areas is that your future opponents will see the deck, and may begin to adopt it, or work on ways to beat it. Your best results always come when you can tune a new deck in secret, then surprise people with it. That’s how you win packs.
The best option for playtest is to meet your playtest partners in some random room. Simply type /join {room name} in the message bar in any main room, and you will go to a room with that room name. If you chose something random, like /join grq178bsw, the odds of anyone randomly stumbling into the room are miniscule. Just remember to let your playtest partners know the room name. 
Using a private room has some big advantages over using the Tournament practice room. The first is that you can set up games by challenging people, but you can make the games watchable without worrying about others learning about your secret tech. You can also write comments in the room screen, where they can be read by everyone in the playtest, without them being a) public and b) drowned in spam, questions and game requests. You can also do this in the game itself, but having them in the room dialog makes them easier to find once the game ends. 
The best reason for playing in a private room is that you can do rational playtesting.
Rational Playtesting
The first thing to realize about playtesting is that you are not playing to win – you are playtesting to learn how to play certain matches. When playing to win, you hope your opponent will make a mistake. In playtesting, you hope they will not. In fact, in paper playtesting, we allow takebacks of any mistakes, and even of any questionable plays. Winning a game because someone forgets to pay for a Pact is fine in a tournament, but not if you are trying to decide whether Green White aggro can really beat Blue Black Control.
In MTGO, you can’t allow takebacks, but you can discuss plays before they happen. In playtesting, I have often told my opponent (and spectators) “I have two removal spells in hand, but can only cast one with the mana available. Do I kill the blocker and swing, or save the instant to kill in response to pump.” Sure, I am giving away info on my hand and options, but the point is that I want to learn which option is better. After some discussion, we play it out – and in the paper world, we may rewind and try the other options, too. 
If you have several playtest partners in your room, you can invite some of them into your room, then discuss it with them. MTGO makes it easy to type comments to some players and not to others. 
Ten Game Sets
When doing serious playtesting, we play ten single games, alternating who plays first. We don’t count any games where a player has to mulligan too many times (which usually means we concede and restart any games with two mulligans for regular decks, and three to four mulligans for combo decks.) Alternating play or draw means that we get realistic results. For example, some super-aggro decks can only beat control on the play – and only if they drop a 2/2 in turn one. Alternating play or draw makes sure you learn about this early. 
We play ten single games because we want to learn how the match plays out, in some detail. After that, we work on sideboarding. Generally, at this point I have saved the main deck, without sideboard. I then discuss sideboard ideas with my opponent. After ten games, or more, we should have a decent idea of how each deck tries to win, and what it’s vulnerabilities are. That should give us sideboard ideas – and we then work on incorporating all of them that we can possibly fit. I’ll make those changes, then save the sideboarded deck under a different name.
For example, let’s assume I am testing red green aggro against UBrg Coalition Relic. I may determine that the biggest weakness the control deck has is it’s mana base. I want to attack that. I want to bring in Ancient Grudge (to kill the Coalition Relics and Prismatic Lenses – and I may even want some (Tin-Street Hooligans). I want to play four copies Magus of the Moon. I may want some Sudden Shocks or even Word of Seizing. It’s obvious that I cannot fit all of that into a normal sideboard, but that does not matter. For the first few sets of sideboard playtest games, I am going to create a new deck with as many of those cards in the maindeck as I can squeeze. Then I’ll save that deck as RG Aggro with UBgr sideboarding, then play that deck for ten single games against the sideboarded UBrg deck. 
The purpose is to see what works, and what does not. After the ten game set, I may discover that some of these cards do nothing, and I’ll cut them back out for the next ten game set. Eventually, of course, I will have to cut back my sideboard to 15 cards, but I want to have more information by then.
When you are playing these games, take notes. Logging out of MTGO wipes your games, and while you can save that log, it’s a pain. It is also a pain to watch the games looking for important info, so take notes. Note what problems you had, and what won. That makes final tweaking and balancing easier, and may help you remember what was important a week, and another hundred playtest games, later. 
The Gauntlet
When I am testing a new deck, I want to test it against a gauntlet of typical Tier One decks. For example, in testing for the upcoming States tournament (paper, held the last weekend in October, with Lorwyn legal and Ravnica block out) I am testing against GW & RG Aggro decks, UWx and UBx control decks, a WB deck build around Haakon, Stromgald Scourge and possibly a couple others. I know of some other deck and archetypes, but they are not in the gauntlet. My pet decks are definitely not in the gauntlet.
