A tuning guide · West Covina, CA

The Truth About Dyno Numbers — Why the Same Car Reads Different

Two shops can measure the same car on the same day and land forty horsepower apart, and neither one is lying. Here's what the dyno is actually telling you, which number to trust, and how shops make the graph say whatever they want.

// Two shops, same car, same day, forty horsepower apart — and neither one lied. Here's what actually happened, and why the number you brag about is the least useful thing the dyno told us.

MUSTANG ~12% under Dynojet HUB ~8–10% over rollers STD ~4% over SAE THE delta, not the peak
The number everyone argues about

Two Shops, Same Car, Forty Horsepower Apart

Two shops, same car, same day, forty horsepower apart — and neither one lied. Here's what actually happened, and why the number you brag about is the least useful thing the dyno told us. The dyno is the most useful tool in the shop and the most misunderstood number in the hobby, and almost every argument about it comes from treating one shop's peak figure as if it were an absolute, universal truth.

It isn't. No chassis dyno on earth measures horsepower directly — every one of them measures speed and either acceleration or force, calculates torque, and derives horsepower from it. Change the machine, the correction standard, the gear, or even the temperature of the room, and the same engine making the exact same power prints a different number. That's not fraud; it's physics. A Mustang dyno reading lower than a Dynojet isn't a weaker car, and a hub dyno reading higher isn't a stronger one.

So the honest way to use a dyno is the opposite of how the internet uses it. The absolute peak number — the one people put in their signature and shout across forums — is the least trustworthy thing the machine produces. The gain between a real baseline and an after pull, on the same dyno, same day, same conditions, is the number that actually means something. This page is about telling those two apart, so nobody sells you a graph instead of a result.

The three machines, compared

The Three Dyno Types — Inertia, Load & Hub

"Dyno" isn't one machine. There are three fundamentally different architectures, and each one produces a different number for the identical car — not because one is right and the others wrong, but because they measure the car in genuinely different ways.

The US standard

Inertia roller (Dynojet)

A single large drum spans the full wheelbase; your wheels accelerate it, and the dyno back-calculates power from the drum's known mass and how fast it spins up. It's mechanically simple, repeatable, and everywhere — which is why it's the number most people compare against. Its blind spot: it can't see the power spent spinning up your own drivetrain and wheels.

⤢ Click to enlarge
Reads ~12% lower

Eddy-current load (Mustang)

Adds a magnetic brake the operator can dial up to apply real, controlled load, and measures torque directly through a load cell. Usually twin rollers with much lower rotating mass, so the absorber — not drum physics — controls the reading. It reads about 12% lower than a Dynojet by design, and it's the machine you actually want for part-throttle and diagnostic tuning.

⤢ Click to enlarge
Reads ~8–10% higher

Hub dyno (Mainline, Dynapack)

The wheels come off and the dyno bolts straight to the hubs — no roller, no tire slip. Because it never has to spin up the weight of your wheels and tires, it reads roughly 8–10% higher than a roller, often 30–40 horsepower on a street car. It's the honest choice on very high-power builds where roller tire-slip gets unmanageable.

⤢ Click to enlarge

None of the three is the "true" number — they're different instruments measuring at different points in the driveline. Which is exactly why I only ever compare a before and after on the loaded dyno in my own shop, never your number against someone else's machine.

Where the confusion comes from

Why the Same Car Reads Different Numbers

Once you know there are three machines, most of the "why is my number different" mystery dissolves — but the machine type is only the first of several variables, and the rest are just as real. Even on the identical dyno, the same car can read differently from one pull to the next.

The machine and the driveline. A Mustang reads ~12% under a Dynojet; a hub dyno reads ~8–10% over a roller. On top of that, a roller dyno loses about 8% of the power just to tire-to-roller slip, and that loss grows as the car makes more torque. So a 400-horsepower Dynojet number, a 352 on a Mustang, and a 440 on a hub can all be the exact same car — three instruments, three honest readings, one engine.

The gear, the sweep, and the weather. An inertia dyno can't see the power spent accelerating your own flywheel, driveshaft and wheels, so a taller gear — 6th instead of 4th — slows that acceleration, shrinks the hidden "inertia tax," and inflates the number without the engine doing anything different. On a load or hub dyno, a slow, heavily-loaded sweep reads higher than a fast one. And air temperature, pressure and humidity move the number every hour of the day. Understanding this is the same discipline behind why a tune sometimes makes no power — you have to control the variables before you trust the result.

A Los Angeles owner's guide

How to Read a Dyno Sheet — A Los Angeles Owner's Guide

Reading a dyno sheet honestly is four questions. Ask them and no shop can sell you a graph instead of a gain.

  1. Question 1 of 4

    What machine, and what correction?

