Widening a champion pool that has quietly shrunk to three picks

Pools narrow by default, not by decision

The mechanism is simple and it's mostly rational. Champion select is a moment of mild pressure — a countdown, four teammates watching, a match about to start — and under pressure people reach for what's familiar. Each individual lock-in is a sensible choice. The problem is only visible in aggregate: a hundred sensible choices in a row and you've trained yourself out of every character you didn't already know.

What makes this hard to notice is that a narrow pool doesn't feel like a weakness while it's working. It feels like mastery, and up to a point it genuinely is — depth on a few picks beats shallow knowledge of forty. The cost shows up sideways: in the games where your main is banned, in the matchups your pick simply loses, in the team comps that needed a tank and got a third assassin because that's what you play. You don't lose those games to a lack of skill. You lose them to a lack of options.

Randomness beats intention here

"I should learn more champions" is a resolution that never survives contact with a lock-in timer, because the decision to try something new has to be made at exactly the moment you're least inclined to make it. Handing that decision to a random draw before you queue removes the negotiation entirely. There's no timer, no teammates waiting, and no chance to talk yourself back into the comfortable pick.

The other advantage of a random pick is that it's genuinely indifferent to your taste. Left to your own judgment, you'd expand your pool toward things that resemble what you already play — another mid-range mage, another aggressive duelist. A draw doesn't care. It'll hand you the enchanter support or the anchor you've been avoiding for two years, and those are precisely the picks that teach you something, because they're the ones your instincts don't already cover.

The Random Champion / Hero Picker covers four games — League of Legends, Valorant, Overwatch 2, and Rainbow Six Siege — each with its own roster rather than one generic list. Pick your game, hit the button, and you get a name. That's the whole ritual, and it takes about three seconds before you queue.

The role box speaks each game's own language

This is the part worth knowing, because it's the one place the tool will quietly do something other than what you meant. The optional role field matches against the class names that game actually uses — and each of the four games uses a different vocabulary. League of Legends organizes its roster by tank, fighter, assassin, mage, marksman, and support. Valorant uses duelist, initiator, controller, and sentinel. Overwatch 2 uses tank, damage, and support. Rainbow Six Siege doesn't use roles at all — it splits into attackers and defenders, which are sides, not positions.

So if you type adc into the role box for League, you won't get marksmen. The roster is keyed on marksman, adc matches nothing, and the picker falls back to drawing from the entire champion list. Same for dps in Overwatch, where the key is damage, or flex anywhere. The fallback is deliberate — an unrecognized filter gives you the full roster rather than an error or an empty result — but it's silent, so a typo looks identical to a successful filter. If you asked for a support and got handed a marksman, you didn't hit a bug: you hit the fallback.

The practical rule: use the game's own word, and if the result doesn't look like the role you asked for, check your spelling before you assume the draw was unlucky. Leaving the box empty is a perfectly good default — that's the widest possible pool, which is the whole point if you're trying to escape a rut.

Drawing more than one

The count field goes up to five, and the picks come back distinct — no duplicates, even across a five-draw, and if you ask for more names than the filtered pool holds you'll simply get the whole pool. That opens up a few uses beyond the solo warm-up. Draw five for a full custom-lobby comp and let the team sort out who plays what. Draw three as a menu rather than a mandate — a soft version of the ritual where you still get a choice, just not the choice you'd have defaulted to. Or draw two and pick the one you like less, which is a genuinely effective way to learn quickly if you can stand it.

Where to actually play the pick

One caveat worth stating plainly: a champion you've never played is a liability to four other people, and the place to work through that liability is not ranked. Take random picks into normals, custom lobbies, or whatever your game calls its low-stakes queue. The point of the draw is to get reps on something unfamiliar, and reps go better when a rough game costs you fifteen minutes instead of a week of climbing.

A reasonable cadence is one drawn pick per session, played for two or three games rather than one. A single game on an unfamiliar character teaches you almost nothing except that it felt awkward — which you already knew. Three games is usually enough to find the thing the character is actually for. Then draw again next session, and let the pool widen a name at a time.

Try it before your next queue

The Random Champion / Hero Picker runs in your browser, needs no account, and works on your phone while you're sitting in a lobby. If you want to push the variety further, the Random Strat Roulette Generator hands you a random restriction instead of a random character, the Random Map Picker settles the map argument before you queue, and the Random Team Generator splits a full lobby into sides. A three-champion pool isn't a character flaw — it's just what happens when every individual decision is made under a timer. Make one decision before the timer starts instead.