M15 Leads Timeframe Mentions, but Says Little About Trade Frequency
By Cree Lawson / Prismatic Algos
An H1 EA does not necessarily trade every hour. The chart interval sets the data cadence; entry conditions, filters, sessions and position limits determine how often orders are placed.
What the dated catalog names
In the October 5, 2026 Lattice export of 6,662 trading-strategy specifications, 4,334 (65.1%) named at least one parseable timeframe in the Timeframes field. M15 was most common: 1,508 mentions (34.8% of the 4,334). H1 followed with 1,317 (30.4%); M5 had 901 (20.8%); H4 and D1 tied at 744 each (17.2%); M1 had 598 (13.8%); M30 had 554 (12.8%).
| Timeframe | Records | Share of 4,334 |
|---|---|---|
| M15 | 1,508 | 34.8% |
| H1 | 1,317 | 30.4% |
| M5 | 901 | 20.8% |
| H4 | 744 | 17.2% |
| D1 | 744 | 17.2% |
| M1 | 598 | 13.8% |
| M30 | 554 | 12.8% |
Other hourly intervals were less common: H2 appeared in 54 records, H3 in 13, H5 in one, H6 in 25, H8 in 21, H10 in two, and H12 in 15. H5 and H10 are rare text labels, not standard MetaTrader chart intervals. Explicit forms such as 1h , 1 hour , 60 minutes and daily were normalized. Generic values such as any and null were excluded; ambiguous 1M was left unclassified. Each interval was counted once per record. Multi-timeframe strategies contribute to several rows.
Trade frequency needs trade logs
How many trades per week does each hourly timeframe produce? This export cannot answer. Across all 6,662 rows, only two records explicitly mentioned trades/week or trades per week ; both were M15 or M5/M15 source claims, not independent observations. There was no adequate trade-rate sample for any hourly label. The count of H1 candles per week is not a trade count.
To calculate a meaningful rate, a report needs completed trades and start/end dates: closed trades ÷ (observation days ÷ 7) . Backtest, demo and funded-live rates should remain separate. Partial fills, duplicate reports and overlapping variants need consistent treatment. The median strategy-level rate and its spread are more useful than a single pooled average dominated by one hyperactive EA.
The takeaway for MQL5 builders is to measure observed activity from complete reports. Do not infer it from the timeframe property in code or a catalog label.
Method note: Snapshot October 5, 2026; unit one strategy-specification row; denominator 4,334 with a parseable timeframe; total 6,662. The source CSV is an internal research export. No hourly trades-per-week rate was imputed.
