Why Your Quiver Picks the Worst Arrows (Selection Bias on the Bench)

Every serious archer I know numbers his arrows. Sharpie on the wrap, 1 through 12. And every one of them, within a month, has a verdict on each number. Three is money. Seven is cursed. Eleven "just doesn't group."

Here's the problem. Almost none of those verdicts are real.

You shot a 3-shot group at 40 yards. Arrow 7 landed two inches out. Two days later it did it again. Case closed, right? Seven gets demoted to foam duty and never sees the hunting quiver.

That's not tuning. That's superstition with better handwriting.

Twelve coins, three flips

Do the math on what you just did. Your honest group at 40 — all twelve arrows, your normal form — is maybe 4 inches. That means every single shot has a decent chance of landing 2 inches from center. You flipped each arrow three times and went looking for a pattern.

Flip twelve coins three times each. One of them is coming up tails all three. Every time. You wouldn't call that coin cursed. But you'll bench an $18 arrow on the same evidence and feel like a precision shooter for doing it.

With twelve arrows and tiny samples, some arrow HAS to be the worst. The bench guarantees a loser even when the dozen is perfectly matched. And once you've decided 7 is the bad one, every loose shot it takes confirms it, and every loose shot from 3 gets blamed on the wind. You built a story, and now the story is doing your scoring.

I proved this to a buddy last summer. He'd banished 4 and 9 from a dozen of his arrows — wouldn't even practice with them. One night I peeled the wraps and renumbered the whole dozen while he was inside. Two weeks later he'd culled two more numbers. Different shafts. Same confidence. Nothing changed but the Sharpie.

The bias runs the other way too

Here's the part that should bother you more. Come November, look at what actually goes in the quiver.

It's the pretty ones. Fresh fletchings, clean wraps, broadheads spun on the night before opener. The arrows with zero history. Meanwhile the arrows you trust most — the beat-up ones with 500 shots of proof behind them — stay home in the target block because they look rough.

You practice all summer with veterans and then hunt with strangers. The bench culls on noise, and the quiver selects on cosmetics. Both filters are pointed the wrong direction.

What real evidence looks like

Before any arrow gets a verdict, it earns one. That means three things.

First, nock tune it. A huge share of "bad arrows" are just arrows with the nock indexed to the wrong side of the spine. Levi Morgan nock-tunes every arrow in the dozen before he trusts any of them, and he'd tell you most fliers are un-tuned, not defective. Rotate the nock a quarter turn and shoot it again before you write anything off. I covered the full process in the nock tuning post — it's 20 minutes.

Second, spin test and check the squareness of the ends. A wobble or an out-of-square insert face is a real, fixable defect. That's a mechanical reason to bench an arrow. "It felt off twice" is not.

Third — and this is the one nobody does — get a sample size. Twenty shots per arrow, minimum, same distance, logged by arrow number, before you call one bad. Not three. Twenty. If an arrow is still printing outside the others after twenty logged shots and a nock tune, congratulations, you found a real cull. That's rare. Usually you find out your dozen is more consistent than you are, which is its own useful truth.

Let the log pick the quiver

This is exactly why I built shot logging into The Forge the way I did. The Range tab lets you tag every logged shot with the arrow number, so over a few weeks you accumulate an actual record per shaft instead of a hunch. And the Shootout feature — where you pit builds head-to-head on group data — has a significance gate baked in. If you haven't logged enough shots, it refuses to declare a winner. People have emailed me annoyed about that. Good. The whole point is that the tool won't let you do the thing you've been doing on the bench for years: calling a verdict on three flips of a coin.

So here's the job. Renumber your dozen tonight — fresh start, no priors. Nock tune each one. Then log twenty shots per arrow in the Range tab in The Forge over your next few sessions and let the numbers pick your hunting six.

My guess: the arrow you hated makes the cut, and one you never suspected doesn't. The data doesn't care about your Sharpie.

Keep reading: for the full picture, see the complete arrow building system — and build a hunting arrow under $100, where to spend on components.