The gauntlet is supposed to represent those decks I expect to face. I generally get my gauntlet decks from the forums on major websites, like StarCity Games. I am not looking for teched out versions – I am looking for generic or stock versions. The purpose is to get experience against the decks I am most likely to play against at the tournament. For that reason, I also look closely at any builds posted by major writers, like Mike Flores, Frank Karsten or the StarCity premium crew. Those decklists are very likely to be copied, complete, by netdeckers.   That is what I want to know how to beat. 
Playtesting also takes a long time. I don’t want to waste it playing against marginal decks that I am unlikely to face in an actual event. In the rare case that I have a new and unique deck, I certainly don’t want to waste time initially playing mirror matches. If this is new tech, then the only way I would face a mirror is if both I and my teammate entered the same event. In that case, if we get paired up, we can draw. After we have won an event or two, then we can expect to see netdeckers and mirror matches, and at that point you can work on the mirror.
Tweaking your Deck
Once you have built the basic deck and played it – sideboarded and unsideboarded – against the field, you need to fine tune it. For me, this generally means cutting the deck down to sixty cards, and the sideboard to 15.
Very often, I end up with two cards that are vying for the last slot in the deck, but I’m unsure which is better. For example, assume that, in my RG Aggro deck, I’m trying to decide between a fourth Seal of Fire and a fourth Rift Bolt. Both are good, in certain circumstances. When this happens, I choose one, then playtest for a while and see how often I draw the card and am happy, and how often I draw it and wish it were the other card. I take notes. In paper magic, I mark the card or sleeve, so I know when I draw the fourth copy (marking the proxie works, as does putting a dot on the sleeve face – I just have to change the sleeve before the tournament, to avoid marked card penalties.) 
In MTGO, you cannot mark the sleeves or use proxies, but you can use foils, or (where possible) other versions. For example, I can play three normal Seals of Fire (the three that are in the deck regardless), and one foil Seal of Fire to represent the card in question. Now, if I draw the foil Seal, I can consider whether a Rift Bolt would be better at that moment. Alternatively, if I was debating playing a fourth Llanowar Elf, I could play three Tenth Edition copies and one Seventh Edition version. This approach can be problematic when the card in question is a chase rare and the foil costs a bazillion TIX, but it is worth considering for commons and so forth. Dropping ninety cents or so on a Foil Rift Bolt might be worth it when you plan on investing a bunch of TIX in tournament entry fees, in hopes of winning a bunch of packs. Besides, you will have a cool foil in any case – and when you play in the tournament, you can cut a non-foil and play the foil, if that’s where you end up.
Cutting the sideboard down to 15 cards is, often, ever harder.   I almost always have thirty or more cards I want to include, and 15 slots. Here are a few thoughts.
First, if you have matches that you just win, even without sideboarding, then don’t include sideboard slots to that matchup. If you already win, sideboard cards often just mean you win more (i.e. pushing a 80-20 matchup to 90-10.)  That’s unnecessary, unless you have nothing better to include. Note, however, that if the other side will be sideboarding something you cannot beat, you may need a sideboard card to deal with that. For example, some Ghazi-Glare decks used to destroy red decks, unless the red decks had Fortune Thief. However, if decks you beat already do not have unstoppable silver bullets, ignore them. You can find better uses for your sideboard space.
Second, make compromises. If you face some decks that have important artifacts, some critical lands and others enchantments that you have to kill, look hard at multi-purpose cards like Creeping Mold. It is slower, but nails all three. Test it. If it is good enough, it can take the place of a Stone Rain, an Ancient Grudge and a Demystify.
Third, if you have a matchup that you cannot beat without thirteen cards from the sideboard, and even then the matchup is still a truly uphill struggle, then simply give it up as a bad job and move on. Sometimes your only hope is to hope you can avoid the matchup. Obviously, this only works if the deck is not super-common, but it does work when the format is varied. Even when certain decks were dominant, some decks managed to T8 simply by not ever being paired with the dominant deck, and preying on the decks tuned to beat that dominant deck. It happens. 
Once you have settled on your 75 cards, it does not hurt to playtest some more – no matter how good and knowledgeable you are, you can always be better. At some point, though, you need to take your deck into the 8-mans or a PE or two. Then, once you win a few games and people start noticing your deck, then the metagame will change and you can start playtesting again.
“one million words” on MTGO


