It would be a mistake to read the chart above as proof that outside money does not work. The causation runs the other way. Money floods a race because it is close, so the races that appear here are, almost by definition, the ones decided by a few points. The safe seats — where a well-funded incumbent faces no serious threat and no super PAC bothers to spend — never enter the sample at all.
What the data does show is narrower, and still striking: once both sides are spending at scale, outspending your opponent buys no reliable advantage. The 2020 and 2022 cycles ran in opposite directions — the better-funded candidate lost three-quarters of the 2020 races and barely one in ten of the 2022 races — which is what a coin flip looks like, not what a purchased outcome looks like.
The money is not buying victory so much as buying presence in the races that decide control of the Senate. That is a different claim, and a more defensible one.