Research · 4 min read
Why Two Properties a Mile Apart Can Face Very Different Security Problems
Repeat-block history can prioritize a closer review, but it is not a forecast or property risk score.
Two properties a mile apart can look identical on a citywide crime map. In three years of Seattle property-offense records, the median block appears exactly twice, while the highest-count 10% of blocks hold 56.2% of everything recorded. Which kind of block you’re on is the question this article answers.
Our analysis of 114,934 records with a valid published 100-block label, July 1, 2023–June 30, 2026.
Key takeaways
- The highest-count 10% of published blocks — 1,975 of them — held 56.2% of 114,934 accepted property-offense records, while the median block recorded two.
- Concentration runs well beyond one extreme: the top 1% of blocks held 20.4% of records, the top 5% 43.2%, and the top 20% 70.0%.
- Weighting recent records more heavily moved the top-decile share only from 56.2% to 57.4%, so recurrence is a stable feature of this data.
- Repeat-block history is a reason to inspect a property more closely — not a forecast or a property risk score.
Is a citywide property-crime average useful for one property?
Not by itself. Across 114,934 accepted Seattle property-offense records from July 2023 through June 2026, the highest-count 10% of published blocks account for 56.2% of records while the median block has two. Recurrence can prioritize review, but it is not a property risk score.
The numbers behind the answer
Selected measures only. Denominators and interpretation stay attached so the headline cannot stand alone.
56.2%
Records on the top 10% of blocks
64,551 of 114,934 accepted property-offense records appeared on the 1,975 highest-count published blocks.
2 records
Median block
The middle published block had two accepted records during the 36-month study period.
27.4%
Blocks with 5+ records
5,405 of 19,741 valid published blocks contained at least five accepted records.
The typical block and the high-count blocks tell different stories
A citywide total answers one question: how much did Seattle publish? Across July 2023 through June 2026, that’s 114,934 accepted property-offense records. What it can’t tell you is whether your address resembles the middle of the city or the top of it — and those are very different places. The median published block recorded exactly two incidents in three years. The 1,975 highest-count blocks, the top 10%, held 56.2% of all accepted records.
That gap makes recurrence worth checking on your block. It doesn’t make it a forecast. Police records count reported events, block labels are blurred approximations, and nothing in the data says how many businesses, visitors, or vehicles were there to be targeted. A high count tells you where to look closer — never what happens next.
It also helps to name what each count actually is. A row is an accepted offense report, not a unique victim, property, or confirmed event — and what gets reported at all varies by offense, neighborhood, and channel. A block’s total therefore blends three things: what happened, what was noticed, and what was reported. Only the third is fully visible here.
Decision rule
Use repeat-block history to decide where to investigate—not to decide that a property is safe or unsafe.
Concentration appears well beyond one extreme percentile
This isn’t a handful of notorious blocks. The top 1% — just 198 blocks — held 20.4% of all records. The top 5% held 43.2%. The top 10% held 56.2%, and the top 20% held 70.0%. However you draw the line, most of what Seattle recorded happened on a minority of its blocks.
Could that be an artifact of one hot year? We reran the top-decile calculation with recent records weighted more heavily. The share moved from 56.2% to 57.4% — barely at all. Recurrence is a stable feature of this data. It still isn’t a prediction.
Evidence visual
Share of accepted records on highest-count blocks
Each bar is cumulative. The groups contain 198, 988, 1,975, and 3,949 blocks respectively.
Turn recurrence into a five-question property review
So how do you use this without overusing it? Start with the public record: does your block recur, and how recently? Then look at what the count is made of. A block dominated by a transit hub or a big venue next door is telling you something different than a block of properties like yours.
Then leave the data behind and go look at the property. The public record can’t see your lighting, your loading dock, your sightlines, or your hours — it can only tell you where to point that inspection. The five questions below are the sequence.
- How many accepted records recur on the published block?
- How recent are they, and does recency weighting change the ranking?
- Which offense subcategories make up the count?
- Could a large venue, transit node, or adjacent parcel dominate the block label?
- Which property conditions can be checked directly?
