PulsCar analysis of NHTSA public data, September 2026 Losing power on a climb: 1,154 owner complaints 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 (ODINO, product type V) received 1 January 2015 to 8 September 2026. NHTSA manufacturer communications (TSB) flat file FLAT_TSBS.zip. Manufacturer bulletins and Part 573 reports downloaded from static.nhtsa.gov and api.nhtsa.gov, listed in the article's sources section. NOTE ON THE EARLIER NUMBER IN THIS ARTICLE. A previous pass (12 September 2026) sampled ten high-volume nameplates, model years 2015-2022, through NHTSA's public API: 74,143 complaints, of which 177 matched. This pass replaces that with the whole complaint file -- every make, every model year, eleven years. Both are correct for their own scope. -------------------------------------------------------------------------- WHICH COMPLAINTS COUNT (the uphill power-loss set: 1,154) -------------------------------------------------------------------------- STEP 1 -- full-text search (FTS5) for a climb word: uphill | hill | hills | grade | incline | climbing | climb | "up a hill" | mountain. 15,267 complaints matched. STEP 2 -- a climb word must appear within 170 characters of a power-loss word, with the two in the same passage: CLIMB WORDS (regex): uphill | up hill | up a (steep) hill/grade/incline/ slope/mountain | going up a/the hill/grade/incline | incline | climbing (a/the) hill/grade/mountain | steep grade | on a grade | up the mountain | mountain pass POWER-LOSS WORDS: lost power | loss of power | losing power | loses power | no power | lack of power | wouldn't accelerate | will not accelerate | failed to accelerate | bogged down | bogs down | hesitat* | sluggish | limp mode | limp home | reduced (engine) power | fail-safe | struggled to accelerate/climb/maintain | couldn't accelerate/maintain/ climb | slowed down 1,220 complaints matched. 172 heavy-truck, bus, RV and motorcycle makes were dropped first. STEP 3 -- EXCLUSION. A passage does not count if the same 340-character window contains roll-back language -- "rolled back", "rolled backward", "rolls forward", "hill hold", "hill start assist" -- or the simile "like being on an incline". These are parking-brake, hill-hold and figure-of-speech complaints, not power loss on a climb. 66 dropped. 1,154 complaints remain. This is the denominator for every percentage. SAMPLE READING (required before publishing any keyword count) BEFORE the roll-back exclusion, 30 randomly drawn matches were read in full: 26 described losing power on a climb, 3 were roll-back or hill-hold complaints and 1 used "like going up a roller coaster" as a simile. FALSE-POSITIVE RATE BEFORE STEP 3: 4 IN 30, ROUGHLY 13%. AFTER the exclusion, a further 30 were read: 28 described losing power on a climb; one was a warning-light complaint that happened to include the words "uphill start assist disabled", and one explicitly said "I am not going up a hill". RESIDUAL FALSE-POSITIVE RATE: 2 IN 30, ROUGHLY 7%. -------------------------------------------------------------------------- RESULTS -------------------------------------------------------------------------- Outcome flags (share answering Y, of 1,154): crash 1.65%, fire 1.13%, towed 3.12%, 10 injuries, 0 deaths. All-complaint baseline (777,040): crash 4.37%, fire 1.82%, towed 5.11%. Mileage (MILES field, 0 < miles < 400,000; 709 complaints report one): median 66,000; quartiles 28,000 and 106,000; 29.6% at or under 36,000 -- that is, inside a typical bumper-to-bumper warranty. Baseline median 70,000. Speed (SPEED field, 0 < speed < 120; 859 complaints report one): median 50 mph; 50.8% at 46 mph or more; 27.1% at 61-75 mph. These are highway losses, not parking-lot ones. SYSTEM NAMED (uphill-power-loss-by-system.csv). A complaint counts under a system when its narrative contains one of that system's phrases anywhere; complaints naming two systems count under both, so the column does not sum to 1,154. transmission (incl. CVT) ........ 393 34.06% a recall is mentioned ........... 218 18.89% engine stalled or shut off ...... 210 18.20% limp / reduced power / fail-safe . 136 11.79% no fault found .................. 61 5.29% spark plugs, coils, misfire ..... 40 3.47% hybrid or electric propulsion ... 39 3.38% engine replaced/rebuilt/seized .. 39 3.38% throttle body or pedal sensor ... 39 3.38% turbocharger or boost ........... 36 3.12% fuel pump ....................... 32 2.77% catalytic converter ............. 30 2.60% air filter or MAF ............... 7 0.61% fuel filter ..................... 4 0.35% oxygen sensor or lean code ...... 3 0.26% THE FINDING, AND ITS LIMIT. The four "cheap fuel and air" causes that most pages put first -- fuel filter, air filter, MAF, oxygen sensor -- are named in 14 of 1,154 narratives, about 1.2%, while transmissions are named in more than a third. This does NOT mean a dirty air filter is a rare cause of a sluggish car. It means the two populations are different: NHTSA complaints are safety reports, filed when a car became dangerous in traffic, and a car that merely feels slow up a hill almost never generates one. The right reading is that the federal record measures the SUDDEN, DANGEROUS end of this symptom, and at that end it is dominated by transmissions, stalls and reduced-power modes. We publish it because it is the only measured evidence available, and because the widely repeated "about 80% of uphill power loss is fuel or air" figure has no source at all -- not a manufacturer document, not a federal file, not a published study. We removed that figure from this article rather than keep repeating it. MODEL CONCENTRATIONS (uphill-power-loss-top-models.csv, top 35 published) Ford Focus 47 complaints, median 52,000 miles, 63.8% name a transmission. Nissan Rogue 45, median 66,500, 57.8% name a transmission. Ford F-150 45, median 80,500, 20.0% name limp or reduced-power mode -- the highest limp-mode share in the top ten. Jeep Cherokee 31, median 22,000 miles, 77.4% name a transmission. Kia Soul 36, median 105,511 miles, 16.7% name a stall. Hyundai Tucson 34, median 52,450, 38.2% name a stall. -------------------------------------------------------------------------- LIMITATIONS -------------------------------------------------------------------------- 1. Complaints are self-reported, unverified, and not normalized by vehicles on the road. Counts by make or model are not reliability rankings and the article makes no such claim. 2. The system counts measure the words owners wrote, which often repeat a shop's verdict and sometimes a forum's. They are not confirmed diagnoses. 3. Selection bias is the dominant limitation and is stated above and in the article: this file records the dangerous end of the symptom. 4. SPEED and MILES are free-text-derived fields with many blanks; the denominators for those two statistics (859 and 709) are printed wherever they are used. ------------------------------------------------------------------------------ CASE SENSITIVITY — SECTION ADDED 19 SEPTEMBER 2026 ------------------------------------------------------------------------------ NHTSA stores complaint narratives in ALL CAPS for records filed through roughly 2020 and in mixed case for records filed from 2021 on. A case-SENSITIVE search of the narrative field loses about 92% of the complaints filed after 2020, so this is the single largest way an analysis of this database can go wrong while still looking finished. Every query behind this file is case-insensitive. SQLite's LIKE and its FTS5 full-text index both ignore case by default, and no query in this work uses GLOB or any case-sensitive comparison. This was not taken on trust. On 19 September 2026 every published folder was re-checked against the same database, comparing the figures here with what a case-sensitive query would have returned. The result: the complaint set published here (74,143) is LARGER than the absolute ceiling a case-sensitive search of the same terms could have returned (5,386 narratives). A case-sensitive query cannot produce a bigger set than itself, so the bug is ruled out for this folder by arithmetic, not by assurance.