660 lb works out to 6x45 + 1x35 + 1x5 per side plus the bar — call it 665 lb in practice, since gyms rarely stock the odd change. At 10 reps the formulas are starting to drift apart — 48 lb between the lowest and highest here — which is the method telling you the estimate is getting soft. Treat the lower end as the safer number. Your 495 lb set represents about 75% of the estimated max.
Why 10 reps widens the answer
Ten reps is the last rep count with a page on this site, and it is the least reliable of them. All seven equations were fitted on sets of two to ten, so ten is the boundary itself; past it they diverge fast enough that we do not publish the pages at all. The 48 lb spread here — 7.3% — is the honest width of the answer. A ten-rep set measures endurance as much as strength, so if you want a number to programme from, repeat the test with four or five reps.
Training loads from a 660 lb max
Programmes are written as percentages of the max, not as absolute weights. From 660 lb, rounded to the nearest 5 lb:
| Percentage | Load | Typical set |
|---|---|---|
| 95% | 625 lb | Heavy single |
| 90% | 595 lb | Doubles and triples |
| 85% | 560 lb | Sets of four to five |
| 80% | 530 lb | Sets of six |
| 75% | 495 lb | Sets of eight |
| 70% | 460 lb | Volume work, ten or more |
660 lb is 6x45 + 1x35 + 1x5 per side plus the bar, 665 lb loaded. At this load the 48 lb the formulas disagree by is bigger than a pair of 10s, which is a good argument for programming off the lower estimate — 619 lb — and letting the bar prove you wrong rather than the other way round. Warm-up jumps also get long here; work up in percentages rather than in fixed plate increments.
The formulas behind this answer, and their limitations
The table above the fold is Epley (1985), Brzycki (1993), Lombardi (1989), O'Conner (1989), Wathan (1994), Lander (1985) and Mayhew (1992), each applied to 495 lb for 10 reps; the headline figure is their median. All seven assume a set taken close to failure with clean technique. At 10 reps you are at the top of the band these equations were fitted on, and beyond it they diverge fast, so 660 lb is a ceiling rather than a target. They also under-rate strong, experienced lifters, who hold a higher percentage of max at any rep count than the regression average. The full account — the populations behind each equation, what a bounced rep does to it, and why we stop at ten reps — is set out on the methodology page. Warm up thoroughly and use safety bars or a spotter before going anywhere near 660 lb. This is a training estimate, not medical advice.