Regarding "overturnedness":
Judge Posner wrote a law review article (29 J. Legal Studies 711 (2000)) in which he examined the reversal rates for the 12 circuits (11 + DC) from 1985-1997.
For that time period, the 9th had the highest percantage (of its total docket) of
summary reversals (which don't occur that often) (the 1st, 7th, and DC had none); the 9th also had highest percentage of
unanimous nonsummary reversal (the 11th* was the lowest); the highest percentage of
nonsummary, nonunanimous reversals was the
DC Circuit (11th was again the lowest). (Posner, a huge critic of the 9th, handwaives that away.)
Mind you: We're talking about the difference between a 99.88% affirmance rate (11th) and a 99.57% affirmance rate (DC).
Regarding circuit splits: For political reasons, a circuit will have at least three states in it. California is approximately half of the 9th's docket. So if the goal really was to split the docket up, you would choose California and two states with smallest dockets. I don't recall presently which states they are, but I remember those three states aren't contiguous. So the "smallest docket" proposal usually is California, Hawaii, and Alaska; however, Alaska doesn't want to be in 9th, so that's probably not viable from a politics standpoint. (In 1995, a bill in Congress suggested the 9th consist of California, Hawaii, and the pacific territories; this violated the "three state" "rule.")
A related issue is the division of judges. Most proposals to split would allocate more judges to the new 12th than would stay behind in the old 9th. Which would effectively
increase the case/judge ratio in the old 9th -- and the "reason" given for splitting the 9th is that the case/judge ratio is already too high.
* The 11th was part of the 5th, but they split in 1981.