[Discussion] Goldfishing Control Decks

[DrJones]

Hi! This will sound weird, but I'm interested to know if you have ever goldfished a control deck, and if so, which things do you think are important to track during those tests? I would also like to hear if you have any trick for knowing when the deck is not behaving good enough (with no clues from an actual opposing deck).

Also, what do you think about this question? is there a benefit in goldfishing a deck which has no clock? should I test only real matches?

Thanks.

[Bardo]

Also, what do you think about this question? is there a benefit in goldfishing a deck which has no clock? should I test only real matches?

I "goldfish" the hell out of all of my decks before playing them against other people. But I goldfish, I call it "testing," my control decks the most. I mainly want to get a feel for how the deck flows and what needs tuning. I've been doing this for awhile and have a general method:

Stage 1. The Mana.

How often do I have four land on turn 4?
Can I play my cards when I want to?
How often are cards stranded in my hand?
Is it because of colorless land? Wrong colored mana? For instance, how reliably can I play Pernicious Deed on turn 3 in a blue-based deck?
Generally: How often does it flood (too much mana)? How often is it screwed (not enough mana)? And why?

Stage 2. The Curve.

Basically, how often am I shipping a turn with nothing to do?
How often is my hand clogged with high cc stuff?
Is there enough card draw that I can stay ahead?
Is it the right "kind" of draw; say FoF vs. AK.

By this point I'll know if a deck is objectively playable or trash.

If I keep working with it though, the manabase has gone back to the shop numerous times and has gone through many tweaks and experiments. Colors are balanced, along with fetchlands, colored mana sources and colorless mana sources. I've adjusted the maindeck spells to have a comfortable and playable curve.

It's pretty much impossible to tell how quickly a control deck is going to win in the abstract, so I don't fuss with that. Anyway, I don't think you can adequately balance your mana and curve when people are trying to beat you, so I think you need to do all of this alone.

Stage 3. Imaginary opponents.

Once I think the deck is vaguely playable, I imagine playing against Goblins with a turn structure and see how often I have an answer to a first turn Lackey and Vial. A turn 2 Piledriver, turn 3 Warchief, etc. Then I'll go through other exercises imagining that my opponent is running Daze. And so on.

I think this is all critical for the development of playable tournament decks. If you take a brand new control deck and test it against tuned decks and real opponents without fixing the fundamentals (curve and mana balance), you're going to lose a lot, get discouraged and throw away an otherwise very good deck, if only you'd fixed some basic problems beforehand.

[umbowta]

Very nice reply Bardo. Oftentimes I've wondered just how in the hell do you goldfish a control deck. I've done many of the things you said but never really tried to put in into an actual structured procedure.

[mattkru]

Imagine you play against Goblins. You make a List from 1 to 6.

For example:
1. Goblin Matron
2. Goblin Lackey
3. Goblin Warchief
4. Goblin Piledriver
5. Siege-Gang Commander
6. Goblin Ringleader

Than you goldfish and in every opposing turn you roll a dice (w6). Than that card is played and you decide, if you counter it or not.

For all the other relevant decks you also make a list from 1 to 6. Or from 1 to 10 if you use a w10.

[DrJones]

Imagine you play against Goblins. You make a List from 1 to 6.

For example:
1. Goblin Matron
2. Goblin Lackey
3. Goblin Warchief
4. Goblin Piledriver
5. Siege-Gang Commander
6. Goblin Ringleader

Than you goldfish and in every opposing turn you roll a dice (w6). Than that card is played and you decide, if you counter it or not.

For all the other relevant decks you also make a list from 1 to 6. Or from 1 to 10 if you use a w10.That idea looks promising, but it looks like it requires more polishing. 1st turn Siege-Gang Commander? Ack! :laugh:

I'm very interested in learn any trick on this topic because I've not found any article about it yet, and because my control decks look less solid than my other decks, which I suspect is due to not goldfishing them properly. I also fear that testing them only against specific decks (instead of abstract concepts like the fundamental turn) will make them too dependant on the meta game.

[Happy Gilmore]

If I'm goldfishing out of bordom I try and simulate having to answer one threat a turn. Its a hard thing for most decks to do, and it lets you know how deep your resources go.

[Bardo]

Imagine you play against Goblins. You make a List from 1 to 6.

For example:
1. Goblin Matron
2. Goblin Lackey
3. Goblin Warchief
4. Goblin Piledriver
5. Siege-Gang Commander
6. Goblin Ringleader

Than you goldfish and in every opposing turn you roll a dice (w6). Than that card is played and you decide, if you counter it or not.

For all the other relevant decks you also make a list from 1 to 6. Or from 1 to 10 if you use a w10.

Nah, no reason to make it so complicated. The right way to play the "head game" is to imagine your opponent playing the cards you least want them to play that turn and then seeing if you don't get wrecked.

[Anusien]

I always do this. It also helps to take a sample hand (like 2 lands and a Brainstorm) and see how often it works out. You can also work out the math to see how often the hand will pay off.

[Iranon]

I think that Goblin example is a little unrealistic and too complicated. I do something similar with a more generic Aggro package, playing against an imaginary Zoo opponent able to cast 3 Kird Apes, 1 Bolt, 1 Watchwolf during their first 3 turns, then drawing nothing relevant for a while. Around turn 5, I check whether I'm still alive, winning or losing and what I'm scared of at this point (being burned out or a renewed creature offense... this is a simple and generic exercise).

[FoolofaTook]

I do a lot of goldfishing of control-oriented decks and the one step I'd add to Bardo's analysis is the worst-case scenario step. I like to take the theoretical matchups and assume god draws for them and then see how often I'm still viable on turn 3.

This gives me a feel for how likely the deck is to go out to a bad matchup randomly without recourse.

Of course most of my experience is now ancient in single-elimination formats so I may overestimate the effect of a bad matchup in a swiss system.

[Maveric78f]

I simply open 2 instances of MWS and I run each deck in each window. I play the best possible with both decks, keeping in mind what should know the player (for cabal or meddling mage it can be a problem for instance). I do it 10 or 20 times preside then that many postside.

This kind of play has for benefit to make you know exactly the MU as you play both sides. You finally know what is lethal. If I want to make statistics, I have to start over the testing (the first ones are necessary to handle well the MU and to know what to side in).