Playgroup.gg logo
Playgroup.gg EDH Tracker
Playgroup Live · Solo playtest

How Sparring Works

Sparring gives a solo playtest phantom opponents that fire real interaction: counters, removal, and board wipes. The pressure is not random. It comes from 21,480 interaction spells measured across 6,934 tracked Commander games, so a solo goldfish finally reacts the way a real table does.

Updated 10 Jul 2026 · from real tracked games
21,480
Interaction casts modeled
Games analyzed
6,934

Finished Commander games behind the archetype mix.

Interaction casts
21,480

Real counters, removal, wipes, bounce, and edicts measured.

Wrath board-read
2.5x

More likely to eat a board wipe with 3+ creatures out.

The short version

A goldfish that finally fights back

Solo playtesting is goldfishing: no resistance, so it never trains the one Commander skill that decides real games, playing around interaction. Sparring adds phantom opponents that occasionally answer you. They counter a spell as you cast it, remove a threat when it lands, or wrath the board when you pass with a wide battlefield.

Every reaction explains why it fired, and the whole model is on this page. Probabilities come from three lookup tables measured from real Playgroup games: what interaction gets cast, when in the game it peaks, and how strongly wipes track board width. There is no rules engine and no AI. The believability is in the data.

What fires

The interaction mix

Across 21,480 interaction casts, spot removal dominates and edicts are rare. Sparring draws each phantom's answer from this same distribution, so you meet removal far more often than a board wipe, exactly like a real pod.

Share of all interaction cast

Percent of every interaction spell measured, by archetype.

58%
19%
13%
7%
3%
Spot removal
Counter
Board wipe
Bounce
Edict
Bar height is each archetype's share of all interaction.
When it fires

Timing curves by round

Interaction is quiet for the first few rounds and peaks in the mid game. Sparring uses each archetype's own curve, so the opening rounds feel safe and the pressure ramps as the game develops. Board wipes peak latest of all.

Spot removal

Share of this archetype's casts by round.

1
2
3
4
5
6
7
8
9
10
11
12+
Round number along the bottom. Last bar is round 12 and later.
Counter

Share of this archetype's casts by round.

1
2
3
4
5
6
7
8
9
10
11
12+
Round number along the bottom. Last bar is round 12 and later.
Board wipe

Share of this archetype's casts by round.

1
2
3
4
5
6
7
8
9
10
11
12+
Round number along the bottom. Last bar is round 12 and later.
The wrath board-read

Wide boards eat board wipes

The single clearest signal in the data: board wipes track how many creatures you have out. Pooled across the game, a wipe was about 2.5 times more likely to be cast against a player holding three or more creatures (49 percent of wipes) than the 20 percent baseline at zero or one creature. From 606 real wipe casts.

The signal first appears around round 5, so Sparring never wraths an empty board or an early game. Tokens were not counted in the offline analysis, so the real effect is at least this strong. That is why over-committing into an open board is the mistake Sparring is built to punish.

How much more likely a wipe fires, by your board width
Creatures you control Relative wipe likelihood
0 to 1 creatures 1.0x
2 creatures 1.3x
3 or more creatures 2.5x
Baseline is 0 to 1 creatures. Same figures the live engine uses to weight board wipes. Round-controlled, the lift is strongest at round 6.
Honest limits

What Sparring is not

Sparring is a training aid, not a full opponent. Being straight about the limits is the point, so you know exactly what you are practicing against.

Not a rules engine

Nothing here adjudicates card legality. A small curated pool of iconic interaction cards is tagged by hand with the card types it can answer, and a strict filter uses those tags. That target tag is why a counter never fires on a creature cast.

Answers are on your honor

When a phantom fires you decide: let it resolve, or say you have an answer and it fizzles. Sparring never looks at your hand and never checks whether your answer is real. Making that call well is the skill it trains.

A curated pool, not every card

The pool is roughly 120 hand-picked staples across the five archetypes, not the whole card pool. It is built for archetype believability and variety, not exhaustive fidelity, so you will see Swords, Counterspell, and Wrath, not every fringe answer.

Coarse threat model, and some noise

The phantom's sense of what looks threatening is deliberately simple in this version, and real interaction data carries some classifier noise. Expect the occasional off read, which is why every reaction has a one-click dismiss.

Sparring FAQ

What is Sparring in Playgroup Live?
Sparring adds phantom opponents to a solo Commander playtest. Instead of pure goldfishing, they occasionally fire believable interaction at you: a counter when you cast a spell, spot removal when a threat lands, a board wipe when you pass with a wide board. It exists to train the core Commander skill of playing around interaction, which no-resistance goldfishing never teaches.
Where do the numbers come from?
From real games tracked on Playgroup. The model is measured from 21,480 interaction spells cast across 6,934 finished Commander games: which archetypes get cast (spot removal, counters, wipes, bounce, edicts), when in the game each one peaks, and how strongly board wipes track how many creatures you have out. The same data file feeds both this page and the live product.
Does Sparring use a rules engine or an AI?
No. There is no rules engine and no AI opponent. Believability comes from the measured lookup tables plus a small hand-curated pool of iconic interaction cards, each tagged by hand with the card types it can legally answer. That target-type tag is what stops a counter from firing on a creature cast.
How does it know whether a card is a legal answer for me?
It does not, and it does not try to. When a phantom fires, you choose: let it resolve, or say you have an answer, in which case the reaction fizzles. That judgment is on your honor, and making it well is exactly the skill Sparring is training. A dismiss button handles the rare misfire with one click.
Why do board wipes fire more when I have a wide board?
Because that is what real tables do. In tracked games, a board wipe was about 2.5 times more likely to be cast against a player holding three or more creatures (49 percent of wipes) than the 20 percent baseline at zero or one creature. The signal first appears around round five, so Sparring never wraths an empty board or an early game.
Can I turn Sparring off or change how aggressive it is?
Yes. Sparring is off by default. In a solo playtest you can set it to Light, Realistic, or Cutthroat, which scale how often phantoms fire. Realistic matches the cadence measured from real games. You can change it mid-game and it takes effect on the next trigger.
Try it yourself

Spar with a deck for free

Playtest a deck

Start a solo playtest in Playgroup Live, turn on Sparring, and practice playing around counters, removal, and wraths. No account needed.

Playgroup Live Live Features How Many Lands