    Before the number means anything, ask what dyno it came from — inertia, load, or hub — and whether it's SAE, STD, or uncorrected. A Dynojet SAE number and a hub-dyno STD number are not the same unit, and comparing them is meaningless. If a shop can't tell you both, the number is decoration, not data.

  2. Question 2 of 4

    Is there a real baseline?

    The only trustworthy figure is the delta — the gain between a before pull and an after pull, on the same machine, same day, same gear, same tires. If a shop hands you an "after" number with no baseline, there's no honest way to know what the work actually did. Always demand the before run, not just the money shot.

  3. Question 3 of 4

    Are the conditions on the chart?

    A real dyno sheet logs ambient air temperature, barometric pressure, inlet air temperature and air-fuel ratio for both runs. Without those, the power numbers are unverifiable — you can't tell heat soak from a real change. I want those columns on every graph I hand you, because they're what make the number honest.

  4. Question 4 of 4

    Does the real world back it up?

    Dyno math can be gamed; physics on the road can't. A big claimed gain should show up in a trap speed or a Dragy or VBOX run. If the dyno sheet screams and the real-world times don't move, the dyno was telling a story. The best validation of a number is that the car is actually faster.

Question 1 / 4
Real LA price bands

What Honest Dyno Time Costs in Los Angeles

Dyno time itself isn't expensive — what you're really paying for is the tuner reading the data correctly and making safe, repeatable changes. Here are the honest LA bands in 2026. Suspiciously cheap "dyno tuning" usually means a canned map and a couple of pulls, not real calibration; the number below buys the discipline, not just the rollers.

Baseline & health check

$150–300
~1 hour in shop

A few logged pulls to read your car's real power, AFR and knock — the honest starting point before any work.

  • Logged pulls
  • AFR & knock read
  • Printed baseline
⤢ Click to enlarge
The real work

Custom dyno tune

$600–1,200
~1 day on the dyno

A full loaded-dyno calibration for your hardware and fuel, logged and verified — with a real before/after delta.

  • Loaded-dyno tune
  • Baseline + after
  • Full data logged
⤢ Click to enlarge

Flex-fuel / E85 tune

$900–1,400
~1–2 days in shop

A dyno tune calibrated across the fuel range, verified on the worst realistic blend and LA heat.

  • Blend-range map
  • Heat-verified
  • Logged & safe
⤢ Click to enlarge

Datalog revision

$150–400
~half day in shop

A follow-up revision from real-world logs after a new part or a season change — dialed on the same dyno.

  • Log review
  • Targeted revision
  • Re-verified
⤢ Click to enlarge

What moves your number: your platform, your power level, and whether you want a full calibration or just a health check. Whatever the job, you leave with a baseline, an after, and the logged conditions that prove the gain is real.

BOOK A DYNO TUNE
Correction factors & the fine print

Dyno Correction Factors — SAE, STD & Uncorrected

Correction factors adjust the raw number for the day's air, so pulls taken in different weather can be compared. There are three you'll see, and which one is right depends on your engine.

SAE versus STD. SAE (J1349) corrects to 77°F, 29.23 inHg and dry air — the industry standard, closest to manufacturer specs. STD (J607) corrects to a cooler, denser 60°F reference, which makes it read about 4% higher for the exact same pull. Neither is fake, but plenty of shops publish STD specifically because it's the bigger number and it makes the money feel better spent. Always ask which one you're looking at before you compare it to anything.

Why turbo cars want uncorrected. Here's the part almost nobody explains: SAE and STD were built for naturally-aspirated engines. A modern turbo ECU already self-corrects for air density in real time, adjusting boost, timing and fuel on its own. Layering SAE or STD on top of that "double dips" the correction and actually makes the graph less consistent. For nearly every turbo platform I tune, the uncorrected number — logged with the day's real conditions noted beside it — is the honest one.

The 5252 rule. One thing that's always true: horsepower equals torque times RPM divided by 5252, which is why every honest dyno graph shows the power and torque curves crossing at 5252 RPM. If they don't cross there on a chart plotted in those units, something about the graph has been massaged.

Mustang 352 Dynojet 400 Hub dyno 440 same car, same day, three machines (whp) →
Dynojet (US baseline) Load / hub // compare deltas, not machines
Which number to trust, by engine type

NA vs Turbo — Which Correction Your Car Actually Wants

The right dyno number isn't a universal answer — it splits by how your engine handles the weather, not by what badge it wears.

Naturally-aspirated cars want correction. An NA engine makes exactly the power the day's air density gives it — no boost controller compensating. That's precisely the case SAE correction was built for: it standardizes the number so a pull on a cold December morning and a hot August afternoon can be compared honestly. On an NA build, an SAE-corrected figure is the right one to trust and the right one to tune against.

