Home / Guides
Why a 90 in one city is not a 90 in another
Restaurant inspection results look alike from city to city. The scales, cutoffs, placards and inspection schedules behind them do not.
People compare restaurant inspection results across cities all the time: in news stories, in apps, and in spreadsheets built from open data. The comparison rarely holds. Cities do not share a scale, and in several the scale runs in the opposite direction. Some publish no number at all. Before you put two cities' results side by side, check four things: which way the scale runs, where the grade cutoffs sit, how often places are inspected, and what the public is actually shown.
Six systems at a glance
| Jurisdiction | Scale | Cutoffs or outcomes | What is posted | Source |
|---|---|---|---|---|
| New York City | Points for violations; lower is better | A: 0-13 points; B: 14-27; C: 28 or more | Letter grade card | NYC DOHMH |
| Los Angeles County | Starts at 100; points deducted | A: 90-100; B: 80-89; C: 70-79; below 70 gets a score card, not a grade | Grade or score card | LA County DPH |
| San Francisco | No numeric score since 2021 | Pass, conditional pass, closure | Green, yellow or red placard | SF.gov |
| Chicago | No score | Pass, pass with conditions, fail | Result only | Chicago data portal |
| Toronto | No score | Pass, conditional pass, closed | Result notice | City of Toronto |
| King County, WA | Average of recent routine inspections | Excellent, Good, Okay, Needs to improve | Rating sign | Public Health Seattle & King County |
1. The scale can run either way
In New York City, inspectors add up points for each violation, and the Health Department says "The lower the score, the better the grade." Its scoring guide sets minimums by severity: a public health hazard "triggers a minimum of 7 points," a critical violation "carries a minimum of 5 points," and a general violation "receives at least 2 points," with more points added for how extensive the problem is.
Los Angeles County does the reverse. Per the county's grading page, "Each food inspection begins with 100 points" and the point value of each violation is deducted from 100. A high number is good.
So a restaurant with a 12 in New York earned an A, and a restaurant with a 12 in Los Angeles County would be far below the 70 needed for any letter grade. Neither number means anything outside its own system.
2. Grade cutoffs are local decisions
Even where two places use letters, the letters are cut from different scales:
- New York City: "Restaurants with a score between 0 and 13 points earn an A, those with 14 to 27 points receive a B and those with 28 or more a C" (NYC DOHMH).
- Los Angeles County: 90 to 100 points is an A, 80 to 89 a B, 70 to 79 a C, and "A facility receiving a score less than seventy percent (70%) will be issued a score card and not a grade card" (LA County DPH).
- King County, Washington: no letters. For the highest-risk category of establishment, the rating comes from the average of the last four routine inspections: 0.00 to 3.75 is Excellent, 3.76 to 16.25 is Good, 16.26 and up is Okay. "Needs to improve" means the place was closed within the last 90 days or needed multiple return inspections, and "is not based on average scores" (King County).
King County's system also changes the unit of measurement. An NYC grade reflects one inspection cycle; a King County rating is an average across several inspections. Even a restaurant in the same county can be rated on two inspections instead of four: lower-risk categories use "the average score of up to two routine inspections."
3. Some places publish no score
San Francisco scored restaurants on a 100-point scale until 2021. Its scores dataset now says "As of 2021, the LIVES standard is no longer being used to evaluate restaurants." The city posts placards instead: green means pass ("zero to 1 major violations"), yellow means conditional pass ("2 or more major violations," corrected during the inspection), and red means closed (SF.gov). Any site still showing an SF score from after 2019 is showing something the city no longer produces. Details in our San Francisco guide.
Chicago never used a numeric score in its public data. Per the dataset's description, "An inspection can pass, pass with conditions or fail." Toronto's DineSafe program is similar: "Each inspection results in a pass, a conditional pass or a closed notice."
A pass in Chicago and a green placard in San Francisco look alike, but the thresholds behind them are each city's own, set in different terms (Chicago's in critical and serious violations, San Francisco's in major violations).
4. Inspection frequency differs, and it changes what a result means
How often a place is inspected shapes how old its latest result is and how many results it accumulates.
- New York City ties frequency to performance. The Health Department's inspection cycle overview puts the next cycle at 11 to 13 months after an initial score of 0 to 13 points, 5 to 7 months after an initial score of 14 to 27, and 3 to 5 months after 28 or more. An initial inspection that does not earn an A is scored but not graded, and the restaurant gets a graded re-inspection; the city says "A restaurant has two chances to earn an A in every inspection cycle" (NYC DOHMH).
- Chicago ties frequency to risk category: "risk 1 establishments inspected most frequently and risk 3 least frequently" (Chicago data portal). See our Chicago guide.
- King County builds frequency into the rating itself by averaging several routine inspections.
The practical consequence: in New York, the restaurants with the worst records generate the most inspection rows, so counting rows over-represents them. In Chicago, rows are weighted toward risk 1 establishments. A raw average across a dataset is shaped by the inspection schedule as much as by conditions in kitchens.
5. What LIVES tried to do about it
The LIVES specification asked every city to express scores "on a 0-100 scale" where "100 is the highest score," and gave each city an optional legend file to describe its own ranges. That puts numbers in the same column, but the legend exists precisely because the meaning stays local. See our LIVES page.
If you need national figures
Rates such as the share of inspections that fail, or the most common violations, only mean something when they are worked out from each city's data on its own terms, one scale at a time. We do not compute cross-city rates on this site, and any figure that blends cities with different scales deserves suspicion.
A checklist before you compare
- Which direction does the scale run, and what is the maximum?
- Is the posted result a single inspection or an average?
- What are the grade cutoffs, and which inspections are graded at all? (In New York, several inspection types are scored but not graded.)
- How often is each place inspected, and does frequency depend on past results or risk?
- Is the city still producing the measure? (San Francisco stopped in 2021.)
- Is the dataset still updating? Check the portal directory.