Seattle’s curve is steep even against the benchmark that defined the field
The local curve holds up against the number that made this pattern famous. The top 5% of Seattle blocks — 988 of 19,741 — held 43.2% of accepted records, and the top 10% held 56.2%. The law of crime concentration reports roughly half of a city’s crime on about 5% of street segments; Seattle’s broader property-category share sits near that mark, and the 14-year Seattle street-segment trajectories found the same micro-places stayed elevated while the citywide total fell. Repeat and near-repeat victimization supplies one engine: a block hit once is more likely to be hit again, and its neighbors rise with it — the burglary-as-a-clue-to-the-future effect.
The cut point moves the headline more than the pattern moves. Crawling from the top 1% (198 blocks, 20.4% of records) to the top 50% (9,871 blocks, 89.0%) traces a smooth curve rather than a cliff — the signature of a heavy-tailed distribution, with a Gini of 0.647 and a median block of two records against a 95th percentile of 21. That is why the curve, not one headline share, is the finding.
Geography sharpens the caveat. The top-decile share ranges from 40.7% in Sandpoint to 71.3% around Roxhill and Arbor Heights, with Downtown Commercial at 61.1% across 411 blocks. But those are counts, not rates: 7,567 records across Queen Anne’s 1,165 blocks and Northgate’s 60.7% say nothing about how many targets each held — the denominator this source simply lacks. Concentration is a stable description of where records recur. It is not a forecast for a specific property.
What this is not
Concentration is a stable description of where records recur. It is not a prediction for a specific property.
Questions property teams ask
Does a high-count block mean my property is likely to be targeted?
No. It identifies repeated published records near a blurred block label. It is a reason to inspect current property conditions, not a property-level probability.
Why analyze blocks instead of exact addresses?
The source publishes block-level location labels rather than precise incident addresses, so the analysis uses the geography actually available.
What qualifies as a repeat block?
This article shows the full distribution and separately reports blocks with at least five accepted records; five is a descriptive threshold, not an official designation.
Should older records carry the same weight as recent records?
Not always. A 365-day recency-weighted test produced a similar top-decile share, but a property decision should still examine recent change alongside the longer baseline.
Is block-level concentration unique to Seattle?
No. Concentration at micro-places is one of the most replicated findings in criminology, and the long-run Seattle street-segment research reaches the same conclusion. The magnitude varies by city, but the pattern does not.
Why doesn’t the study control for how busy a block is?
The public source carries no exposure denominator — no counts of businesses, parcels, visitors, or vehicles. Concentration is therefore descriptive: it shows where records cluster, not how much risk each target carried.
Our methods, limits, and sources
How we calculated this
This is original 911 Sentinel research — we gathered the records, ran every calculation below, and published the aggregate dataset.
We pulled Seattle police property-offense records, kept only rows with a usable published block label, and ranked all 19,741 distinct blocks by record count.
- We retained accepted property-offense rows in the July 2023–June 2026 window.
- We excluded 1,880 rows without a usable published block label.
- We counted records by normalized block, ranked blocks, and calculated cumulative concentration bands.
- We re-ran the top-decile share with 365-day recency weighting and alternate label normalization to check stability.
What this analysis cannot establish
- A record is not a unique victim, property, call, or future event.
- Published block labels are approximate and may combine activity from several properties or public spaces.
- No exposure denominator controls for businesses, parcels, visitors, vehicles, or operating hours.
- Concentration is descriptive and does not identify why records recur.
Sources
The raw records come from the sources below; the study design, analysis, charts, and conclusions are our own.
- Seattle Police Department — SPD Crime Data: 2008-Present
- Published aggregate result (JSON)
- Source manifest (JSON)
- David Weisburd (Criminology) — The Law of Crime Concentration and the Criminology of Place
- David Weisburd, Shawn Bushway, Cynthia Lum & Sue-Ming Yang (Criminology) — Trajectories of Crime at Places: A Longitudinal Study of Street Segments in the City of Seattle
- Shane D. Johnson & Kate J. Bowers (European Journal of Criminology) — The Burglary as Clue to the Future: The Beginnings of Prospective Hot-Spotting
- Bureau of Justice Statistics — National Crime Victimization Survey (NCVS)
Related questions and practical guides
No obligation · free property walk
Put the block history beside the property itself
A site assessment can compare public recurrence with entrances, parking, loading, lighting, camera views, and operating patterns at the actual property.