There is no single best timeframe for gold. What works is pairing them: use a higher timeframe (H1 or H4) to establish direction and key levels, then a lower one (M15 or M5) to time the entry. Trading one timeframe in isolation is what produces random results.
Traders hunt for a magic timeframe hoping it will supply an edge. It won't. A timeframe is a zoom level, not a strategy. The edge comes from knowing where the important levels are and waiting for price to interact with them — and that requires more than one zoom level.
Use H1 or H4 to answer: what direction is the market in, and where are the significant highs and lows? Then drop to M15 or M5 to answer: has price reached one of those levels, and has it rejected? The higher timeframe supplies bias and location; the lower one supplies timing and a tight invalidation.
On M1 and M5, most of what you see is noise — spread, small orders, and algorithmic churn. The signal-to-noise ratio is poor, so entries fire constantly and most fail. Higher timeframes have fewer signals but a much higher proportion of meaningful ones.
If you can't watch screens all day, trade H1 or H4 structure with wider stops and fewer trades. If you can sit through the London–New York overlap, M15 entries against H1 levels is a sound combination. The worst choice is a timeframe that demands attention you can't give.
Yes, as an entry timeframe — provided your direction and levels come from H1 or higher. M15 alone produces too many low-quality signals.
You can, but spread and noise consume much of the edge, and it demands constant attention. Most traders do better on higher timeframes.
Typically several: a higher timeframe for bias and levels, a lower one for execution timing.