Our approach
Ballistics Board exists to make factory ammunition comparable. Every figure on the site is either a number the manufacturer published or one derived from published numbers by a formula shown below. Where a manufacturer doesn’t publish something, we say so rather than fill the gap with a guess.
This page documents how that works. Ballistics Board is a comparison and education tool — it is not reloading data, and nothing here should be used to assemble ammunition. See What Ballistics Board is not.
Where the numbers come from
The load database is built from factory-load specifications published by the ammunition manufacturers — muzzle velocity, energy, bullet weight, test-barrel length, and so on, as listed on their product pages and catalogs. Each load is entered and reviewed by hand.
A “factory load” is a complete cartridge sold ready to shoot. That is distinct from a custom handload, which you can build yourself in the Custom Load Builder using your own components and chronograph data.
Caliber and bullet diameter
A cartridge doesn’t use a single bullet diameter — it uses a small family of them that shooters treat as interchangeable. A .45 ACP, for instance, is loaded with both .451″ and .452″ bullets. So when you pick a projectile for a load, the list shows every diameter in that caliber’s family, and each row is labeled with its exact diameter so you can tell them apart.
Those diameters are grouped from a hand-curated table, not by rounding to a tolerance. Real bullet diameters sit too close together for arithmetic to separate them safely: .308″ and .311″ are only three thousandths of an inch apart, but they belong to different cartridges — .308 Winchester and .303 British — and their bullets are not interchangeable. A numeric “close enough” rule would mix them, so the table only groups diameters that are genuinely the same caliber.
When velocity isn’t published
Not every manufacturer publishes a muzzle velocity for every load. When one doesn’t, Ballistics Board shows “Not published” instead of inventing a figure. Those loads stay in the database — you can browse them, search them, and see their bullet and box details — but they are held out of anything that needs a velocity to be meaningful: side-by-side comparison, the ballistics solver, and sorting by velocity or energy.
We don’t estimate the missing number. Muzzle velocity is a product of the maker’s specific powder charge and test barrel, not something you can infer from the cartridge and bullet weight — the same cartridge and bullet can vary by more than 150 fps between manufacturers, which is over a minute of angle at 500 yards. A guessed velocity that someone then dopes a rifle from would be worse than no number at all.
Test barrel length
A muzzle velocity is only half a fact without the barrel it was measured in. Velocity climbs with barrel length — the same load can differ by well over 100 fps between a 20″ and a 24″ barrel — so comparing a figure from one against a figure from the other is partly comparing barrels, not loads.
Where a manufacturer publishes the test barrel behind a velocity, we record it and show it beside that velocity. Where they don’t, we show the cartridge’s typical test barrel length marked with a tilde — ~24″ — to make clear it’s a general figure for the cartridge and not a specification for that load. Where we have neither, the field is blank. When a comparison puts loads from different barrels side by side, we say so underneath it rather than let the velocities read as like-for-like.
We don’t fill the gap with a house standard. There is no single industry test barrel to fall back on: SAAMI specifies a length per cartridge, handgun cartridges are tested in 4″ and 5″ barrels, .300 Blackout in 16″, and manufacturers often publish from their own barrels rather than a SAAMI one at all. Our own data shows it plainly — among the loads that do record a barrel, .308 Winchester appears at both 20″ and 24″, and .300 Winchester Magnum at 24″ and 26″. Assuming a number here would quietly overwrite that variation with a figure nobody published.
When ballistic coefficient isn’t published
The same rule applies to ballistic coefficient, and for the same reason. BC is what determines how fast a bullet sheds speed, how far it drops and how far the wind pushes it — the two numbers a shooter actually dopes from. Where a manufacturer publishes none, Ballistics Board shows “Not published” and computes no trajectory. Those loads stay browsable and searchable, with their velocity, weight and box details intact, but they are held out of comparison and the solver.
This affects handgun and rimfire ammunition most. Makers of defensive handgun loads generally don’t publish a BC, because it isn’t a meaningful figure at the distances those rounds are used at — and a component BC often doesn’t exist for a proprietary bullet you can only buy already loaded. That is a real gap in what is published, not an oversight in our data, and we would rather show the gap than paper over it with an assumed number that would put a confident-looking curve on the screen and be wrong.
