Stay or switch (BASIC)
Your Garchomp is on the field and the opponent just brought in a high-SpDef Steel wall. Next turn you've got two impulses: stay and hit it, or pull your wallbreaker out. The problem is they're thinking the exact same thing. You switch, they switch, no move lands this turn — and the board is now a completely new matchup. That's a double switch, the sweatiest moment in singles.
Core: a 50/50 is a bet, not a read (BASIC)
Picture a double switch as flipping cards at the same time. There's no "think harder and you win" option:
- Stay in: they stay → you get damage; they switch → you wasted the turn
- Switch out: they stay → you handed them a free turn; they switch too → double switch happens, matchup resets
None of the four boxes is a guaranteed win. Beginners think they lost because they "didn't read the opponent." You're actually facing a gamble under incomplete information. Good players don't guess right every time — they only gamble when winning the gamble pays off.
Worked: is this 50/50 actually worth it (INTERMEDIATE)
Garchomp is in after one Swords Dance (+2 Attack); the wall is at full HP. Your Earthquake is STAB, so ×1.5. It deals 89%–105% to that wall — and because damage rolls in 16 fixed steps, roughly 5 of those 16 rolls are an outright OHKO, with the rest landing 89%–99%.
Now split the branches:
- You stay, they stay: ~31% to flat-out KO; even on a non-KO roll they're left in the red and basically defanged.
- You stay, they switch (into their Water answer): your Earthquake hits air, and that Water-type one-shots your +2 Garchomp next turn.
The payoffs are lopsided: win the gamble = small gain or a KO, lose it = you throw away an already-boosted win condition. That's a high-risk 50/50 — shrink it.
Your turn (faded): same board, but now your Earthquake only does 44%–52% to the wall (a 2-hit KO), and guessing wrong just costs you one turn of chip — nobody faints. High-risk or low-risk? Gamble or fold?
Decision: if-X-then-Y (INTERMEDIATE)
Use the cost of being wrong as your only yardstick:
- Wrong call costs only HP, nobody faints → low risk, gamble, apply pressure
- Wrong call drops your sweeper / win condition → high risk, shrink, find a safer landing spot
- Can't tell → ask: "If I'm wrong, can I still win this game?" If yes, gamble; if shaky, fold.
Exceptions: when you don't have to guess at all (INT/ADV)
- Wide coverage deletes the gamble. If your wallbreaker beats every realistic thing they'd double into, staying in is never a loss — you're not even gambling. A mon that threatens multiple defensive cores costs more precisely because it erases these 50/50s for you.
- You've read their habit. The first 10 turns aren't just trading damage — they're intel. Does this player stay in when threatened, or pre-switch? Once you know, the 50/50 stops being 50/50.
- Speed is hard info. If you're sure you outspeed (and aren't paralyzed), you get to see one more thing before committing. Ranked Battles has no Tera and no Dynamax, so speed tiers stay stable and that read is worth more — but watch out: paralysis isn't just the 12.5% full-stop, it also halves your Speed, so the lead you counted on may already be gone.
Where beginners die (INT/ADV)
The mistake: treating every 50/50 as a battle of wits you must "win," so you switch constantly chasing the perfect turn. You keep gifting free turns, the opponent sets up, a +2 sweeper comes online, and you get swept.
The fix: in most double-switch spots, just staying in and clicking damage is the higher-EV play. Switch out only when you're sure you can't break through and staying is death, or when you've genuinely read their switch. Otherwise, don't trade guaranteed damage for a coin flip. Stop gifting turns first; get fancy later.
Now do this (ADVANCED)
Recall — or go play — a singles game, and freeze on the turn where stay-or-switch genuinely twisted your gut. Open Coach and run it through this lens:
- Draw the four boxes — I stay / I switch × they stay / they switch — and write each outcome.
- Tag the cost of "I guessed wrong": lost HP, or a lost win condition?
- Run the key hit's damage range in the calculator (with the boost and STAB) and see where the KO chance falls across the 16 steps.
- Shrink the high-risk ones, gamble the low-risk ones — next time you face that same mon, you already have the answer and don't have to gamble at all.