Trend Following — Live Status
Live operation record of an automated bot that enters on volatility breakouts and, with a moving-average filter, excludes from its buy candidates any market whose opening price that day is at or below its moving average.
Daily Cycle
Today's Breakout Board
Once a day, at midnight KST, the bot decides market by market whether to trade. Only the ones that pass wait for a breakout that day.
Dots: fills today
- BTCSitting out
- ETHSitting out
- XRPSitting out
- SOLSitting out
- NEARSitting out
- SUISitting out
Prices and charts are based on Bithumb 60-min candles.
Open Positions
No open positions right now. Next midnight check in --:--:--
Recent Trades
This system switched to live trading on August 21, 2026. Records before that date are paper trades, marked with a PAPER badge in the history below. Live statistics count only fills executed after the switch, so the sample is very small at this stage. Win rates or average returns based on too few trades are not a meaningful measure of a strategy.
Performance
Fees include buy-side fills, so a nonzero fee can appear with zero closed trades.
Earlier records 3
How the bot behind this page decides
This dashboard shows exactly what the live bot is doing right now. What follows explains the rules behind the numbers on this page — not trend-following theory in general, but the actual conditions written into the code that runs this page.
When it buys — volatility breakout
Every day at midnight it computes one reference line for the day: yesterday's high minus yesterday's low, multiplied by 0.4, added to today's opening price. If the current price crosses above this line, it buys. If not, it does nothing for the rest of the day.
The 0.4 multiplier isn't an arbitrary number — it sets the breakout sensitivity. A lower value triggers buys on smaller moves, which means more false signals; a higher value can miss real upward moves entirely.
This reference line is calculated only inside the bot and is not shown on this page.
Most days, it buys nothing
A filter runs before the breakout check. If the day's opening price is at or below its moving average, the target price isn't even calculated, and that market is skipped for the day.
During downtrends, price often spikes briefly above the breakout line — most of those spikes are just short bounces inside a larger decline. The moving-average filter blocks those signals before they can trigger a buy.
As of this writing, several watched markets are sitting behind this filter — the ones marked "Sitting out" on the breakout board above. Buying nothing is not a malfunction; it's the filter working as designed.
When it sells — three exit conditions
There are three exit conditions, checked in a fixed order of priority.
First, once the peak unrealized profit reaches 5%, a trailing stop switches on. From that point, giving back 3 percentage points from that peak triggers an immediate sell. For example, if profit climbs to 8% and then drops back to 5%, the position closes.
Second, if the trailing stop hasn't switched on yet, a 6% drop from the entry price triggers a stop-loss. Once the trailing stop activates, this fixed stop-loss line no longer applies — the trailing stop fully replaces it.
Third, even if neither condition above is hit, the position is closed at the next day's opening price once the date changes. This bot never holds a position for more than a day.
When a position is open, the open-position card above shows its entry price, current price, and unrealized P&L.
A stopped-out market sits out for two days
Once a position is closed by a stop-loss, that market is blocked from re-entry for the next two days. A stop-loss means the market didn't move the way the setup expected — re-entering the moment a new breakout signal appears the next day makes it easy to repeat the same loss.
The results so far aren't good
The live-trading stats at the top of this page show a win rate of 48.2% and a cumulative return of +20.25% (per-trade average +0.24%). These returns include Bithumb fees (0.04% per side). Those figures only count fills executed after the switch to live trading, so the sample is very small — far too small to judge the strategy's real performance.
Trend following is inherently a low-win-rate strategy. It takes many small losses and relies on a handful of large winners to carry the overall result, so a run of consecutive losing trades falls within its normal range. How long that run lasts can't be known in advance, and nothing here promises the numbers on this page will improve going forward.
That's also why after-cost figures — fees and slippage included — are shown alongside gross returns: looking only at gross returns makes real performance look better than it actually is.