I've been running AUTO since I started, but I keep seeing references to Manual mode. What's the actual advantage? When would I use it over just letting AUTO run?
Manual mode is one of the most underrated parts of the bot. AUTO is great for hands-off volume, but Manual puts you in the driver's seat with the same engine underneath:
- Full visibility on the Pool. Every active option shows its fair value (the ensemble's estimate) next to the current market price, with the edge and confidence. You see exactly why each entry is or isn't attractive.
- You approve each entry. Nothing goes in without your click — take only the setups you like.
- Focus on the cities you understand. You can narrow to a handful of cities, or even a single one, and follow it closely. Some operators run one city they know well and simply check in as the picture updates.
- Fair values update with every model run. The ensemble re-computes as GFS / ECMWF / ICON / NOAA release new runs through the day. An option that wasn't attractive this morning can develop a clean edge after the next run. In Manual you watch those fair values shift and step in exactly when the edge opens.
It's ideal for operators who want to learn the rhythm of the markets, or who prefer a smaller, hand-picked set of positions over broad AUTO coverage.
I run almost entirely Manual now. My workflow: I picked four cities where the tool consistently shows the cleanest, most stable edge, I keep an eye on the Pool a few times a day, and I approve entries only when a fresh model run pushes the fair value well above the market. Slower than AUTO, but every position is one I actually understand. P&L has been steadier for me this way.
How do I know which cities the tool is strongest on? Is there a list?
There's no fixed list — it depends on station data quality and how tightly the models agree for each city. The practical method: run for a bit and watch which cities consistently surface HIGH-confidence entries with a solid edge in your Pool. Those are the ones where the ensemble maps cleanly to a reliable observation station. Narrow your City Filter to those.
We're also building machine-learning auto-tuning that will weight cities automatically based on realized performance — that's on the roadmap. Until then, manual city selection based on what you observe in the Pool already works very well.
+1 on that. Even a single well-chosen city, checked consistently as the fair values update, can be a complete strategy on its own. You don't need to trade everything to do well.