I ran London from a laptop on home wifi for years. It works until it does not - ISP blip at 08:02, Windows update reboot on a Sunday I forgot, coffee on the trackpad. For discretionary scalps measured in seconds of decision time, those are not anecdotes. They are process breaks.
Mandatory VPS, for me, means: the trading terminal and the flatten hotkey live on a London-region box that I remote into. Charts can still be local. Orders are not. I tested failover twice: kill the local session, confirm the VPS session still manages the open ticket. If that drill fails, I am not live the next morning.
Local-only is acceptable for journal work, for watching Asia, for building the night-before sheet. It is not acceptable as the only path to flatten when EURUSD is in a Tier-1 spike. That is the whole argument. Not latency bragging. Just one place where the kill switch still works when the house connection does not.
I still keep a phone login as a last resort, but I do not pretend that is a plan. The plan is VPS first, local second, ego nowhere near the order ticket.
Why I chose VPS mandatory over local machine acceptable in a scalping context
-
LondonNewsTrader
- Posts: 421
- Joined: Mon Sep 21, 2026 9:30 am
-
Shadow Trader
- Posts: 959
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Why I chose VPS mandatory over local machine acceptable in a scalping context
Kill the local session and confirm the VPS still manages the ticket is the drill most people skip. I failed it once because the VPS terminal had auto-locked and my flatten hotkey lived only on the laptop. That morning I was not live until it was fixed.
Charts local, orders on the London-region box is exactly how I run it. Latency bragging never interested me. What I care about is that the flatten still works when the house ISP blips at 08:02, which it has.
Phone login as last resort is fine if you have actually timed a multi-ticket close on mobile data. Mine took over two minutes for four positions on one platform with no close-all button. That number changed how many separate tickets I allow when the VPS is the primary path.
Local-only for journal and Asia watching is where I draw the line too. The moment risk is on, the order path has to survive a home outage. Ego near the ticket is how people keep trading through a freeze instead of failing over.
How often do you re-run the failover drill? Mine is every Sunday evening before the week starts.
Charts local, orders on the London-region box is exactly how I run it. Latency bragging never interested me. What I care about is that the flatten still works when the house ISP blips at 08:02, which it has.
Phone login as last resort is fine if you have actually timed a multi-ticket close on mobile data. Mine took over two minutes for four positions on one platform with no close-all button. That number changed how many separate tickets I allow when the VPS is the primary path.
Local-only for journal and Asia watching is where I draw the line too. The moment risk is on, the order path has to survive a home outage. Ego near the ticket is how people keep trading through a freeze instead of failing over.
How often do you re-run the failover drill? Mine is every Sunday evening before the week starts.
-
Shadow Trader
- Posts: 959
- Joined: Wed Sep 09, 2026 5:19 pm
Re: Why I chose VPS mandatory over local machine acceptable in a scalping context
LondonNewsTrader, the failover drill change I made was putting the flatten hotkey on the VPS terminal itself, not only on the local session. Killing local used to leave me with a locked VPS box and no keystroke path until I remembered the on-screen button.
I re-run that drill every Sunday and once after any VPS reboot or platform update. The phone stays last resort, timed on mobile data, never the plan for a multi-ticket book.
Charts can stay local. Orders and the kill path stay on the London-region box. That split is what survived the house ISP blip, not a latency contest.
Ego near the ticket still shows up as staring at a frozen local chart instead of failing over. The drill exists to make failover boring muscle memory.
One London box plus phone is enough for my stack today. Curious whether you keep a second VPS login on a spare device for the same account.
I re-run that drill every Sunday and once after any VPS reboot or platform update. The phone stays last resort, timed on mobile data, never the plan for a multi-ticket book.
Charts can stay local. Orders and the kill path stay on the London-region box. That split is what survived the house ISP blip, not a latency contest.
Ego near the ticket still shows up as staring at a frozen local chart instead of failing over. The drill exists to make failover boring muscle memory.
One London box plus phone is enough for my stack today. Curious whether you keep a second VPS login on a spare device for the same account.