PulsCar analysis of NHTSA public data, September 2026 The silent no-start: what 495 owner complaints and 2,331 manufacturer bulletins say Licensed CC BY 4.0. Queries run 18 September 2026. -------------------------------------------------------------------------- SOURCE FILES -------------------------------------------------------------------------- NHTSA ODI complaint flat files COMPLAINTS_RECEIVED_2015-2019, _2020-2024, _2025-2026 -- https://static.nhtsa.gov/odi/ffdd/cmpl/ (dictionary CMPL.txt) 777,040 distinct vehicle complaints received 1 Jan 2015 to 8 Sep 2026. NHTSA manufacturer communication flat files TSBS_RECEIVED_2015-2019, _2020-2024, _2025-2026 -- https://static.nhtsa.gov/odi/ffdd/tsbs/ NHTSA recall flat file FLAT_RCL.zip, investigation flat file FLAT_INV.zip. -------------------------------------------------------------------------- CASE SENSITIVITY -- READ BEFORE REUSING ANY QUERY -------------------------------------------------------------------------- NHTSA complaint narratives are ALL CAPS through the end of 2020 and mixed case from 2021 onward. Measured on the file: of 58,465 complaints received in 2020, 58,386 are identical to their own uppercase form; of 62,775 received in 2023, only 939 are. An uppercase-only regular expression or an uppercase SQLite GLOB silently discards most post-2020 complaints while still returning a plausible-looking number of rows. Every query here is case-insensitive -- Python regexes with re.IGNORECASE, SQL text matching with LIKE over lower(), which is case-insensitive in SQLite for ASCII. -------------------------------------------------------------------------- WHICH COMPLAINTS COUNT (the silent no-start set: 495) -------------------------------------------------------------------------- Stage 1 -- SQL prefilter. Complaints whose narrative contains any of: "not start", "no start", "wont start", "won't start", "fail to start", "failed to start", "would not start", "will not start". 17,252 complaints. Stage 2 -- exclusions applied to the WHOLE narrative, in this order: (a) the engine CRANKS -- "cranks but", "turns over but", "engine cranks", "cranked fine/ok/normally/strong", "will crank but". That is the crank-no-start symptom and it belongs to a different page. 832 complaints dropped. (b) a CLICK is reported and not negated -- any form of "click" without "no click", "not even a click", "without a click", "didn't click". 529 complaints dropped. (c) the car STALLED or lost power WHILE DRIVING and then would not restart. That is the stalling symptom; the no-start is downstream of it. 1,239 complaints dropped. (d) EV charging-session failures and clearly mechanical engine failures (coolant into the cylinders, seized or hydrolocked engine, rod knock, misfire with reduced power). 354 complaints dropped. Stage 3 -- sentence test. Split at . ; ! ? and keep the complaint if any sentence contains a NO-CRANK phrase AND a no-start / dead-battery / jump phrase appears in that sentence or the one either side of it. NO-CRANK phrases: "no crank", will/would/does/did/could/can not crank, won't/wouldn't/doesn't/didn't crank, will/would/does/did not turn over, won't/wouldn't/didn't turn over, engine/motor/vehicle did not turn over, "no click", "not even a click/sound/noise", "without a click", complete / total / absolute / dead / just silence, was/is/were/remained completely or totally silent or dead, turning or pressing the key/ignition/start button within 60 characters of nothing / no response / no reaction / no sound / no noise / no click / dead / silent (in either order), "nothing happened when/after/upon". Sentences matching "no sound or warning", "no noise or light" and "without any warning" are skipped, because that phrasing describes a stall arriving without warning rather than a silent starter. Result: 495 complaints. -------------------------------------------------------------------------- SAMPLE READING (the false-positive rate we found) -- READ THIS FIRST -------------------------------------------------------------------------- This set was read three times, in three versions, and the honest headline is that it is the noisiest set we have published. Round 1 (1,141 rows, before the crank and stall exclusions): 30 read in full. Roughly a third were wrong -- crank-but-won't-start narratives, stalls, and engine failures. Those readings produced exclusions (a) and (c). Round 2 (534 rows): 30 read in full. About one in three were still misfits, mostly narratives about a car that stalled on the highway and then would not restart, plus EV charging failures. That produced exclusion (d) and the adjacency requirement in stage 3. Round 3 (the final 495): 30 read in full. Four or five were still misfits -- an acceleration-loss narrative, a Sirius radio complaint, a transmission complaint, and a couple where a no-start was mentioned in passing inside a long list of unrelated faults. **Treat every count in this file as carrying roughly one mis-sort in six to seven (about 15%).