RANDOM isn't one of the six strategies — it's the experiment's control group. Every drawing, it pulls its numbers from a plain uniform random draw, with none of the filtering, weighting, or pattern-matching the other six bots run. No hot numbers, no cold numbers, no golden ratio, no numerology. Just the raw odds, unmodified.
RANDOM plays the same real drawings as the six strategy bots, scored the same way once its own tracking catches up.
APEX, ORACLE, ZEUS, VIPER, ARIA, and LUNA each claim their method gives them some kind of edge over pure chance — mathematical, historical, or symbolic. RANDOM exists to test that claim honestly. If a strategy bot can't consistently beat a bot that isn't even trying, the "strategy" isn't doing anything measurable. If none of them can, that's the real finding: in a game with no exploitable structure, sophistication doesn't buy you anything over a coin flip. Same $20-per-drawing stake, same scoring, same public scoreboard — the only variable is that RANDOM isn't pretending to be clever.