A chart crossing and a working alert are separate facts
You may see a chart pass a level and later find no notification. Before concluding that the alert failed, establish which price observations are available and when the alert was created. A line on a historical chart does not by itself show that the alert already existed, had become active or used that same price field.
TradingView’s introductory documentation describes alerts for price and other conditions, with several notification methods. This worksheet does not reproduce those settings. It answers a narrower question from supplied records: where does each observed threshold transition sit relative to the creation instant? Real activation, delivery and provider logs remain separate evidence.
Sources: TradingView — Introduction to alerts.
Make the crossing rule explicit
For an upward comparison, the earlier price must be strictly below the threshold and the later price at or above it. For a downward comparison, the earlier price must be strictly above and the later price at or below. An earlier price already on the threshold does not start a new pair under these rules.
This convention is stated so that the result is reproducible. It is not presented as the exact rule of every charting service, exchange or app. A provider may evaluate bid, ask, last, an indicator, a bar close or another stream. Keep one product and one price definition throughout your supplied observations; otherwise even perfect time arithmetic answers the wrong question.
A pair provides an interval, not the crossing instant
Suppose you observed 99 at 01:10 and 101 at 01:20 with a threshold of 100. The entered records support an upward transition between those observations. They do not identify whether it happened at 01:11, 01:19 or at several times in between. This guide calls the observation interval a bracket.
The bracket uses the earlier boundary exclusively and the later boundary inclusively: (01:10, 01:20]. The earlier observation is still below the level under the upward rule. No straight-line interpolation is used. A market may jump between recorded values; the worksheet is classifying observed changes, not proving that a trade printed at every intermediate price.
Compare both bracket boundaries with creation
A bracket ending before creation is classified as before creation. A bracket whose earlier observation is at or after creation lies entirely after the supplied creation instant. If the earlier observation is before creation and the later one after, the bracket spans the boundary and the ordering of the actual transition is unknown.
When the bracket ends exactly at creation, the worksheet uses a separate category. Two records sharing one timestamp do not reveal server event order, activation latency or precision beyond what was recorded. Treating that equality as a successful after-creation trigger would claim more than the data supports.
The repeated London hour changes the answer
GOV.UK lists 25 October as the UK clock rollback date for 2026. It explains that clocks go back one hour at 2am and the UK returns from BST to GMT. In the fictional scenario, 01:20 with offset +01:00 is 00:20 UTC, while 01:20 with offset +00:00 is 01:20 UTC. Identical clock faces can therefore identify instants an hour apart.
The created record is 01:15 with +00:00. The first 01:10–01:20 pair is entirely before it; the repeated-hour pair spans it. Reading only the hour and minute would erase this distinction. The worksheet requires an offset or Z and displays the resulting London time with an offset, preserving the original timestamp in CSV.
| Observation pair | Recorded offsets | Relation to 01:15 GMT creation |
|---|---|---|
| A → B | 01:10 and 01:20, both +01:00 | Bracket ends before creation |
| C → D | 01:10 and 01:20, both +00:00 | Bracket spans creation |
| E → F | 01:25 and 01:30, both +00:00 | Bracket entirely after creation |
Sources: GOV.UK — When do the clocks change?.
A missing observation must not disappear
A blank price makes each adjacent comparison involving that row unknown. The tool does not skip over the row and connect the previous known price to the next known price as if no gap existed. Missing saved evidence has the same effect. The entered row remains in the export, making the limitation visible to the next reviewer.
Even complete-looking rows do not prove that you supplied every tick. Two equal prices around a long gap can hide multiple crossings. A no-crossing-pair result means only that the adjacent supplied values do not meet the stated rule. It is not evidence that the market never crossed the threshold during the interval.
Being beyond the threshold is not a new crossing
If the first available observation is already above an upward threshold, its prior state is unknown. This worksheet cannot assign an earlier crossing to creation or count one just because the current level is high. A later drop below followed by a return to or above the threshold can form a new observed pair.
The same distinction matters after editing an alert. This tool accepts one creation record and one fixed threshold; it does not infer an amendment history, compare two alert versions or replay provider settings. If the level or condition changed, preserve those versions and review each clearly bounded configuration separately.
Preserve order and evidence instead of repairing silently
Rows must be strictly increasing by their actual UTC instants, not sorted text labels. Equal instants are rejected because their ordering is unresolved. Duplicate IDs are rejected rather than counted twice or overwritten. These safeguards protect the chronology; they do not authenticate an ID or prove that two clocks were synchronised.
The report as-of instant prevents a later observation from being mixed into an earlier report. A creation date after report as-of also stops the calculation. An absent creation reference leaves crossing pairs visible but their relationship to creation unsupported. Original references remain text; a pasted label is never opened as a link or treated as an instruction.
Use the report to narrow the next evidence check
Review the before, boundary, spanning and after categories, then compare them with the provider’s actual activation and notification records if you have them. An after-creation bracket alone is not a failed-alert diagnosis. An omitted price field, a different sampling policy or unknown server timing may still matter.
The local CSV contains original observations and UTC bracket boundaries. No account connection, automatic monitoring, alert creation or delivery test occurs. Editing the input clears stale results; manual copying remains available if clipboard access is unavailable. The full method and fictional example remain readable without JavaScript, before the optional offer link.
Worked examples
Example 1
The fictional upward rule uses a threshold of 100 GBP. A→B crosses in the first 01:10–01:20 hour, before the recorded creation at 01:15 GMT. C→D crosses in the repeated 01:10–01:20 hour and spans creation, so its exact ordering is unknown. E→F crosses entirely after creation. There is one bracket in each category; no actual notification or exact crossing time is proved.
Inspect example inputs
{
"product": "DEMO-ASSET/GBP",
"price_field": "last price in GBP",
"level": "100",
"direction": "up",
"created_at": "2026-10-25T01:15:00+00:00",
"creation_reference": "DEMO saved alert creation record",
"as_of": "2026-10-25T02:00:00Z",
"rows": "id|observed_at|price|reference\nDEMO-A|2026-10-25T01:10:00+01:00|99|DEMO observation A\nDEMO-B|2026-10-25T01:20:00+01:00|101|DEMO observation B\nDEMO-C|2026-10-25T01:10:00+00:00|99|DEMO observation C\nDEMO-D|2026-10-25T01:20:00+00:00|101|DEMO observation D\nDEMO-E|2026-10-25T01:25:00+00:00|99|DEMO observation E\nDEMO-F|2026-10-25T01:30:00+00:00|101|DEMO observation F"
}Questions about the method
Does an after-creation bracket prove the service should have notified me?
No. Activation, provider condition, sampling and delivery evidence are not measured by this worksheet.
Why does an exact timestamp match get a separate category?
A shared timestamp does not establish event order or finer precision. The tool leaves that boundary unresolved.
Can I use London time without an offset?
No. Some local times repeat or do not exist around clock changes. Supply a known offset or Z from your evidence.
Primary sources and scope
- GOV.UK — When do the clocks change?Checked 2026-10-21. 2026 UK clock rollback is 25 October, one hour back at 2am; BST/GMT distinction. This supports the dated repeated-hour example, not market observations or alert activation.
- TradingView — Introduction to alertsChecked 2026-10-21. Alerts can monitor price and other conditions and use notification methods. No claim that this worksheet reproduces provider trigger, timing or delivery semantics.
Examples explain a method. They do not predict prices or establish where a trading service is available. Check the named source before relying on future schedules.