Best article I've seen in ages by runeliger at Tue, 10/16/2007 - 07:43
runeliger's picture

Literally hands down the most informative article on PureMTGO I've ever read.

 However, I'd like the mention for the gauntlet testing, MAKE SURE YOU HAVE DECENT PLAYTEST PARTNERS PILOTING THE DECK! I've seen countless times where people disregard a matchup as being "70/30 and 60/40" (meaning basically in their favor), but that was a result of their problems with who they're playtesting against. Things happen like this all the time. The first two examples that come to mind is the belief by some people who play classic that Threshold has a great matchup against the Red burn deck (FYI, it doesn't) and a PTQer of the previous block constructed who believes that Pickles destroys UB teachings hands down. Both misconceptions come from a lack of a good partner. The classic players obviously are rather good at thershold, but don't have a good red burn player to test against, and the PTQer who pilots Pickles hasn't faced a good UB teachings player in their playtesting.

Ideally, you want to test against Omnichord with LSV, Enduring Ideal against Andre Mueller, GW aggro against PV, but that's not happening, so you need to find the next best thing. Make sure that your playtest partners are not only good friends who will help with the process, but also are great players too.

 Oh also, being involved in a group where your cardpools can be shared is priceless (especially if you think Orim's chant would be tech in your new deck...)

Abso-friggin-lutely Amazing! by tempesteye at Tue, 10/16/2007 - 15:47
tempesteye's picture

I can only echo every one elses comments; That was prolly one of the bedst damn MTGO articles I have ever read, across all the major websites.

I don't know what Heath is paying you but it's not enough. That was some serious quality writing.




Random Rooms by Me5794 (Unregistered) (not verified) at Tue, 10/16/2007 - 10:54
Me5794 (Unregistered)'s picture

When you challenge someone in any "/Join <insertrandomroomhere>" , even if you are on the main splash screen, your matches will be watchable from the "Anything goes" room from my experiences. Also if you are the challenger and are in any other room your matches will be viewable from there.

 Honestly, if you are afraid of letting secret tech out,  your best bet is just to go to the Anything goes room and make your matches unwatchable.

by hamtastic at Tue, 10/16/2007 - 10:21
hamtastic's picture

Great job, and I agree completely with runeliger, competent playtest partners is key.  A lot of misconceptions about certain matchups stem from incorrectly testing them.  This is where your statement of backing up play mistakes comes in handy in paper, but isn't nearly as easy in MTGO.  In MTGO you really want to find a good player for the deck you want to play test as that will give you the most accurate results. 

by SpikeBoyM at Mon, 10/15/2007 - 22:08
SpikeBoyM's picture

Great article!  Something that all players of any level could really use.  Bonus points for the Packers reference. 

I  think the part about gauntleting is very important.  So often in PDC, you see people bringing whatever they feel like regardless of current metagame trends.  Additionally, some players will so heavily metagame against a good match up that they weaken their sideboards against their poor matchups.  Great job of bringing this information to the MTGO audience.