Zero weekend risk: what we do, step by step
You said you cannot carry any risk on client positions while markets are closed. This page shows exactly how the system works under that rule, symbol by symbol and hour by hour. It also shows what the rule costs, based on a new strategy D added to the simulation.
1. Why there is risk, and the one rule that removes it
When a client buys 1 BTC from you, you are now short 1 BTC. If BTC goes up, the client wins and you lose. That is your risk.
To remove it, you immediately buy 1 BTC somewhere else (at your LP, an exchange or CME). This is called a hedge. Now if BTC goes up, you lose to the client but win the same amount on the hedge. Your profit is only the spread (markup) you charged the client, minus what the hedge cost.
The problem is the weekend. Your LP is closed for gold, silver, copper and Brent. If a client trades and you cannot hedge, you carry the risk. So the rule is:
No open hedge market → no client trade.
Every client order must be copied at once on a market that is open. If no market is open for that symbol, or the open market is full, the order is refused.
2. What each symbol does on the weekend
Friday 22:00 to Sunday 22:00 UTC. On weekdays every symbol trades normally and is hedged at your LP.
| Symbol | Can clients trade on the weekend? | Where we hedge | Limit | Weekend spread |
|---|---|---|---|---|
| BTC, ETH, SOL, XRP, TON | Yes, normally | Crypto LP or exchange (open 24/7) | Same as weekdays | 1.5× |
| Gold (XAU) | Yes, up to a limit | CME 1-oz gold futures (open 24/7) | $500k of new weekend exposure | 2.2× |
| Silver (XAG) | Yes, up to a limit | CME 100-oz silver futures (24/7; confirm live) | $200k of new weekend exposure | 2.2× |
| Copper (HG) | No: closed | None (no weekend market exists) | 0 | — |
| Brent | No: closed | None. WTI oil is similar but not the same, so using it leaves risk. | 0 | — |
Why copper and Brent cannot even "close only": on Friday the client's copper position is hedged at your LP. If the client closes on Saturday, the client is out, but your LP hedge is still open and the LP is closed, so you cannot close it. Now you hold copper risk alone until Sunday. So copper and Brent must be fully closed on weekends: no opening and no closing. Clients see the price but cannot trade, exactly like today.
Why a limit on gold and silver: CME weekend contracts are small and only absorb about $500k (gold) and $200k (silver) for us. Once that is used, further orders that add exposure are refused. Orders that reduce it are always accepted.
3. What happens to one client order
This is the "hedge first, then fill" flow. It runs inside the bridge between MT5 and your LP or CME.
Because the client gets the price the hedge actually got, a price jump between the click and the fill cannot hurt you. The client sees a refusal or a requote instead.
4. One week, hour by hour
5. What the simulation says (strategy D)
Same 300 simulated years, same clients and same price shocks as A, B and C, including the bad-news weekend in week 32.
| A. No hedge | C. Unified router | D. Zero weekend risk | |
|---|---|---|---|
| Average profit per year | $6.69M | $5.51M | $3.38M |
| Bad year (1 in 20) | $2.07M | $5.08M | $3.35M |
| Worst weekend | −$668k | −$9k | +$13k (never a losing weekend) |
| Weekend profit per year | −$983k | +$1.35M | +$904k |
| Spread income | $9.49M | $8.28M | $8.17M |
| Cost of hedging | $0 | −$2.82M | −$4.79M |
| Client volume refused by limits | $4M | $1.63B | $1.41B |
| Unhedged risk during stress weekend | $millions | small | $0 on every symbol |
Per symbol, D earns about $2.8M a year from crypto and $0.6M from gold, silver, copper and Brent together. Copper and Brent still earn on weekdays, but nothing on weekends.
6. Risk that still remains (and how to handle it)
"Zero risk" here means zero market risk on positions. These smaller risks still exist and need a rule each:
Client gap (negative balance)
Copper and Brent can jump at Sunday's open. You are hedged, so the hedge pays you what the client loses. But if a client's loss is bigger than their deposit and you must offer negative-balance protection, you cannot collect the difference. Fix: on Friday at 21:00 require higher margin (lower leverage) on positions held over the weekend.
Gold/silver price difference
CME futures and LP spot gold move almost, but not exactly, together. In the simulation this cost about $17k a year on gold weekends. Fix: keep the 2.2× weekend spread, which covers it.
Contract size
CME silver trades only in 100-oz contracts. Small client trades are added up until they make one contract, so a gap of less than one contract stays open for a short time. Fix: on weekends, hedge silver in whole contracts and refuse what does not fit.
Hedge failure
The LP or CME connection drops or rejects an order. Fix: "hedge first, then fill" (section 3). If the hedge does not fill, the client order is refused, so no exposure is created.
LP or FCM failure
The firm holding your hedges fails. This is credit risk, not market risk. Fix: use regulated counterparties and split hedges across two LPs where possible.
Weekend settlement
CME weekend trades get Monday's trade date, and the FCM may ask for extra margin. Fix: keep spare cash at the FCM (included in the calculator below).
7. Money needed for strategy D
| What the money is for | How it is calculated | Amount |
|---|---|---|
| Margin for crypto hedges | Crypto limits total $5.2M × margin % | |
| Margin for commodity hedges | Commodity limits total $4.3M × margin % | |
| Running costs in cash | Hedge cost $4.79M a year for the chosen months (spread income pays it back) | |
| Reserve | No market-loss reserve is needed, only a buffer for the risks in section 6 | |
| Total own capital |
These are the same limits as strategy C, so the number of clients served is about the same (see the guide's client calculator, about 10,000 active clients). To serve more clients, raise the limits. Margin rises with them, but risk stays at zero, because every extra trade is hedged too.
8. What to build, in order
Open the hedge accounts
- A crypto LP or exchange account with 24/7 API trading for BTC, ETH, SOL, XRP and TON.
- An FCM account with CME Globex weekend access for 1-oz gold and 100-oz silver. Ask: are 100-oz silver weekend contracts live, what is the weekend margin, and what are the account minimums?
Set up the MT5 symbol groups
- A weekday group: all 9 symbols, normal spreads and limits.
- A weekend group: crypto at 1.5× spread; gold and silver at 2.2× spread with weekend limits; copper and Brent "trade disabled".
- A scheduler (Manager API) that switches groups at Fri 22:00, during CME maintenance and at Sun 22:00 UTC.
Build the "hedge first, then fill" bridge
- For each client order: check the symbol is open, check the limit, send the hedge, then fill the client at the hedge price plus markup.
- Refuse the client if the hedge is refused or times out.
Build the weekend price feed
- Gold and silver: CME futures price minus the Friday difference between futures and spot.
- Pull the quote (stop trading) if the feed is older than a few seconds.
Add the Friday and Sunday jobs
- Fri 21:00: confirm every position equals its hedge, and raise weekend margin on copper and Brent positions.
- Sun 22:00: move CME hedges back to the LP.
Build the risk screen
- Live per-symbol client position, hedge, and difference. The difference must always be zero. Alert the dealer if it is not.
Test on demo, then go live small
- Run two or three weekends on a demo server with real prices. Then go live with low limits and raise them when the numbers match.
- Re-run this simulation on your real MT5 deal history to set the limits.
All figures come from a simulated backtest (300 paths of 52 weeks). They are not live trading results or investment advice. Confirm venue hours, contract availability and margin with your LP and FCM.