Scope. This page provides a controlled reader-run protocol and does not claim site-run measurements. For cooking, cooling, storage and reheating, follow the current official guidance linked on the page.
A recipe is developed on one appliance and read on thousands of others. The author's instruction was accurate in their kitchen. Whether it survives the journey depends almost entirely on a variable neither party is thinking about: what kind of heat is under the pan.
At high heat this hardly matters. Everything gets hot. At low heat it matters enormously, because the three common domestic heat sources fail in three different ways when you ask them to deliver a small, steady amount of power. One of them may not be able to go low enough. One may go low enough but not steadily. One may go low enough and steadily but take an age to change its mind.
Anyone who has moved house and found their reliable dishes suddenly behaving differently has run this experiment already, without instrumentation. The protocol below shows how to run it with instrumentation.
What the test asks all three hobs to do
A comparison is only worth anything if the task is identical and the protocol is stated before the results.
The task. Bring a stated volume of water in a stated pan to a target temperature in the low bands, then hold it there for thirty minutes without touching the controls. Then a second task: make a step change to a new target and log how long the pan takes to arrive and settle.
The constant. The same pan for every run, on every hob. The same water volume. The same probe, in the same position, at the same depth. If the pan is a variable, nothing else in the comparison means anything.
The pan. One heavy pan with a ferrous base, so it works on induction as well as the other two, recorded by material, base thickness and diameter.
The instruments. A calibrated probe logging liquid temperature continuously, plus a separate surface measurement at the pan base, because the gap between what the water is doing and what the pan base is doing is one of the central things the test shows. Both instruments checked at the ice point before and after each session, with offsets logged.
Ambient. Kitchen temperature and extractor state recorded per session. Draughts are not trivial over a thirty-minute hold.
Zone matching. Pan diameter matched sensibly to burner or zone size on every hob, and the pairing recorded, because a wide pan on a small ring is a different experiment.
Replicates. Multiple runs per hob per task, on separate days, with the spread recorded. Discarded runs appear in the log with the reason.
Three measures, defined in advance:
- Floor. The lowest stable liquid temperature the hob can hold with this pan and this volume. Not the lowest dial setting; the lowest temperature it can actually sustain.
- Stability. Drift and swing across a thirty-minute hold, expressed as the range between the highest and lowest logged values, not as a single spot reading. A spot reading on an oscillating system tells you nothing except when you happened to look.
- Responsiveness. Time from a control change to a settled new temperature, in both directions, plus the size of any overshoot.
Reader worksheet
The table below is a worksheet for recording your own readings. Nothing in it is filled in, because we have not taken the measurement; every value belongs to a run you did yourself, on your own hobs, to the protocol above.
| Measure | Gas | Induction | Electric radiant |
|---|---|---|---|
| Floor: lowest stable liquid temperature held, in °C | — | — | — |
| Stability: highest-to-lowest range across a thirty-minute hold, in °C | — | — | — |
| Responsiveness downward: time from control change to settled, in minutes and seconds | — | — | — |
| Responsiveness upward: time from control change to settled, in minutes and seconds | — | — | — |
| Overshoot after a step change, in °C | — | — | — |
| Cycle period at the lowest usable setting, in seconds | — | — | — |
| Peak-to-trough swing at the pan base over one cycle, in °C | — | — | — |
Use one copy of the worksheet per run rather than averaging as you go, and record beside it the date, the pan, the water volume, the burner or zone the pan sat on, the kitchen temperature, the extractor state, and the probe offsets from the ice-point checks before and after. A reading without those conditions cannot be compared with anything, including your own next reading.
Celsius canonical, Fahrenheit alongside, raw logs kept.
The floor: how low each one can actually go
The floor is where the differences start, and it is not a design decision in every case.
On gas, the floor is set by flame stability. Below a certain gas flow the flame becomes unreliable and eventually will not sustain, so the lowest usable setting is a physical limit rather than a chosen one. Some hobs include a small dedicated low-output burner precisely because of this; many do not. The result is that plenty of gas hobs have a floor that sits above the temperature a delicate task needs, and no amount of careful dial work will get below it.
On induction, the floor is a control-system decision. The hob can deliver very little average power, and the question is how it does so, which is the cycling problem below.
On electric radiant, the floor is bounded by the element's own behaviour. A hot element radiates whether you want it to or not, and the residual heat in the element and the hob surface sets a floor of its own even after you have turned the control down.
Stability: holding a target for thirty minutes
Stability is the measure that matters most for the work this site is about, and it is the one that almost never appears in appliance comparisons, because it takes half an hour per run to obtain and cannot be lifted from a spec sheet.
A hob that arrives at the right temperature and then drifts steadily upward over twenty minutes has failed at the job even though it looked correct when you checked it. A hob that oscillates around the right average has also failed, if the peaks are high enough to matter to whatever is in the pan.
We expect gas to be relatively steady in the sense of delivering continuous rather than intermittent power, while being vulnerable to draughts and to the mismatch between a ring of flame and a flat pan base. We expect radiant electric to drift slowly with its own long cycles. We expect induction to be the most stable in average terms and the least stable instantaneously. All of which is a hypothesis until it is logged.
Responsiveness: how fast each reacts
Responsiveness is the reason gas cooking feels the way it does. Turn the dial, the flame changes immediately, and the heat input changes with it. The pan still has thermal mass and still takes time to follow, but the input is instant.
Induction is comparably immediate at the input, because the power delivered into the pan base changes at once.
