Forex trading bot Switzerland

FINMA and the banking licence requirement

The Swiss Financial Market Supervisory Authority regulates financial markets in Switzerland, and the arrangement here is unusual: firms offering forex trading to retail clients generally require a banking licence rather than a lighter investment-firm authorisation.

That is a substantially higher bar than most jurisdictions apply, with capital requirements to match, and it means there are relatively few domestic providers. Swiss residents also commonly trade through brokers regulated elsewhere, in which case the protections that apply are the ones from that jurisdiction rather than FINMA's.

Running a bot on your own account is not a regulated activity. The licensing question applies to the provider.

15 January 2015, and why every automated trader should know about it

The Swiss National Bank had maintained a floor under EUR/CHF, committing to defend a minimum exchange rate. On 15 January 2015 it removed that floor without warning. EUR/CHF fell by a very large margin within minutes, in conditions where liquidity effectively disappeared.

The consequences for automated trading were severe and specific. Stop-loss orders did not execute at their levels — there was nothing to execute against — and filled far below. Accounts went deeply negative. Several brokers were pushed into insolvency, and one large retail broker required an emergency loan to continue operating.

The lesson is not that stops are useless. It is that a stop is an instruction to trade at a price, not a guarantee of that price, and a strategy whose worst case assumes stops always fill has not modelled its worst case at all.

What that means for how you build a bot

• A backtest cannot show you this. Historical data contains the gap as a price movement; it does not contain the absence of anyone to trade with, which is what actually determined the fills.
• Position sizing should assume a stop can be exceeded, not merely reached. The question is not 'what if this trade loses' but 'what if it loses several times the intended amount'.
• Pegged and managed currencies carry this risk structurally. They trade quietly for years and then move violently once. A strategy fitted to the quiet period holds no information about the event.
• Negative balance protection, where it applies, is the specific protection against this scenario. It is worth knowing whether your account has it before you need it.
• Diversification across uncorrelated instruments limits how much of an account a single such event can reach.

Trading CHF pairs today

The franc remains a currency where policy is an active factor rather than background noise, and CHF pairs are commonly used as a safe-haven destination during risk-off periods, which produces sharp moves driven by events elsewhere.

For an automated strategy that means volatility arriving in bursts rather than steadily. Volatility-scaled stops handle this considerably better than fixed-distance ones, because they widen as conditions change instead of being sized for a calm that has ended.

Common questions