How to Improve Your Average of 5 (Ao5)
What Ao5, Ao12, Ao100, mean, and standard deviation actually measure, why averages matter more than single PBs, and practical ways to bring your Ao5 down.
By Cubeathon Team
A lucky single is a story. An average is the truth. If you want to know how fast you actually are — and get faster — the number to watch is your Ao5.
What the statistics actually mean
Ao5 (Average of 5). Take your last five solves, drop the best and the worst, and average the middle three. This is the WCA's official format, and it's designed to filter luck in both directions: one lucky skip can't inflate it, one pop can't ruin it. One DNF counts as your worst solve; two DNFs make the whole average a DNF.
Ao12 and Ao100. Same idea, bigger windows — Ao12 drops the best and worst of twelve; Ao100 drops the best five and worst five. Ao5 tells you about today. Ao12 tells you about this week. Ao100 tells you who you actually are as a cuber right now.
Mo3 (Mean of 3). A plain average of three solves with nothing dropped — used officially for events like 6×6, 7×7, and blind.
Mean. The plain average of every solve in your session, no trimming. Honest, but sensitive to disasters.
Standard deviation (σ). How consistent you are. Two cubers can both average 20 seconds — one solving 19–21 every time, one swinging between 14 and 26. The first cuber's next PB average is much closer. A shrinking σ is progress, even when your mean stands still.
PB. Your best single. Fun to chase, terrible to train by — it mostly measures luck density.
Why averages beat singles
Improving your PB single by two seconds might take one lucky scramble tonight. Improving your Ao5 by two seconds requires your typical solve to genuinely improve. That's why competitions rank by average, and why serious practice does too.
How to actually bring your Ao5 down
- Kill the disasters first. An Ao5 forgives one bad solve — not two. Most bloated averages come from panic: a mis-scramble, a wrong algorithm, a lockup you fought instead of breathing through. Consistency lowers averages faster than raw speed does.
- Slow down to speed up. Do sessions 10–20% below your comfort speed with zero pauses. Averages are made in the transitions between steps, not in turning speed.
- Scramble correctly. A botched scramble ruins the solve and the data. If your cube ever doesn't match the preview, use the scramble guide to replay the scramble and find exactly where you diverged.
- Watch σ, not just the average. If your standard deviation is more than ~15% of your mean, your gains are hiding in consistency, not speed.
- Use inspection fully. Fifteen seconds is enough to plan the entire cross (and your first pair, with practice). Rushed inspection is the most common self-inflicted +2 there is.
- Track honestly. Take your +2s, count your DNFs. Statistics only help if they're true.
Milestones worth celebrating
Your first completed Ao5. Your first Ao12. The first time your worst solve in an Ao5 would have been your PB a month earlier — that one sneaks up on you.
Cubeathon tracks all of this automatically — Ao5, Ao12, Ao100, Mo3, mean, σ, best averages, and today's and lifetime PBs, computed the official WCA way, saved on your device. Do five solves and you'll have your first real number to beat.