Electric radiant is the outlier. The element has to heat up or cool down before the pan sees anything, so a control change is a request that gets honoured a minute or two later. In practice this means an experienced electric-hob cook learns to anticipate: to turn the heat down before the pan needs it, and to lift the pan off the ring entirely when they need an immediate change. That is not a workaround for a defect. It is the correct technique for the appliance.
The number that matters is time-to-settled after a step change, in both directions, and it must be measured on your own setup.
Induction's cycling behaviour, and why it matters below a simmer
This deserves its own section because it is the least understood behaviour in the domestic kitchen and it directly contradicts what the control panel implies.
Many induction hobs cannot deliver a genuinely low continuous power. To produce a low average, they deliver full power for a period, then no power, then full power again, in a repeating cycle. Set to a low number, the hob is not running gently. It is alternating between running hard and running not at all.
Whether this matters depends entirely on how much thermal buffering sits between the hob and the food.
- A large pot of water: barely matters. Water is a huge thermal buffer and smooths the cycle almost completely. The average is what the food experiences.
- A thin pan with a delicate sauce: matters a great deal. The pan base tracks the cycle closely, and the sauce touching that base experiences the peaks, not the average. This is how a sauce can catch and scorch at the bottom while the bulk of it sits below the temperature you were aiming for.
- A heavy pan: matters less, because the pan's own mass does the damping. This is one of the few situations where a thick base is not merely nice but load-bearing.
Two consequences follow. First, an induction user who reads "hold at a bare simmer" and sets a low number may be doing something quite different from the recipe author's intention. Second, the classic advice to stir constantly is only a partial fix, because stirring moves the contents around but does not lower the peak temperature at the metal.
What to record: the cycle period at each low setting, the peak-to-trough swing measured at the pan base, and how much a heavier pan or an added water buffer reduces it.
If any of this describes your hob, what medium-low measures on each hob has the protocol for finding your own settings, and it is worth running before you conclude anything about your appliance from anyone else's.
This is three appliances, not three classes
We have to be explicit about this, because comparison articles routinely overclaim and the overclaim is what makes them useless.
This protocol compares individual hobs, not classes of appliance. Induction hobs vary in their low-power control strategy between manufacturers and model years, and the strategy is not published. Gas hobs vary by burner design, whether a simmer burner is fitted, and how the gas valve is set up. Electric radiant hobs vary by element wattage and by how much mass sits between element and pan.
So this page declares no winner. A verdict of the form "induction is best at low heat" would require a sample nobody has and a definition of "best" that ignores how differently the three fail. What it can honestly give you is the protocol, and how to run it on yours.
That last part is the useful bit. The value of this test is not anyone's published numbers. It is that after reading it you know which three properties to check on your own hob, and how.
What to do if your hob cannot hold the band
Most readers arrive at an article like this because something is not working, so here is the practical part, which holds whatever your own log says.
If your hob's floor is above the temperature you need, or it will not hold steady, you have five options and none of them requires a new appliance.
- Add thermal mass. A heavier pan damps everything: cycling, overshoot, hot spots. If you cook gently and own one thin pan, this is the highest-value change available to you.
- Put something between the burner and the pan. A diffuser spreads and moderates the heat reaching the base. How much it actually helps is a measurable question rather than a matter of opinion, and what diffusers actually do shows how to measure it.
- Cook the food in water rather than in the pan. A water bath decouples the food from the hob's behaviour entirely: the water buffers the cycling, and the food sees a stable temperature even when the hob is misbehaving. This is the single most effective workaround and it costs nothing if you own two pans.
- Use the oven as a holding box. An oven is a large, slow, well-insulated thermal environment, which makes it a better holder than most hobs even though it is a worse cooker. Its own accuracy at low settings is a separate question, and rarely as good as the dial claims.
- Coast off the heat. With a heavy pan, take it off the burner and let it drift, returning it briefly when the number falls. Laborious, free, and surprisingly effective for a short hold.
All five are compared in how to hold a low temperature anyway, with a protocol for measuring their stability on your own hob.
In Plain English
Gas, induction and electric hobs all get food hot. They differ in how low they can go, how steadily they hold, and how fast they respond, and those three properties are what decide whether a gentle recipe works on your appliance.
Gas responds instantly but has a floor set by flame stability and is vulnerable to draughts. Induction responds instantly and holds a good average, but at low settings many hobs deliver that average by cycling full power on and off, which can scorch a thin pan even when the number looks right. Electric radiant is the slowest to respond in both directions, so you have to anticipate rather than react.
This matters most if you have recently changed hob type, or if a recipe that others report as reliable keeps failing for you in the same way. Use it by identifying which of the three failure modes you are up against, then applying the matching workaround: thermal mass, a diffuser, a water bath, the oven, or coasting off the heat.
What to do next: put a pan of water on your lowest setting, leave it for half an hour, and probe it at the start and the end. If the number moved, you have learned something no appliance specification would have told you.
Where to go from here
Translate the recipe yourself, which is the only translation that accounts for your appliance. Put a measured quantity of water in the pan you actually use, set the dial one notch at a time from its lowest position, and record the temperature the water settles at once it has stopped changing at each setting. That table is the map between your dial and a recipe written for someone else's hob. It is rough, it is specific to that pan and that quantity, and it is still better than a number transplanted from an appliance you have never used.
If the conclusion you reach is that your pan is the weak link rather than your hob, the thermal behaviour of different pan materials sets out how to measure yours, including how much difference base thickness actually makes.
Related technique: The Four Temperature Bands Every Cook Should Know.