Congratulations! A bolt from the blue has struck and left in its wake a great idea for a new deck. It's time now to get turn that idea into reality! Of course, as any player knows, that's only the start. Once built, a deck needs to be tested. This begs the question, how do we actually test a new deck? Most players would answer with "Playtest it, of course."

Which is fair. That's a necessary and invaluable part of deck building at any level. However, what most players actually mean by playtesting is to grab someone to jam games with. Which, again, is great. Magic is a card game after all. However, just playing games isn't really playtesting. In fact, jumping right to games is almost certainly the wrong move. Players at all levels and in all formats tend to skip an important step before playing games.

Don't Neglect to Goldfish

This may be shocking to some players, but goldfishing is an essential part of deckbuilding and testing. Yes, it doesn't replicate games and so doesn't replicate the actual play experience, but it's not supposed to. Goldfishing does serve an important purpose though. If playtesting is learning how a deck performs in a match, then goldfishing is learning if a deck works at all and how to actually play it. Skipping goldfishing is the equivalent of taking an idea right off the drawing board to a full product release. It might work, but it's more likely to be an avoidable disaster.

To make sure that everyone is on the same page here, goldfishing is where a player plays a game without an actual opponent. The name (probably) comes from asking if a deck can beat a goldfish, an opponent who does actual nothing. Being able to beat said goldfish is the basic test of competency for any Magic deck, regardless of competitiveness or format.

Ideas Must Be Tested

It is possible that some players can build a deck correctly on the first try. I've neither met one nor heard of one. I'm not one either. I, therefore, feel confident in saying that every newly built deck is an unproven deck. Theorycrafting a deck and building to the theory is great, but when theory and reality conflict, reality always wins. Goldfishing is how to find out if the ideas that went into the deck actually work. Goldfishing reveals answers to these questions:

1. Does the mana work? Is this deck capable of playing Magic, and what do I need to do to make it happen?
2. Does the deck do The Thing? Whatever The Thing might be if a deck can't do what it's supposed to do against no opponent, how will it do so against actual resistance?
3. How quickly does the deck do The Thing? Is it as fast as it's supposed to be?
4. Do I know how to do The Thing? Do I know what I'm doing?
5. Is The Thing worthwhile to do? Is this idea dead on arrival because it just isn't very good?

Make no mistake, goldfishing will not provide definitive answers to these questions. That's playtesting's job. Actual games with actual interaction will impact a deck's ability to function. However, goldfishing will reveal the larger flaws in a deck which will make the later playtesting more productive.

#1: Does the Mana Work?

There are all sorts of guides out there for how to build mana bases in all formats. Many players have their own heuristics and guides based on their experiences and preferences. However, none of that matters for a given deck. It may be trying to cast spells that make atypical demands. Or there may be heavy color requirements that force concessions. A deck whose mana cannot cast its spells cannot function.

As an example, I am spending most of my playtest time working on Modern Merfolk. After I built my first test deck, I started running into a problem casting my spells. Specifically, I had a lot of double blue spells, and Mutavault coupled with Cavern of Souls made it hard to cast them. Cavern was fine for the merfolk themselves, but it was very awkward with sideboard cards. Mutavault was worse and made me mulligan more than I thought necessary. I replaced the 'Vaults with Islands and the problem disappeared. Doing so was heresy once, and maybe playtesting will say it's wrong, but the goldfish test indicates that it's not.

This is also the chance to learn not just if the mana works but how it works. Is there a specific order that lands must be fetched and will I lose if I can't or don't? How important land sequencing is and how many lands the deck needs in play to function are just as important as what lands are in the deck. It's better to figure out how to play the lands when there's no pressure than trying to figure out during a game.

#2: Does the Deck do The Thing?

A deck is built to do The Thing. It doesn't matter what The Thing is, but it does matter whether or not the deck can do what it's been built to do. This can be archetype specific. For example, with combo decks, does that combo come together on purpose? Can an aggro deck present a decent clock? Do the synergies I think work actually work in the deck? If the answer to these questions when playing against no resistance is anything other than a definitive yes, the deck needs more work.

Control decks can struggle with this one, but I have a solution. How often and how does it need to interact against a given deck? Imagine that scenario and see what happens. Play the control deck against an imaginary opponent. To avoid cheating, before the game begins write an opposing sequence that will win the game and try to stop it. If a control deck struggles to actually present the interaction necessary to survive a theoretical deck, it can't possibly do so against a real one consistently.

Also, this is a chance to find out if the deck does any other Thing. There may be unintended synergies in the deck or alternative ways to win. Goldfish the deck to see what it can do, not what it will do in a game.

#3: How Quickly Does the Deck do The Thing?