When two sources publish different BCs
The same bullet often carries two different ballistic coefficients — one from the company that makes the projectile, another from the company that loads it into a cartridge. Swift lists its 6.5mm 130-grain Scirocco at .571; Remington, loading that identical bullet, lists .547. Neither is a misprint.
A published G1 BC isn’t a fixed property of a bullet. It is a curve fit, and the curve is a flat-based artillery shell profile from the 1880s that no modern hunting bullet resembles. Force a boat-tail spitzer’s real drag onto that shape and the best-fit numberdrifts with velocity — so a single published figure is really a figure for one part of the speed range. Some makers say so outright: Sierra publishes banded BCs, a different value for each velocity band of the same bullet. Two more things move the number. BCs are referenced to a standard atmosphere, and the older Army Standard Metro yields a value roughly two percent above the ICAO standard for the identical bullet. And a figure measured by Doppler radar, one measured over chronograph screens, and one back-fitted from observed drop will not agree — the last also absorbs any error in the velocity measurement it was fitted to.
So we keep both. The bullet maker’s figure sits on the projectile, because it describes the bullet. The loader’s figure sits on their load, because it was fitted to that cartridge out of that barrel and predicts that round better. A comparison across manufacturers then lines up on like-for-like projectile data without discarding the number that fits each individual round.
It is worth knowing which direction the disagreement runs, because it isn’t consistent. Across the Scirocco loads we hold, Remington reads lower than Swift on four bullets and higher on two. That pattern rules out the easy explanation that bullet makers simply publish optimistic numbers, and points at velocity band and measurement method instead. This is also the strongest practical argument for G7: for a boat-tail bullet a G7 BC stays far steadier across the velocity range, and the spread between sources largely collapses.
Which BC we actually solve with
Keeping several figures raises an obvious question: when a bullet has a G1 and a G7, and the company loading it publishes a third number, which one does the chart use? The rule is short, and it never varies by cartridge, manufacturer, or which page you came in on.
Where a load’s own maker publishes a BC for that specific round, we use theirs. They measured that bullet, at that velocity, out of that barrel. If they publish both a G1 and a G7 for it, we take the G7.
Otherwise we use the projectile maker’s figure, and where they publish both, we take the G7. That follows from the section above: a G1 BC is fitted against a flat-based 1880s shape, so the number drifts with velocity and a single published figure really only describes one part of the speed range. G7 is fitted against a boat-tail profile — the actual shape of the bullets anyone publishes a G7 for — so it stays far steadier across the range a trajectory is solved over. A manufacturer who publishes a G7 has done that work, and it is the better number.
We never mix the two. A BC only means anything paired with the drag curve it was fitted against, so we never take a G1 from one source and a G7 from another, or pair a loader’s number with the bullet maker’s. Whichever source answers, answers with both halves. Every bullet page names the model that was used, so you can always see which curve is behind the figure.
One consequence worth stating plainly, because it is a limit rather than a feature: this means we cannot choose G1 for a bullet whose maker published a G7. We think that is right — a G7 BC is only meaningful for a boat-tail, so publishing one is itself a statement about the bullet’s shape. If a manufacturer ever publishes a G7 for a flat-base projectile, that figure would be the problem, and we would rather not carry it than quietly solve around it.
Energy and sectional density are calculated
Muzzle energy and sectional density aren’t stored — they are calculated from their inputs every time they are shown, so they can never drift out of step with the velocity, weight, and diameter they come from.
- Muzzle energy (ft-lbs) = bullet weight (grains) × velocity (fps)² ÷ 450,436
- Sectional density = bullet weight (grains) ÷ (7,000 × bullet diameter (inches)²)
The energy constant, 450,436, is the value the ammunition industry uses — it folds in the grains-to-pounds and gravity conversions — and we verified it against thousands of loads that publish both velocity and energy.
