Hellickson’s changeup varies greatly in the amount of movement it has. It ranges anywhere from no-tail slideresque to full blown screwball. That would be a large reason why he makes it so effective against same handed hitters as he can get the equivalent of a harder breaking ball moving away from them, especially important considering how mediocre his curveball is.
As curveballs go, that is not spectacular. Kershaw, for example, allowed a .182 OPS against his curve last year. That’s not batting average, that’s OPS.
A curve is by it’s nature hard to hit. It is made to deceive a batter by moving in a way they don’t expect. Plus, as a breaking ball, one would expect a pitcher to use it more in favorable counts. Pitchers also usually use curveballs less frequently than other pitches, meaning a batter is less used to seeing it and less able to jump on it.
Some pitchers allow an overall OPS better than the .605 Hellickson allowed on his curve, and that is considering that curves should probably have a better OPS-against than other more frequently used pitches. This to me suggest his curve is mediocre at best.
I wonder if those numbers even mean anything. Pitches are effective in context, the changeup is effective because of how it’s mixed in with the fastballs. Same with the curveball, do batter numbers against the curveball actually matter? If the curveball is always outside of the zone and never swung at it might still be an effective pitch if it screws up batters’ ability to hit the fastball, yet you’d be giving the fastball all the credit and perhaps concluding that the pitcher should never throw the curveball, which in this strange hypothetical I’ve created wouldn’t be accurate. That’s obviously an extreme and implausible example, but I think it illustrates a point that a pitch can be effective for how it makes batters react to the NEXT pitch. These stats for a specific pitch don’t capture that at all.