Does my combo deck win quickly enough to not die to aggro? Will this aggro deck kill quickly? Kill speed is the quintessential reason to goldfish a deck. It's why players started goldfishing decks in the first place. If an active deck can't win in the timeframe their chosen format demands, they're not going to be viable. For competitive formats, this would be a kill on turns 3-5.

Again, for control decks, this question is hard to answer. Control's whole deal is slowing down the game and ending the game after it's already won. Just as with #2, contriving imaginary scenarios is useful here. Against a purely hypothetical opponent, can the control deck actually establish control, turn the corner, and proceed toward winning the game.

#4: Do I Know How to Do The Thing?

Can I play my deck? Not just make normal Magic actions, but carry out my deck's plan. Can I execute the combo or string together 20 damage? I built a deck filled with internal synergies, but can I get them together and maximize their effect? This is a very basic question, but for a surprising number of players, the answer is no.

For example, Steve Sadin won Grand Prix Columbus in 2007, but didn't actually know how his deck worked until most of the way through Day One. He was playing the Aether Flash/Protean Hulk combo, and expected that when he Flashed in Hulk, his opponents would concede. Which happened the first four rounds. Then, in Round 5 his opponent made him actually play out the combo, which he'd never done before. Steve didn't know what to do, messed up, and lost the match.

Had Steve actually tried out the deck before the GP, he would know how the combo worked and how to execute it. In order to learn how to play a deck, we have to actually play it. Goldfishing doesn't reveal how to play every scenario, but at least we'll know how it works. I resent having to spell this out, but I've seen players, even in the past week, lose matches because they didn't understand how to win with their deck beside the opponent conceding. Don't let that be you.

#5: Is The Thing Worthwhile to Do?

This applies more to competitive than casual or Commander decks, but is this new deck a worse version of an existing one? Worse is relative, but if the new aggro deck we've made is similar to but slower than Burn with no compensating value like disruption, why bother? Just play Burn. Playing a worse version of an existing deck is not going to win a tournament.

Even for non-competitive play, just because something seems fun, in theory, doesn't mean it actually will be fun in practice. If the deck does the thing it's supposed to, but we'd rather it hadn't, that's a strong sign something's wrong. Better to find out before subjecting ourselves and others to something nobody will enjoy.

A Bonus

As a bonus, goldfishing a deck will make us faster at playing the deck. I noted in my full article on the subject, that practice is the primary way to speed up. Learning how the deck works in a vacuum makes it easier and therefore faster to make those sequencing decisions that aren't dependent on the matchup. Land sequencing in particular should be rote for most decks, so just get the practice in.

Additionally, goldfishing combo decks will make it easier to execute the combo. A new player needs to think through the whole combo. An experienced player knows what needs to happen without having to think about it, which means they shave time off the combo.

Now to Playtest

Having goldfished the deck enough to have worked out the really rough edges and figured out how to play the deck reasonably competently, it's time to actually playtest. After all, it's the actually fun part. Playing games of Magic is the only way to learn and get better at playing Magic. Without playtesting, there's no way to know how a deck will perform in actual games. It's also the only way to learn matchups.

However, be careful. There are many stories of players and entire pro-level teams' playtesting so heavily that their testing became recursive. They convinced themselves that the matchup worked one way based on their testing, but their opponents had gone a completely different direction.

The key in playtesting is to play enough to understand how a matchup works and how to win. Play matches with the starting decks, make some tweaks as needed, but primarily learn what matters in the game. That's far more important than specific details about interactions or how to beat the specific build. Every player will have a slightly different configuration. Preparing too much against a certain version will leave a deck hopelessly inbred and weak.

Seek Answers

The other thing to do during playtesting is to test out various scenarios. If something isn't coming up in normal games, but it could be important, then set it up and find out. For example, one of the hardest things to determine is whether sideboard cards actually work. They're played in small numbers for specific scenarios, so there's just less opportunity to find out if they do what they're supposed to do. Therefore, set up both decks in the best-case scenario for that card and see if it does what's intended. If not, abandon it. If it does work, then start working back to a more typical scenario and see if it's still effective.

I know that there will be objections to this on the basis that it isn't a real-world test. So what? Science is all about contriving scenarios that are rare in order to see what happens. Don't get caught up splitting hairs about it being artificial. If there are questions about something in a matchup, answer them. It's better to have the answers than continuously worrying about them.

One Is Good, Two Is Better

Goldfishing and playtesting are both necessary to making good Magic decks, regardless of format. To do one is helpful but doing both is critical. Think of it as the equivalent of lab testing and field testing. Both reveal weaknesses and flaws, but different ones. Doing one without the other is leaving a huge blind spot.