The expensive thing about a stale chart is that it does not look stale. There is no grey overlay, no warning banner, no obvious tell. There is a candle, a price, and a pair of order buttons, all of them behaving exactly as they did four minutes ago when the numbers were still true.
You price a limit off that screen. One of two things then happens. Either the market has moved away from you and your order sits there unfilled while the trade you wanted goes without you, or the market has come towards you and your limit is now through the market, which means it executes immediately at whatever the best available price is rather than at the price you chose. Both outcomes are the chart's fault and neither is visible until afterwards.
Where Live sits, and what you can actually verify from the screen
The Terminal's top toolbar is dense, and it is worth knowing the order of it because the check below depends on reading two things that sit inches apart.
At capture the toolbar ran from a BTC/USD symbol selector through timeframe buttons for 1m, 5m, 15m, 30m, 1H, 4H, 1D and 1W, then a readout stamped 25 Aug 08:15 showing an open of 78,963.99, a high of 78,963.99, a low of 78,715.63, a close of 78,808.48 and a volume of 96.29, then HA, an Indicators button, a plus, the Live control, a link to the BTC page, and then LONG and SHORT.
What the Live control does underneath is not something I am going to assert, because I have not seen its behaviour documented and inventing product mechanics is how bad advice gets written. What I can tell you is what the screen lets you verify without knowing any of that, and it is enough.
The bar stamp is your evidence. On the capture it read 25 Aug 08:15 on a 15 minute chart. While a bar is in progress, its close is the current price, which is why that capture shows an open of 78,963.99 equal to the high and a close of 78,808.48 well below it: the bar opened at its top and had traded down by the moment of the screenshot. A close that is moving means data is arriving. A close that has not changed while you watch, on a liquid instrument, means it is not.
The arithmetic of a limit priced off a frozen screen
Take the capture at face value. Bitcoin at 78,808.48, and suppose you want in with 2,000 dollars.
Your chart has been frozen for ninety seconds and the market has moved 0.3 percent, which on that price is about 236 dollars a coin. Two cases, and they fail in completely different ways.
If price has fallen, your buy limit at 78,808 is now above the market. A limit above the market is marketable, so it does not rest anywhere. It executes straight away against whatever is offered, and you are a taker paying taker fees on a trade you thought you were placing patiently. The direct damage is small, a few dollars on 2,000. The real damage is that you have no idea it happened until you look at the fill, and you will repeat it.
If price has risen, your buy limit is now 236 dollars below the market and it will simply sit. You will notice in a few minutes, when the entry you were trying to take is 0.3 percent worse or gone entirely. On a trade you expected to make 3 percent on, missing it costs 60 dollars of expectation. That is ten times the fee difference in the other case, and it is the one people never count because a trade you did not make does not appear on any statement.
This is the asymmetry to hold on to. Stale data on the entry side mostly costs you fees. Stale data on the miss side costs you trades. The second is bigger and quieter.
The ten second check before you press LONG or SHORT
Three steps, and none of them requires knowing anything about how the feed works.
Read the bar stamp against your own clock. On a 15 minute chart the stamp should be the current 15 minute bucket. On a 1 minute chart it should be this minute. If the stamp is behind, stop there.
Watch the close for five seconds. On anything liquid, the close of the in-progress bar should tick. If it is completely still, that is your answer regardless of what any indicator says.
Cross-check one number against a second source before you commit. Your broker's own ticket, the venue page, or the linked asset page. The point is not that the two must match to the cent, because different sources aggregate differently. The point is that they must be in the same neighbourhood. If your chart says 78,808 and your broker ticket says 79,050, the chart is not the thing to price your order off.
If the check fails, the fix is the ordinary one. Toggle Live off and on, or reload the page, then repeat the three steps before doing anything else. What you should not do is reload and immediately click, because a chart that has just come back may still be backfilling.
Why this matters far more on some timeframes than others
The check costs ten seconds. Whether ten seconds is well spent depends entirely on which timeframe button you are on, and being honest about that will stop you performing a ritual that does nothing.
On 1m and 5m, staleness is the dominant execution risk on the screen. A minute of missing data is a meaningful fraction of your entire holding period, and a limit priced off it is close to a coin flip about which side of the market it lands on. Run the check every time. No exceptions.
On 1D and 1W, the picture inverts. If you are entering a position you intend to hold for weeks, a chart that is two minutes behind changes your entry by an amount that is invisible against your expected move. Running a careful staleness check before a weekly-horizon entry is theatre. The thing that will actually cost you on that horizon is size and stop placement, and attention spent on the quote is attention taken from those.
The awkward middle is 15m to 1H, which is where most people actually trade. My own rule there is simple: run the check if the position is large enough that 0.3 percent of it is an amount of money I would be annoyed to lose. That threshold is personal and it should be, because it depends on your size rather than on the chart.
Two habits that make staleness impossible to miss
The check works but relies on you remembering it, which is a weak foundation. Two habits do better because they change what you see rather than what you must recall.
The first is to always trade from the same screen you priced from. The failure I see most often is someone reading the level off the Terminal, switching to a broker window, and entering a number they memorised thirty seconds earlier. Now you have two possible sources of staleness, yours and the platform's, and you cannot tell them apart. Read the price and place the order from the same surface within a few seconds of each other, or do not read it at all.
The second is to make your first order of any session a small one. Not as superstition, as a probe. A small marketable order returns a real fill price from the real market, and comparing that fill against what your chart said at that moment is the single most direct test of whether your data is current that exists. It costs the spread on a small position and it tells you something no visual inspection can. On a session where that probe comes back wide of the chart, you have learned to distrust the screen before you had size on rather than after.