NEW MINI-INVESTIGATION: How NYC reports a violent-crime decline that didn’t happen. Let’s look at the last 28 days.
🚨 There was no 5.6% drop in violent crime. NYPD’s own report produces that number by comparing two different kinds of data. Match them, and the decline vanishes — apples-to-apples, violent crime is flat, with three of four categories actually higher than a year ago.
Here’s the trick.
For the latest 28-day period, NYPD compares fresh, preliminary 2026 numbers against 2025 numbers that have had a full year to be revised upward. Both reports admit it in writing: figures are “preliminary and subject to further analysis and revision.” The current year hasn’t been revised yet. Last year has — all year long.
Compare the same kind of data — 2026 preliminary against the preliminary 2025 numbers NYPD itself published for the identical 28-day period a year ago — and the 5.6% decline doesn’t shrink. It disappears.
As the report frames it (2026 fresh vs. 2025 revised):
• Murder: 22 vs. 25 (−12.0%)
• Rape: 153 vs. 153 (0.0%)
• Robbery: 1,040 vs. 1,253 (−17.0%)
• Felony Assault: 2,517 vs. 2,523 (−0.2%)
Total violent felonies: 3,732 vs. 3,954 = −5.6%
Same kind of data (2026 fresh vs. 2025 fresh):
• Murder: 22 vs. 14 ( 57.1%) — small numbers, so read the counts, not the percent
• Rape: 153 vs. 146 ( 4.8%)
• Robbery: 1,040 vs. 1,191 (−12.7%)
• Felony Assault: 2,517 vs. 2,371 ( 6.2%)
Total violent felonies: 3,732 vs. 3,722 = dead flat.
Three of the four major violent-felony categories are higher than a year ago. The “decline” is an artifact of the comparison, not the crime data.
Here’s where the 5.6% comes from. Over the past year, NYPD revised the 2025 numbers upward — its own counts for the identical 28-day period, a year apart:
• Murder: 14 → 25
• Rape: 146 → 153
• Robbery: 1,191 → 1,253
• Felony Assault: 2,371 → 2,523
Total: 3,722 → 3,954 ( 232 crimes)
Inflate last year’s baseline by 232 crimes, hold this year’s un-inflated, and you’ve conjured a 5.6% “improvement” out of thin air. Nothing about actual crime changed. Only the maturity of the data did.
The revisions themselves are routine. The spin isn’t. And this isn’t a one-administration problem — CompStat has reported preliminary-against-revised under every mayor. That’s the point: the comparison is built to flatter whoever’s in office, every single week.
If 2026 follows the same pattern, these “fresh” 2026 numbers will get revised upward too — and today’s flat line will tilt positive.
One window, equal length, both sides 28 days, both numbers straight from NYPD’s own reports. (I’m not touching year-to-date — the 2026 report runs through June 14, the 2025 report through June 15, so YTD isn’t a valid one-to-one comparison. The 28-day window is.)
The report’s framing: down 5.6%.
Apples-to-apples: flat, with three of four categories up.
Same source. Same crimes. The decline only exists if you grade two different years on two different curves.
ALT NYPD misleading crime stats