** We are publishing the number rather than smoothing it, because "car will not start" is one of the most overloaded phrases in the complaint file and no regular expression cleans it completely. -------------------------------------------------------------------------- STRUCTURED FIELDS AND THE BASELINE -------------------------------------------------------------------------- All 777,040 complaints: towed 5.11%, crash 4.37%, fire 1.82%, median mileage 70,000 (392,935 usable mileages). The 495 silent no-starts: towed 1.01%, crash 0.40%, fire 1.01%, 3 injuries, 0 deaths, median mileage 65,000 (302 usable mileages, first quartile 32,000, third quartile 105,000). Note the gap inside our own data: the structured tow field says 1.01%, while 127 of the 495 narratives (25.66%) contain the word "tow". The structured field is filled in far less often than the story is told. We report both and we do not average them. -------------------------------------------------------------------------- WHAT OWNERS NAME (silent-no-start-what-owners-name.csv) -------------------------------------------------------------------------- Case-insensitive searches over the whole narrative; overlapping, so the percentages do not sum to 100. These are counts of WORDS OWNERS WROTE, not of verified causes. Owners frequently repeat what a shop told them, and several narratives name a part that was replaced and did not fix anything. battery named anywhere 293 59.19% towed (narrative text) 127 25.66% names a recall or campaign number 123 24.85% battery replaced, or was new 114 23.03% dash, lights or radio worked (power present) 114 23.03% intermittent -- it sometimes starts 98 19.80% everything dead, no power at all 91 18.38% jump-start tried 78 15.76% starter or solenoid named 71 14.34% key fob, smart key or push-button named 56 11.31% battery tested good / was not the problem 50 10.10% alternator named 49 9.90% ignition switch, cylinder, EIS or steering lock 43 8.69% fuse, relay or fusible link 30 6.06% immobilizer, anti-theft or security system 23 4.65% shop could not find it, or no codes stored 22 4.44% battery cables, terminals, ground or corrosion 21 4.24% starter replaced, per the owner 20 4.04% TIPM, integrated power module, junction box or BCM 20 4.04% jump-start did not help 19 3.84% neutral safety, park/neutral, range or clutch switch 2 0.40% brake pedal switch 1 0.20% Two of these deserve a warning label. 1. Battery terminals, cables and corrosion appear in only 4.24% of narratives. That is NOT evidence that loose terminals are a rare cause. A terminal you tighten yourself in a driveway never becomes a federal safety complaint. The number measures what reaches NHTSA, and it cannot measure what a wrench fixed before breakfast. 2. The neutral safety / park-neutral switch appears twice in 495 complaints, and in 9 of 2,331 no-start bulletins. Same caveat, doubled: it is a well-understood, quickly diagnosed failure. We report the numbers because they are the numbers, and we do not convert them into a frequency claim in either direction. -------------------------------------------------------------------------- MANUFACTURER BULLETINS -------------------------------------------------------------------------- Manufacturer communications whose summary contains "no crank", "no-crank", "does not crank", "will not crank", "won't crank", "no start", "no-start", "will not start", "does not start", "fails to start" or "won't start", counted DISTINCT by NHTSA record id: 2,402. Removing heavy-truck, bus and stationary-generator manufacturers by name leaves 2,331. mentions battery 714 30.63% mentions software, reprogramming or calibration 566 24.28% mentions a ground, connector, corrosion, wiring or terminal 248 10.64% mentions a starter or solenoid 148 6.35% mentions an immobilizer, key, fob or security 130 5.58% mentions a body or power-distribution module 111 4.76% mentions a relay or fuse 89 3.82% mentions a neutral, park or clutch switch 9 0.39% By manufacturer, Jeep files the most (322), then Chevrolet (202), Ford (157), GMC (143) and Cadillac (109). That ranks documentation, not reliability. -------------------------------------------------------------------------- LIMITATIONS -------------------------------------------------------------------------- 1. About one in six to seven rows is a mis-sort (see the sample reading). This is the least clean set we publish, and the reason is linguistic, not statistical: "won't start" describes at least four different symptoms. 2. Complaints are self-reported and unverified. 3. There is no denominator. Nothing here ranks any vehicle's reliability. 4. Free fixes are systematically invisible. Anything an owner solved in a driveway for nothing -- a terminal, a gear selector, a fob battery -- is under-represented by construction. Read low percentages accordingly. 5. Mileage is owner-entered and missing in 193 of 495 rows. 6. The complaint file ends 8 September 2026. Contact: support@pulscar.io