Steelmanned

An integrity referee for public debates. We score how each side argued, never who won.

Ben Shapiro

2 scored debates

32 challenges faced · 13,460 words scored

89
Debate integrity · record

Range 87–94 across 2 debates. Weighted by challenges faced, not words spoken: how hard each performance was tested.

Where these scores sit

050100

Ben: 94 (v panel of 5) · 87 (v Sam Harris) · shaded to our ±5–8 re-run band · all 46 scored performances shown

Inside our ±8-point re-run band of this record: Chris Cuomo's 89, Mehdi Hasan's 89, Sam Harris's 89, Viva Frei's 89. We treat these as indistinguishable from it.

Component standings

Confronted the point94site 52–97
Argued clean90site 6–100
Dealt straight67site 0–87
Let them talk96site 75–100

Ben's strongest component is “Let them talk” at 96. Across the 46 scored performances Steelmanned has published, that component runs 75–100.

A 100 on "Argued clean" means no fallacy points were charged, not that no fallacies were logged. The scorecard lists every logged event, including zero-valued ones.

The record

David Pakman ShowPanel debate0h 41m

David Pakman, Ben Shapiro, Don Lemon, Tara Palmeri | Reuters NEXT Conference Debate, December 2025

Ben Shapiro
94
DIS ±8
On a panel with Don Lemon, David Pakman, Tara Palmeri & Mitali Mukherjee
CRUX45% · N=9
93
CLEAN30% · 2,194W
100
DEALT15% · N=9
77
TALK10% · 26 TURNS
100

Confronted 9 challenges (20.5/22); owned an error that creators his company brought on came 'some for good, some for ill' with 'regrets, we have a few', named common ground ('I totally agree with you'), and conceded creators are 'not the ones who are doing a lot of the core reporting'; no fallacies recorded across 2,194 words.

9 challenges faced · 2,194 words scored · rubric v0.11
The Free PressPanel debate2h 13m

Trump or Kamala? Ben Shapiro and Sam Harris Debate.

Ben Shapiro
87
DIS ±8
Opposite Sam Harris · their score 93
CRUX45% · N=23
95
CLEAN30% · 11,266W
86
DEALT15% · N=23
63
TALK10% · 107 TURNS
95

Confronted 23 challenges (48.3/51 weighted); answered his Never-Trump reversal head-on, owned his 2016 predictions as wrong, conceded 2020 was 'illegitimate' and 'I don't think that he won,' and named himself 'quite anti-tariff'; recorded 7 weighted fallacy points (0.62/1000 words) including pivoting to 'who does Iran want to win' when pressed on Putin; left the cult-of-personality/loyalty-test crux (P15) engaged-not-answered.

23 challenges faced · 11,266 words scored · rubric v0.11

Portrait credits