When a bigger water change helps, and when it just stresses the tank

"30% weekly" is a sticker number. It's a reasonable default, but it isn't an answer to "how big should my water change be", because that depends on what you're trying to move and how fast it's piling up. A planted tank with light stock and a hungry plant mass might coast for a fortnight on nothing. A grow-out tub of young cichlids can need more than 30% twice a week just to stay level. The sticker doesn't know which tank is yours.

So before reaching for a bigger bucket, it helps to know what the water change is actually doing to each parameter. They don't all move the same way, and most advice glosses over that.

The dilution maths, honestly

A water change does one thing: it swaps a fraction of your tank water for new water. If you change a fraction p of the volume, anything dissolved in the tank that your source water does not carry gets removed in that same proportion. Replace 30% of the water and you remove 30% of the accumulated nitrate, because the replacement water brings 0 mg/L of nitrate with it.

C_after = C_before × (1 − p)

So a tank sitting at 40 mg/L NO₃, after a 30% change:

40 × (1 − 0.30) = 28 mg/L

That's the clean case, and nitrate is the clean case, because tap and RO water normally arrive nitrate-free.

Here is where it goes sideways for a lot of people. KH and GH don't get diluted toward zero. They move toward whatever your source water carries. A change averages your tank against your tap rather than subtracting from it. The same formula gains a second term:

C_after = C_before × (1 − p) + C_source × p

If your tank KH has drifted down to 2 dKH and your tap runs at 8 dKH, a 30% change raises KH:

2 × 0.70 + 8 × 0.30 = 3.8 dKH

Keep a soft-water tank on remineralised RO at 3 dKH instead, with the tank crept to 5 dKH from rock leaching, and the same change lowers it. A water change pushes GH and KH up or down depending entirely on the gap between tank and source. (If GH versus KH isn't second nature yet, the gh-vs-kh guide is the short version.) This is how a keeper on hard tap doing "big changes to fix high nitrate" can also be hammering their soft-water shrimp with a GH swing they never measured.

30 vs 50 vs 70 on nitrate

For a one-off swap on a tank at 40 mg/L NO₃, the single-change numbers are easy:

Change size	NO₃ after one change	Removed
30%	28 mg/L	12 mg/L
50%	20 mg/L	20 mg/L
70%	12 mg/L	28 mg/L

The bigger swap clearly knocks more off in one go. But almost nobody does one-off changes. The number that matters is where a routine settles, and that's less intuitive.

Say your stock and feeding add roughly 20 mg/L of nitrate to the water between weekly changes. Each week you add 20, then remove a fraction. Over many weeks the tank converges on a sawtooth that bounces between a steady high (just before the change) and a steady low (just after). The floor it settles to is:

NO₃_high ≈ A / p          NO₃_low ≈ A × (1 − p) / p

with A the weekly nitrate input and p the change fraction:

Weekly change	Settles between (low → high)
30%	47 → 67 mg/L
50%	20 → 40 mg/L
70%	9 → 29 mg/L

That's the useful comparison. Going from 30% to 50% weekly doesn't just shave a few mg/L off, it roughly halves where the tank lives, because you're fighting a continuous input, not a fixed pile. If your nitrate keeps creeping up despite weekly 30% changes, this table is why: at that cadence and input, 30% isn't enough to balance what's going in, and no single heroic change fixes a flow problem.

Why frequent and smaller usually wins

Bigger isn't free, though. A water change doesn't only move nitrate. It moves GH, KH, TDS, pH, and temperature all at once, and livestock feel those swings far more than they feel a slightly elevated, stable nitrate.

A fish or a shrimp osmoregulates against the water it's sitting in. Drop 70% of the volume and replace it with water that's a few degrees cooler, a couple of dGH softer, and a notch off on pH, and you've handed the animal a sudden environmental step to absorb in minutes. Caridina and Neocaridina shrimp are especially unforgiving here; a large fast change is a classic trigger for a post-change moult death. The shrimp-safe-parameters guide goes deeper on the bands that matter.

Compare two routines that remove the same nitrate over a week:

• One 50% change. Big single swing in GH/KH/TDS/temp, then six days of drift.
• Two 30% changes. Each swing is smaller, the tank spends less time at the nitrate peak, and the average parameters are steadier.

The second is almost always kinder, and in my experience it's less work to recover from, because you're never dragging the tank a long way back to baseline. Stormy watches the seven-day window, and steadier-but-smaller is exactly the shape that keeps that window calm: small regular dips instead of one cliff and a long climb. The water-change-impact tool runs this same maths interactively for your tank volume if you want to see the trade-off before committing.

When a bigger change actually is right

There are real cases where you want the big bucket, and they're mostly one-offs rather than a new routine:

• Acute nitrate. A test that comes back at 80+ mg/L on a stocked tank is worth a large change now, often 50% today and another 50% tomorrow, because the single-change table above tells you one 30% barely touches it. Two staged 50% changes beat one panicked 80%, for the swing reasons above.
• Post-incident. A dead fish you found late, an overfeeding, a substrate disturbance that clouded the tank with mulm. Dilute the spike, then let the filter catch up.
• Post-medication. Most treatments expect a large change (or a run through carbon) once the course is finished to clear the residual. Follow the medication's own instructions here, not a generic percentage.

In all three, the change is a correction, not a habit. Once the acute reason is gone, go back to the steady routine.

When to change the cadence, not the size

If your nitrate sits fine on the day you change water but is too high by day six, the lever is frequency, not volume. Two 30% changes a week hold a tank far steadier than one 60%, for the same total water moved and a fraction of the parameter shock. Logging the values across the week tells you which it is: high right before every change is a cadence problem, high right after a change points at source water or input instead. The cycle-envelope guide is about reading exactly that high-to-low swing on the chart.

When the problem isn't the water change at all

A bigger change is the reflex for most problems, and for a surprising number of them it's the wrong tool entirely.

Low KH is a remineralisation problem, not a dilution problem. If your tank keeps drifting toward a pH crash on soft water, the answer is putting buffer in, not taking water out. Diluting won't help if your source water is soft too, and on RO it makes low KH worse. Fix the new water before it goes in.

Algae is a light-and-nutrient balance problem. A bigger change can knock back a nutrient spike that's feeding an outbreak, but if the cause is too much light for the available CO₂ and plant mass, you'll dilute the symptom and the algae returns next week.

Persistent ammonia or nitrite is a cycle problem. Above-zero NH₄ or NO₂ means the filter isn't keeping up. Water changes buy time and protect livestock, so do them, but they don't fix the bacterial shortfall. That's a stocking, feeding, or filter-maturity issue, and the nitrogen-cycle guide covers why.

In every one of these, a big swap adds stress without touching the cause. Find the actual lever first.

The non-negotiables, whatever the size

A few things hold regardless of how big the change is. Temperature-match the new water to within a degree or so. Dechlorinate every time, because chlorine and chloramine harm fish and wipe filter bacteria. And know your source water's GH and KH before it goes in, so you're not surprised by which way they move the tank. A 20% change with mismatched, cold, chlorinated water can do more harm than a careful 50%.

Size was never really the question to start with. What you're moving, how fast it accumulates, and what your source water carries: write those three down and the right size mostly answers itself.