Why a vario cannot just tell the truth.
A digital vario differentiates pressure — and differentiation amplifies noise. The static line breathes with turbulence, the sensor has its own jitter, the TE compensation spikes on every gust. Differentiate all of that raw and the needle becomes a blur with no usable content. So every variometer ever built — pneumatic bottle or digital sensor — averages over recent time. The only question is how much.
That average is the response time in your vario settings, and it buys smoothness with delay at a fixed exchange rate. A short filter answers fast and trembles; a long filter is calm and late. There is no setting that is both, because the trade is mathematics, not engineering: to know whether a climb is real, the instrument must watch it for a while — and while it watches, you fly.
Three delays, stacked.
The filter is only the middle of the story. Before it, the glider itself is a filter: a thermal edge does not lift a fifteen-metre wing instantaneously, and your airspeed converts thermal size into signal duration — at 150 km/h a 100 m core is a 2.4-second event, at circling speed it lasts four. After the filter, the pilot adds the final lag: hear, decide, bank. Add them up and the honest budget from entering lift to established turn is several seconds — in which you travel a hundred metres or more.
This stack is why experienced pilots fly the thermal entry with their inner ear and use the vario as confirmation: the seat feels the surge one filter earlier than the needle reports it.
Cross the thermal. Watch the needle arrive late.
Two experiments. Narrow the thermal to 60 m and watch the needle under-report a genuinely strong core — the filter never has time to charge up, which is why narrow spring thermals read weaker than they climb. Then widen the thermal back to 150 m and set the response time to 4 s: lovely calm needle, and the peak arrives almost three seconds — some seventy-five metres — past the core. Where the needle peaks is where the lift was, corrected by nothing.
Netto, relative, averager: three questions.
Modern instruments derive three siblings from the same compensated signal, and they answer different questions. Netto subtracts your polar sink at the current speed and shows what the airmass is doing — the cruise instrument, telling you whether the air under this cloud street is worth slowing down for. Relative (or super-netto) shows what you would climb if you stopped and circled here — netto minus your circling sink — which is precisely the number the stop-or-press-on decision needs. And the averager integrates the real vario over the last twenty or thirty seconds: the only honest measure of what a thermal is actually delivering, immune to the surges that fool the needle.
The practical division of labour: cruise on netto, decide on relative, stay or leave on the averager — and let the needle do what it does best, which is finding the strong side of the circle you are already flying.
Flying ahead of the needle.
None of this argues for the shortest possible setting. It argues for a known one. Pick a response time that matches your air — shorter for narrow, rough thermals, longer for wide smooth ones — then learn its delay as a distance at your circling speed, and lead every correction by that amount: widen toward lift you felt two seconds ago, not lift the needle announces now. The audio helps more than the display here; its pitch starts rising the moment the filtered value moves, several visual reaction-times before the needle catches your eye.
A vario you know the lag of is a precision instrument. A vario you believe literally is a history teacher.
Standalone digital vario — response time is a setting, not a fate, and the audio carries the news faster than the needle.
Fast pressure sensors and inertial data in one unit — the raw material for a vario that can afford a short filter.
The racing computer — vario, netto and averager side by side, each answering a different question about the same air.