Turbo cars mostly want uncorrected. Almost every platform I build is turbocharged — and a modern turbo ECU already adjusts boost, timing and fuel for air density in real time, so it self-corrects before the dyno ever applies a factor. Correcting on top of that double-counts the adjustment. For a turbo car, I read the uncorrected number and log the actual conditions beside it, which is the same honesty behind matching a custom ECU calibration to what the car really sees. It's also why the stage a shop quotes only means something with real numbers behind it — the same discipline as knowing what the stage numbers actually mean.

How the graph gets gamed

5 Dyno Tricks Shops Use — And How I Do It Differently

A dyno graph is easy to manipulate if you know the levers. These are the five I see most, and how I keep the number honest:

How I do it differently

1. Publishing STD to inflate the number

STD correction reads about 4% higher than SAE for the identical pull, and shops lean on it because the bigger number makes the money feel well spent. I tell you which standard you're looking at — SAE for NA, uncorrected for turbo — so the number isn't quietly padded.

How I do it differently

2. Zeroing the smoothing factor

Set a graph's smoothing to zero and a brief false spike in the raw data displays as your "peak" horsepower. I run a smoothing factor of five, the way it's meant to be read, so the peak is a real reading and not a one-sample blip.

How I do it differently

3. Running a taller gear for a bigger number

A pull in 6th instead of 4th slows the engine's acceleration, shrinks the inertia tax an inertia dyno can't see, and inflates the number — sometimes marketed as the "true 1:1 gear." I run the same sensible gear on every pull so the before and after are actually comparable.

How I do it differently

4. Faking a gain with pull order

Baseline the car hot so catalyst protection is pulling timing, swap a part while it cools, then pull again once the protection lets go — the number jumps and the part gets the credit it didn't earn. I control temperature and pull order, and I log the conditions so a real change can't hide behind heat soak.

How I do it differently

5. Padding the correction or the sensor

An operator can manually inflate the Total Correction Factor, or park the weather sensor near a hot engine so the software corrects upward. I keep the sensor honest and the correction appropriate to the engine — and I'll back the number against a trap speed if you ever doubt it.

Why LA makes this worse

Dyno Numbers in Los Angeles — Heat, 91 Octane & Real-World Truth

LA conditions make the dyno-number game both more tempting and more consequential. The heat and the 91-octane cap mean the honest number moves more here than in a cooler, higher-octane market, and the difference between a marketing peak and a repeatable result matters more when you actually drive the car hard on these roads.

Heat soak is a bigger factor here. A car that makes a clean, repeatable pull at 72°F loses consistency once the dyno cell heat-soaks to 95°F on a summer afternoon, and back-to-back pulls in LA heat can drop power just from rising inlet temps and catalyst protection. A shop chasing the highest possible peak will pull the number early and cold; I care more that the tune holds up on the third pull, because that's the pull that matches your car sitting in August traffic on the 10.

The road is the final judge. This is where the real-world validation matters most: a Dragy or VBOX run, or a trap speed, cuts through every dyno argument. On these roads a car either pulls or it doesn't, regardless of what correction factor made the graph look good. I'd rather hand you a slightly lower number that repeats every pull and shows up in your trap speed than a hero peak you'll never see again — because the number is only worth what the car actually does when you drive it.

Baseline, tune, verify

How I Baseline, Tune & Verify on the Dyno

Every dyno session I run follows the same disciplined arc — the number is a result of the process, never the point of it.

  1. Step 1 / 5

    Warm it up and run a clean-out pull

    I bring the engine to proper operating temperature first — a cold engine can read ten horsepower low — and treat the very first pull as a clean-out, not data. A car that's not warmed up or a first stab at the drum tells you nothing reliable, so I never diagnose off run one.

  2. Step 2 / 5

    Establish a real baseline

    Then I take honest baseline pulls — same gear, same tire pressure, conditions logged — so there's a true before to measure against. On the road I'll confirm two pulls land within a few horsepower of each other. Without a real baseline, no "gain" I report later would mean anything.

  3. Step 3 / 5

    Tune under real load

    I calibrate on the loaded dyno so I can hold the engine at part-throttle and specific RPM, not just wide-open — which is where a car actually lives on the street. Fuel, timing and boost get dialed for your hardware and your fuel, watching AFR and knock the whole time, never chasing a peak at the cost of safety.

  4. Step 4 / 5

    Verify across repeated pulls

    A number only counts if it repeats. I run the tune through back-to-back pulls in real heat to confirm it holds up when the cell — and your engine bay — are hot, not just on a cold first stab. The tune isn't finished until the third pull looks like the first.

