Flat Spot or Fluke?
What your velocity ladder can really tell you, and how many shots per rung it needs
Download the full study (PDF)A velocity ladder loads a few charges in small steps and looks for a “flat spot”: a stretch where the speed stops climbing even though each rung adds powder. This study asked one question: when a ladder shows a flat spot, is it real, or is it ordinary shot-to-shot variation?
We took the shot-to-shot spread from 79 real Garmin Xero strings (89 sessions in the data file), then had a computer fire 20,000 ten-rung ladders with a real flat spot built in and 20,000 without, for every combination of rung size, SD and shots per rung, and read each one the common way: find the three rungs in a row whose speed changed least from first to last, and call it flat if the first and third are within 5 fps.
The answer in one table
Rungs about 10 fps apart, no drift during the session. The “real flat spot” is three perfectly level rungs, the easiest kind to find. Both columns use the same test, so read each pair side by side.
| Shots per rung | Very consistent load (SD 5) | Good load (SD 8) | Author's average (SD 11) | |||
|---|---|---|---|---|---|---|
| No flat spot, looks flat anyway | Real flat spot, picked and looks flat | No flat spot, looks flat anyway | Real flat spot, picked and looks flat | No flat spot, looks flat anyway | Real flat spot, picked and looks flat | |
| 1 shot | 12% | 71% | 51% | 54% | 64% | 42% |
| 3 shots | 0.1% | 83% | 9% | 73% | 32% | 63% |
| 5 shots | <0.1% | 90% | 1% | 78% | 12% | 71% |
| 10 shots | <0.1% | 98% | <0.1% | 86% | 0.9% | 78% |
A stricter reading, all three rungs within 5 fps of each other, gives the same answer for one shot per rung: at SD 8 a load with no flat spot looked flat 17% of the time, and a real flat spot was picked and looked flat 16% of the time.
What it means at the bench
- One shot per rung: on a good or average load, a load with no flat spot looked flat about as often as a load with one was read correctly (51% against 54% at SD 8; at SD 11 it was worse, 64% against 42%). Only a very consistent load did much better (12% fooled at SD 5).
- Three to five shots per rung make a ladder much harder to fool, but how much depends on your SD: at five shots, 1% at SD 8 and 12% at SD 11.
- A flat spot is a place to look, not an answer, at any number of shots. Check it separately: a longer strings at the pick and at the rungs on either side of it, or the whole ladder fired again on another day. Firing a one-shot ladder twice and keeping only a flat spot that showed up in the same place both times cut the false rate from 51% to 9% (SD 8), but it also read a real flat spot correctly less often (24% against 54%), and the model had no drift.
- Two rungs at the same speed is weak evidence. On a load with no flat spot, the two closest neighbouring rungs were within 5 fps in 85% of five-shot ladders (SD 8).
- Drift doesn't average out. A ladder is fired in order, so barrel heat, fouling and changing conditions are mixed in with the charge. With speed wandering an assumed 4 to 6 fps from rung to rung, even ten-shot ladders were fooled 9% to 29% of the time (SD 8, 10 fps rungs).
- Don't read a flat spot into one slow rung. At least one rung read slower than the one before in 91% of one-shot ladders, from ordinary variation alone.
Is a velocity flat spot real?
This study can't say whether your load has one. It says how often a ladder is fooled, which is a different question. Hornady, Applied Ballistics and Outdoor Life reported that, with larger samples, the flat spots their ladders showed mostly did not repeat; Applied Ballistics reported one statistically significant result, and experienced competitors report flat spots that hold up for them. If you have one you've confirmed across barrels and seasons, nothing here says it's a fluke.
Know your own numbers
- Your fps per rung: average speed of your top rung minus your bottom rung, divided by the number of steps between them.
- Your real SD: from a string of 10 or more shots of one load at a steady pace, not from one three-shot rung. The author's Garmin Xero figured SD the “divide by the number of shots” way, so a three-shot rung read about 18% low: SD 6.5 on that Garmin is about SD 8 in this study. Other units may differ.
FAQ
How many shots per rung does a velocity ladder need?
It depends on your shot-to-shot SD. With rungs about 10 fps apart, a load with no flat spot still looked flat 32% of the time at three shots per rung on an average load (SD 11), 12% at five and 0.9% at ten. On a good load (SD 8) it was 9% at three shots and 1% at five. Drift during the session can still fool a ladder at any shot count.
Is a one-shot-per-rung ladder reliable?
On a good or average load, no. At SD 8 with 10 fps rungs, a load with no flat spot showed one 51% of the time, about as often as a real flat spot was picked and looked flat (54%). Only a very consistent load (SD 5) did much better (12% fooled).
If my ladder shows a flat spot, is it real?
The ladder alone can't say. The study measured how often a ladder is fooled, not how likely a given flat spot is to be real. Treat a flat spot as a place to look, then check it separately: longer strings at the pick and at the rungs on either side of it, or the whole ladder fired again on another day.
Is a rung that reads slower than the one before a problem?
Not by itself. With rungs adding 10 fps, at least one rung read slower than the one before in 91% of one-shot ladders from ordinary shot-to-shot variation alone. This study is about speed only and says nothing about pressure; watch for pressure signs as your load manual describes.
The full study
The PDF has the method, every table, the sensitivity checks, the limits and the references: Flat Spot or Fluke? (PDF).
Disclosure: the author makes Loadscope, software that scores load development ladders. Loadscope gives velocity flat spot a 5% weight in its overall score by default (users can change it), a setting made in July 2026, before this study. No outside funding.