A calculated energy can sit a fraction of a percent off a manufacturer’s own published figure, because makers round the velocity they publish and some use a slightly different energy constant. The gap is far too small to affect a comparison. Sectional density, which depends only on bullet weight and diameter, matches published values to the printed digit.
The ballistics solver
Trajectories are computed by numerically integrating the projectile’s flight (a fourth-order Runge–Kutta method), using the G1 or G7 drag model — standard drag curves published by JBM Ballistics — according to the bullet’s ballistic coefficient. Atmosphere follows the ICAO standard model.
Unless you change them, the solver uses these reference conditions:
- Sea level, 59°F, 29.92 inHg, 0% humidity
- 100-yard zero, 1.5″ sight height
- 10 mph full-value (90°) crosswind
It calculates drop, retained velocity and energy, time of flight, and wind drift (using the standard lag-time approximation). It does not model spin drift or the Coriolis effect — corrections that matter only at long range and that would demand far more precise inputs than a comparison tool has. Treat the output as a high-quality comparison and planning aid, and confirm your own dope on paper.
How recoil is estimated
Free recoil is conservation of momentum: the rifle’s kick equals the momentum of everything that leaves the muzzle — the bullet and the powder gas. It therefore depends on bullet weight, muzzle velocity, powder charge, and rifle weight (we default to an 8-pound rifle and a 4,700 fps gas velocity).
Manufacturers don’t publish the powder charge, so we estimate it from the cartridge’s case capacity — and case capacity here always means grains of water (the volume a fired case holds, measured with water), never a powder-charge weight. Recoil is shown only for loads whose cartridge has a water-capacity figure on file; where we don’t have one, recoil is left blank rather than guessed. Treat the result as a sound estimate for comparing rifle loads against one another, not an exact measurement.
We don’t estimate recoil for handgun cartridges. The method above is built around a rifle: it assumes the case is nearly full of powder, and it assumes an eight-pound gun. Neither holds for a handgun. A straight-wall pistol case runs a fraction full — a 9mm burns roughly a quarter of what the same volume of rifle powder would be — and a pistol weighs a third of what a rifle does, which matters because recoil energy rises as gun weight falls. Those two errors push in opposite directions and partly cancel, which is exactly what makes the result untrustworthy: it looks reasonable and isn’t. Felt recoil in a handgun also depends heavily on grip, action type and bore height, none of which a free-recoil calculation accounts for. So for handgun loads we show no recoil figure at all rather than one we can’t stand behind.
How the popularity figures are measured
Load pages carry a “compared n×” figure and the home page shows what is trending. Those come from our own counting, so here is exactly what is and is not behind them.
What is counted. When a factory load appears in a comparison, we add one to a tally for that load on that day. That is the whole record — a load, a date, a number. We do not store who compared it, or when during the day, or anything that could be traced back to a person; there is no row belonging to any individual. Custom handloads are never counted at all.
What is excluded, and why it matters. Links we wrote ourselves — the comparisons featured on the home page, and the “run this comparison” links in articles — are tracked separately and left out of every published figure. They have to be. Featuring five loads on the front page sends traffic to those five loads, and counting that traffic would produce a chart of our own editorial choices wearing the costume of a popularity ranking. The same goes for the crawlers that index the site.
Where the numbers stop. Days are counted in UTC, so a comparison run late in a US evening lands on the following day; this is invisible at the monthly resolution any trend uses, but it is the boundary we use. Small totals are withheld rather than published, because a thin enough slice stops describing a trend and starts describing one person’s afternoon. Every counter on the site was reset to zero when it launched in August 2026, so nothing you see was accumulated under looser rules than these — there is no residue of our own testing or our own featured links sitting underneath the numbers.
None of this measures sales, and it is not a survey. It measures what visitors to one website chose to put side by side, which is a real signal about what people are researching and a poor proxy for what they bought.
What Ballistics Board is not
Ballistics Board is not load data. It does not tell you what powder, primer, or charge to use, and its factory-load figures are not a substitute for a current, published reloading manual. Always verify any figure against the manufacturer before you rely on it, and follow established safety practice when handling ammunition and firearms. The tool is provided for comparison and education.