  5. Step 5 / 5

    Hand you the honest picture

    You leave with the baseline, the after, the logged conditions, and the correction standard spelled out — plus an honest read on what the car will do in the real world. A dyno sheet from me is a record you can trust and verify, not a marketing poster with one big number on it.

Step 1 / 5
Questions, answered

Dyno Numbers Questions, Answered

Why did two shops give me completely different horsepower numbers for the same car?
Different dyno types measure and calculate power differently. A Mustang or other eddy-current load dyno reads roughly 12% lower than a Dynojet inertia dyno, and a hub dyno — which removes the wheels entirely and bolts to the hubs — reads roughly 8 to 10% higher than either, because it never has to spin up the weight of your wheels and tires (typically 30 to 40 horsepower on a street car). None of the three numbers is "wrong"; they're measuring at different points and in different ways. That's exactly why you can never compare your number to another car on a different dyno — only a before-and-after on the same machine tells the truth.
What's the difference between SAE and STD correction on a dyno printout?
Both are legitimate correction standards that adjust the raw reading for the day's air temperature, pressure and humidity — they just correct to different reference conditions. SAE (J1349) corrects to 77°F and is the industry standard, closest to what manufacturers quote. STD (J607) corrects to a cooler, denser 60°F reference, which reliably makes the number read about 4% higher for the identical pull. Neither is fake, but a lot of shops publish STD specifically because it's the bigger figure. Always ask which standard a number uses before you compare it to anything — and for a modern turbo car, ask whether it should be corrected at all.
Should my turbo car's dyno number be SAE corrected, STD corrected, or neither?
Neither, honestly — for a modern turbocharged car. SAE and STD correction were designed for naturally-aspirated engines, to compare pulls taken in different weather. A modern turbo ECU already adjusts boost, timing and fuel for the day's air density on its own, so it self-corrects before the dyno ever applies a factor. Layering SAE or STD on top double-counts that adjustment and can actually make the graph less consistent from pull to pull. For a turbo build, the uncorrected number — logged with the day's actual air temperature and pressure noted alongside it — is the most accurate picture of what the car really made.
Why does my car make less power on the third or fourth back-to-back pull?
Heat soak, and it's normal. As you run repeated pulls, inlet air temperatures climb and air density drops, so the same tune makes a little less power. On top of that, if the catalytic converters get hot enough, the ECU deliberately richens the fuel mixture to protect them, which costs power on that pull. It's thermal management, not the car "getting tired." It's also why pull order can be used to fake a gain — baseline hot, let it cool during a part swap, and the next pull reads higher for reasons that have nothing to do with the part. I control temperature and log conditions so real changes can't hide behind heat.
Can changing gears on the dyno actually change my horsepower number?
Yes, on an inertia dyno specifically, and it's one of the oldest tricks around. The dyno can't measure the power spent accelerating your car's own drivetrain — flywheel, driveshaft, wheels — so a taller gear makes the engine accelerate more slowly, shrinks that hidden inertia penalty, and inflates the reported number even though the engine made the exact same power. A shop running a fast car in 6th and calling it "the true 1:1 pull" is using this, not doing you a favor. A reputable tuner picks one sensible gear and runs every pull in it, so the before and after are genuinely comparable.
How do I know if a dyno graph has been manipulated?
Ask three things and watch how the shop answers. First, what smoothing factor was used — it should be around five, not zero, because zeroing it lets a random one-sample spike display as your "peak." Second, is the number SAE, STD, or uncorrected, and is that appropriate for your engine. Third, are ambient temperature, barometric pressure, inlet air temperature and air-fuel ratio logged on the chart for both runs — without those, the number is unverifiable. And the ultimate check is the real world: if a big claimed gain doesn't show up in a trap speed or a Dragy run, the dyno was telling a story the road won't back up.
Where I serve

Honest Dyno Tuning Across Greater Los Angeles, CA

My shop and dyno are in West Covina, in the San Gabriel Valley. Owners bring me their cars from the near ring, the mid ring and the South Bay for dyno work that comes with a real baseline, logged conditions, and a number they can actually verify. Tap your city:

The tools behind an honest number

Brands We Trust

An honest dyno number depends on honest tools. These are the dynos, tuning platforms and logging hardware I rely on to read a car correctly — chosen because they produce repeatable, verifiable data, not because there's a poster on the wall.

Dynojet inertia dyno Mustang load dyno SuperFlow dyno EcuTek calibration COBB Accessport HP Tuners logging Innovate wideband Zeitronix data logging Dragy real-world

// A number you can verify beats a number you can brag about.

Let's put a real number on your car

Tell me your car and your goal. You'll get a real baseline, a logged after, the correction standard spelled out, and a tune that holds up on the third pull and on the road — not just a hero peak.