Finding a load for a velocity

Most of the app works forward: pick a powder and charge, get a predicted velocity. The candidate explorer runs the model the other way — enter a cartridge, a bullet, and a target (or measured) velocity, and it surveys the powder-and-charge combinations that could reach it.

It is a candidate explorer, never a recommender. It answers one narrow question — which loads could the model reach this speed with? — and nothing more.

Velocity is not precision

Matching a velocity says nothing about how a load will group. How a load shoots depends on the powder, bullet, seating, brass, and your rifle — none of which a velocity can recover. A candidate that reaches your number is a starting point to investigate, not a load that will shoot well, and never one that has been checked in your firearm.

What every candidate carries

Each candidate is shown with the same honesty as a forward prediction:

• a modeled muzzle velocity with its interval — the spread, not just a point;
• %MAP — pressure as a percentage of SAAMI maximum, keyed to the modeled high end of the interval and shown neutrally (never green, never “safe”), with an over-MAP caution when that high end crosses the limit;
• case fill; and
• a calibration grade (Estimated → Provisional → Calibrated) — how much to trust that candidate’s numbers.

Candidates are ranked by model confidence, never by predicted accuracy. There is deliberately no “best”, no node, no precision score anywhere — the model cannot know which will group, so it never pretends to.

Using a candidate

Use candidate only seeds the forward workbench with that powder and charge — it never saves a load and never invents a chronograph number. From there you design the charge as usual: reduce, and work up from current published data. Nothing is committed until you choose to save it.

Develop a load

Starting from an ammo, Develop a load pre-fills the survey with that ammo’s cartridge, bullet, and target velocity, so the candidates are already aimed at the goal you set. Pick one and save it as a new load under that ammo — one more powder-and-charge to try toward the same target, alongside any loads already there. (One ammo carries many loads; see Ammo & firearms.)

If the ammo’s bullet is one you typed in by hand rather than a catalog or saved bullet, the app first offers to save it as a custom bullet — developing a load needs only the bullet’s diameter and weight, so a brand-new bullet works exactly like a catalog one. (Ballistic coefficient is for trajectory, not for finding a charge, so it stays optional.)

When nothing comes back

A survey can come back empty, and when it does the app tells you why rather than leaving you to guess:

• the bullet isn’t matched — the ammo names a bullet the library doesn’t have by that name; pick one, or save it as a custom bullet;
• the cartridge isn’t catalogued — reverse only explores cartridges in the catalog;
• the target is too high — reaching it drove the modeled pressure over SAAMI maximum; lower the target;
• the target is too low — the charge falls below the case-fill floor; raise the target;
• your filters excluded everything — loosen the temperature-stable, minimum-grade, or published-data filters.

It is the same honesty as everywhere else: an empty result that explains itself, and never a nudge toward a load that crosses the pressure limit.

Before you load any of them

Every candidate is a modeled hypothesis, not load data. Check each one against current published data for your components, and work up from a reduced charge while watching for pressure signs. The model cannot know your firearm, your brass, or your lot of powder — verifying that is yours to do.