Is the ACPL system broken for chess960?

I played a 960 game which resulted in an ACPL of 8, whereafter the analysis stated my opponent's ACPL was 30. This seems a little bit crazy because my opponent hung multiple pieces. I imagine that my ACPL stayed low largely from picking up dropped pieces and trading down, strange!

You played on so long while your opponent had no pieces that every move he played from then on out was "accurate" by force.

@grapejam is correct.

Analysis done with stockfish 7, 19 plies depth:
White:
Total moves: 48

Average CP loss: 71.94
T1 moves: 17 (35.42%)
T2 moves: 10 (56.25%)
T3 moves: 8 (72.92%)

---
Black:
Total moves: 47

Average CP loss: 248.82
T1 moves: 18 (38.30%)
T2 moves: 4 (46.81%)
T3 moves: 7 (61.70%)

----
That is, of course, unfiltered analysis.

Consider that the evaluation immediately preceding mate scores was +14.1 (1410 CP).

1410 / 48 moves is approximately 30.

Cheers guys, the more I read into ACPL statistics the more I find it a largely irrelevant concept aside for it's use as a convenient standard to measure a single move's + / -

The thing that bugged me was that over three 960 games I have averaged 8.5 ACPL, which seems insane as I know in the past when I have played standard variants I often float 50-100. I understand that 3 games is not a large enough sample to make this accurate at all though.

Thanks again, the input was much appreciated!