Plate Maths: Why 165lb Isn't 45 + 10 + 5

Ask most lifters to load 165lb and they will put a 35 and a 25 on each side without thinking. Ask a naive computer program to do it and you get a 45, a 10 and a 5 — the same weight, three plates instead of two, and an extra trip to the rack.

The difference is a nice illustration of why the obvious algorithm is not always the right one.

The obvious approach, and where it fails

The intuitive way to load a bar is to work down from the heaviest plate that fits: take as many 45s as you can, then as many 35s, then 25s, and so on. This is a greedy algorithm, and for most weights it is both fast and correct.

It relies on an assumption, though: that each denomination divides neatly into the ones above it. Kilo plates satisfy this. The set runs 25, 20, 15, 10, 5, 2.5, 1.25, 0.5, 0.25 — evenly spaced through the useful range, and working down from the top is genuinely optimal for every weight the set can make.

Pound plates do not. The set runs 45, 35, 25, then drops straight to 10. There is nothing at 20 or 15. That gap is where greedy loading goes wrong:

• 60lb a side (165lb on the bar): greedy takes 45, then 10, then 5 — three plates. Optimal is 35 + 25 — two.
• 105lb a side (255lb on the bar): greedy takes 45 + 45 + 10 + 5 — four. Optimal is 45 + 35 + 25 — three.

Across every loadable pound weight up to 200lb a side there are exactly four of these cases. Not many — but 165lb and 255lb are ordinary working weights, not curiosities, and any lifter would spot the answer immediately.

What the calculator does instead

The Plate Loading Calculator does not work down from the heaviest plate. It searches for the loading that gets closest to your target and, among those, uses the fewest plates. On the kilo set the result is identical to the greedy answer for every weight, because the kilo set really is evenly nested. On the pound set it fixes those four cases.

It also refuses to round up. If the plates you have cannot make the weight you asked for, it loads the closest it can at or below the target and tells you what is left over per side. A loading tool that silently adds a kilo to a heavy single is a worse tool than one that admits it came up short.

Why the leftover figure matters

The leftover is the most useful output for anyone training somewhere with an incomplete plate rack. Ask for 61kg on a 20kg bar with nothing smaller than 2.5kg available and you will learn immediately that you cannot have it — the bar makes 60kg and you are 0.5kg a side short.

That is worth knowing before you build a programme around 2.5% weekly increments that your gym physically cannot deliver. Put your real plate inventory into the calculator and the increments available to you stop being a guess.

Loading in practice

A few habits that make the whole problem smaller:

• Load symmetrically and count out loud. The most expensive plate error is not a suboptimal combination, it is an asymmetric bar. Check both sides before you get under it.
• Keep the heaviest plate innermost. Not a maths point but a handling one — the bar is easier to control and collars sit better.
• Learn the two or three weights you use most. If your working sets sit at 100, 120 and 140kg, you should not be computing anything. The chart on the Rep-Max Plate Chart page does that for a whole block at once.
• Collars count. Competition collars are 2.5kg each. Training clips are usually negligible. If you are chasing precision, know which you have.

The general lesson

The greedy approach fails here for the same reason it fails at making change with an awkward set of coins: it works whenever the denominations nest, and stops working the moment they gap. Kilo plates nest. Pound plates gap between 25 and 10. Everything above follows from that one fact about how the plates in your gym happen to be sized.

Collars, clips and what the bar really weighs

Every calculation above assumes the bar plus the plates and nothing else. In a real gym there is usually something else.

Competition collars weigh 2.5kg each — 5kg on the bar, which is a whole plate's worth and is counted in a competition total. Training clips and spring collars typically weigh a few hundred grams and are conventionally ignored. Screw collars sit between the two, sometimes half a kilo each.

If you are chasing precision on percentage work, decide once whether your collars count and be consistent. Adding 5kg of collars to some sessions and not others introduces a variation larger than everything else discussed here combined.

Plate thickness and how much fits

There is a physical limit that the arithmetic ignores. A standard Olympic sleeve is about 415mm of loadable length, and how much you can fit depends on what the plates are made of.

Calibrated steel competition plates are thin — a 25kg plate is around 26mm. Rubber-coated commercial plates are much thicker, often 40–55mm for the same weight, and bumper plates thicker still. A bar loaded to 250kg in thin steel plates fits comfortably; the same weight in rubber bumpers may not fit at all.

This is why the minimum-plate answer is not merely tidier. Fewer, heavier plates leave sleeve space, sit closer to the collar, and make a heavy bar easier to control. At the top end it is the difference between loading the weight and not.

Loading for drop sets and strip sets

One case where minimum plates is the wrong objective: any protocol where you strip weight between sets without resting. There, what you want is the loading that can be reduced fastest — several smaller plates on the outside so a drop is one plate per side rather than a rebuild.

For a drop set from 100kg to 80kg to 60kg on a 20kg bar, loading 40kg a side as 20 + 10 + 10 lets you pull one 10 for each drop. The minimum-plate answer, 25 + 15, requires unloading both and starting again. Same weight, very different session.

The calculator optimises for the general case, which is loading a bar once and lifting it. Knowing when your situation is not the general case is the useful part.

A note on asymmetry

Every result here is per side, and it is worth stating plainly why that matters. An unevenly loaded bar is the most consequential loading error there is — far worse than a suboptimal plate combination or a rounding decision. It rotates under load, it loads one side of the body unpredictably, and on a squat or a press it is genuinely dangerous.

Load one side fully, then the other, then look at both before you get under it. Every experienced lifter has a story about the day they did not.