Box scores are built to flatter scorers. Points are counted every time they happen; defence mostly isn’t. A player who gets 18 a night looks productive on paper even if he’s a turnstile on the other end, and a player who does real defensive work but doesn’t score much can look replaceable, even when he isn’t.
This is a real blind spot, and it’s worth being deliberate about correcting for it rather than trusting the eye test alone, because the eye test has the same bias the box score does: scoring is visible, defense is often not.
The two honest signals box-score data actually gives you for defense are steals and blocks , not a complete picture of defensive value (help positioning, closeouts, and physicality never show up in a box score), but a real, verifiable floor. Combine that with scoring and playmaking output, and you get a genuinely useful split: who contributes on both ends, who’s carrying the offence but giving it back on defense, who’s a legitimate defensive presence without much offensive value, and who isn’t doing much of either.
The players worth the most roster flexibility are usually the ones in that first group, real offensive production alongside real defensive activity, because they’re the ones who don’t force you into lineup trade-offs. A pure scorer forces you to hide him defensively. A pure defender forces you to find offense elsewhere. A two-way player doesn’t force either compromise.
This is why MelonIQ’s Two-Way Value chart plots scoring-plus-playmaking output against steals-plus-blocks for every player in a league , not as a full defensive rating (box-score data can’t honestly support one), but as a clear, real signal for a question that’s otherwise easy to answer with a gut feeling instead of a number.