Was Ferraro actually good at killing penalties? No, and you were right to ask
You caught me being lazy, and you were right to. I wrote that the Sharks "lose Ferraro's 268 penalty-kill minutes off a unit that ranked 28th" as if the minutes themselves were the value. They are not. Most minutes is a usage stat, not a quality stat, and on a 28th-ranked kill, the busiest guy is at least as likely to be a symptom as a loss. So here is the actual question answered: was he any good at it?
The numbers, defensemen on the kill
The cleanest test is how many expected goals the unit allowed per 60 with each defenseman on the ice shorthanded (lower is better), on the same team, same system, same goalies. Here is the Sharks blue line:
| PK defenseman | PK min | xGA/60 (suppression) | Actual GA/60 |
|---|---|---|---|
| Desharnais | 166 | 6.52 | 6.16 |
| Ferraro | 268 | 7.75 | 7.84 |
| Dickinson | 64 | 7.76 | 9.41 |
| Mukhamadullin | 69 | 7.77 | 6.96 |
| Orlov | 62 | 9.42 | 9.64 |
Read the top two rows together, because that is the whole answer. Ferraro killed roughly 100 more minutes than anyone else on the blue line, and did it while allowing chances at a meaningfully worse rate than Desharnais (7.75 against 6.52). On the penalty kill, a gap of more than a full expected goal per 60 is large. The coach handed the heaviest shorthanded workload to his second-best option, not his best.
Was he good, in absolute terms? No, he was average.
Ferraro's 7.75 lands in league-average penalty-kill territory, give or take. He was not a sieve, his actual goals against (7.84) matched his chances almost exactly, so he was neither bailed out by the goalies nor robbed by bad luck. He killed penalties at exactly a middling level. That is the honest verdict: not bad, not good, average, and overused. Desharnais (6.52) is the only Sharks defenseman who actually suppressed chances at a genuinely good rate, and quietly, so did the forward Collin Graf (6.52 over 196 minutes). Those are the two the kill should run through, not Ferraro.
So is losing him addition by subtraction?
Yes, conditionally. Removing 268 average minutes is only an upgrade if they go to better killers.
If those minutes shift to Desharnais and Graf (the two real suppressors) and a competent addition, the kill gets better by subtraction, exactly your point. If they just get redistributed to the leaky end of the group (Orlov at 9.42, Goodrow and Sherwood north of 9.0), nothing improves. So it is not "we lose Ferraro and the kill suffers." It is "we lose an average, overused killer, which is a chance to upgrade the role rather than a hole to fill."
And the deeper correction to my email, and to the "bring everybody back" trap: the 28th-ranked kill was not mostly a Ferraro problem, and it was not even mostly a skater problem. Look at the actual-goals-against column across the unit: Dickinson allowed 9.41 against 7.76 expected, Dellandrea 8.65 against 7.46, and the small-sample forwards were worse. Across the kill, the Sharks conceded a lot more than the chance quality said they should, which points straight at the goaltending behind it. You fix a 28th-ranked penalty kill two ways: give the minutes to your best suppressors (Desharnais, Graf), and get better goaltending on it (Askarov). Ferraro walking out the door does not hurt either of those, and it might help the first one.
The one-line answer
No, Ferraro was not good at killing penalties. He was a league-average killer given the most minutes, a full expected goal per 60 worse than Desharnais, and letting him walk is closer to addition by subtraction than to a loss, provided the minutes go to the players who were actually good at it.
Shorthanded (4-on-5) on-ice expected goals against and actual goals against per 60, MoneyPuck 2025-26, defensemen with 30-plus penalty-kill minutes. Same team, system, and goaltending, so the comparison is apples to apples. Team PK ranked 28th (75.8%) per state/counting_stats.md.
