I've spent a lot of time working on a trend following system this year, and started trading it in November. The amount of retail systematic trend followers is small, and I enjoyed reading @globalarbtrader 's journal so I figured I would share my journey occasionally as well. Me: Tech background when I used to work a day job. Always been part time in the markets in one way or another since I bought my first paper share certificate when I was a teenage McDonalds worker. Been full time self sufficient since COVID, and gradually moved to systematic trading. I've done alot of futures seasonality trading, with spreads and direct directional swing trading but this was very manual. My goal is to combine this with systematic trend following. Software: RealTest & CSI Data (though I use Norgate data for stocks). RealTest's futures capability has been under rapid development by Marsten Parker during this process. He's been amazing in implementing everything and helping with all the hurdles along the way. I can't recommend any other software as much as this. Design ideas are all learned from information that others have shared. Turtles of course, Marsten Parker, both Moritz's from Takahe capital, Rob Carvers' writings, Andreas Clenow, Jerry Parker, Paul Mulvaney, Tom Basso, and the great podcasts (Top traders unplugged, Algorithmic Advantage). So much good information out there these days. System: Classic style binary trend 1 entry, 1 exit, 1 stop, but diversified into 3 systems short, medium, and long term. Moving average for trend direction, breakout for entry rule. My own seasonality rules that I used to use manually. It won't get used on all timeframes so the system won't miss a trend, but it's particularly useful for shorts to not enter during a seasonal bullish period. Universe: ~100 futures markets. Any CSI data market offered by IBKR that has enough volume and can accept API execution. Also as many micros as I can get for granularity. Backtest swaps full size to micro on appropriate dates. Holding time for winners is tested to be approx 50, 75, 110 bars for short, med, long. For losers, exit on stop or reverse breakout rule. Position sizing based on ATRs. Stops based on ATRs. ATR multiplier fanned out to different distances to allow different timeframes of the 3 systems. Unrealised gains are not marked to market until closed, position sizing risk% will only spend initial capital and closed equity on new positions. When looking at the drawdown on the backtest keep this in mind - it does not show the pain I will feel when giving back profits (this I have already started to feel giving back half of several positions). Backtest is done without compounding, so that every year/regime was tested on an equal footing. Note that some of these metrics won't make sense for futures, as it's also a stock testing system. But it gives me a basic idea on what to expect, assuming an even higher drawdown in live trading will come. Execution: RealTest orders, with its order management tool that connects to IBKR API. Once daily, about 2 hours after the markets open (it doesn't affect results to be late).
Trading started on 6th November. As per advice from Andreas Clenow's book, and my own testing - all current positions from the backtest were entered from the first day to ensure my holdings would match the backtest. Diversfiication has been great from this respect. Current closed & Open positions, as of yesterday's market close (17:00 NY time). The #Positions represents how many systems were in, short + medium + long etc. BarsHeld are market days, eg 5 bars is 1 week. Coffee was quite painful, giving back what was quite a decent open profit but that's the price to pay if I want the potential to catch a longer trend.
Number of contracts formula is: PercentRisk * AccountAlloc / RiskPoints. RiskPoints is the value in USD of X number of ATRs (X is the average distance to my stop losses, approx 4). Percent risk is between 0.4 and 0.9 percent depending on what it is, how risky it is or how many I have. What to trade is the entry rules: on the right side of moving average, and passed a breakout rule of one of the 3 systems eg Donchian channel for the long term system. I've been experimenting with different rounding rules, to allow me to enter a contract if the position size is < 1 but closer to 1 than some threshold. Possibly the diversification benefit is worth the larger size on some positions, but there will definitely be a few contracts that are too large.
I added contract rounding up for the sizing of new positions: in the case where sizing is between 0.75 and 1.0 contracts the system will now take the 1 contract instead of skipping the trade. Made a change to remove BZN (Brent Crude on NYMEX CLEAR). I already have ICE Europe's Brent Crude, and NYMEX CLEAR data comes in almost an hour later than everything else. Out of 100 markets, I think it's not worth waiting an extra hour for just 1 market. I do think it's good to have multiple versions of the same commodity like Robusta and Arabica coffee, but in the case of Brent the ICE version is priced in USD and basically the same price as the CLEAR version. So I rolled my BZN position into COIL (ICE Brent), removed BZN from the universe and swapped it out in my trades CSV records. Also fixed a bug fix with CSV playback of trades for forex conversion, and roll costs/commissions was done so some of the closed trade losses changed ~$900 less loss. New signals for today: Short, Short term: Exit Rule 6x Mexican Peso (executed) Long, Short term: Exit Rule 12x Lumber (queued for open) Long, Medium term, Enter 1x TSE 60 Canadian stock index (executed) Long, Long term, Enter 1x Cocoa (executed) Positions from 17:00 market close (doesn't include new trades until tomorrows data):
This ends Month #1 of the system, nothing but ecstatic about what the market did to my positions yesterday. Without a doubt the biggest one-day upside volatility I've had so far. Strongly aware this is incredible starting luck for this system.. other systems (stock systems, mean reversion) sometimes had much worse starting luck for me. Based on prior experience, I feel some of this will be given back come Monday open. Strong single day moves like this often revert a bit. P&L was actually $10k higher intra-day and reverted slightly towards close. Mondays orders: Roll UC (Chinese Renminbi) from Dec to March Roll both MP (Mexican Peso) from Dec to March Roll both RY positions (Euro/Yen cross) from Dec to March Roll KMD (Mini Korean Wong) from Dec to Jan Roll RP (GBP/EUR Cross) from Dec to March Roll SEK (Swedish) from Dec to March Roll EURO from Dec to March Bit of a pain to get rolls on the weekend, because I roll manually I normally prefer to do it before submitting the new contracts stop losses into the market.
Had some thoughts on the weekend that gains would pull back, I was expecting it to be Coffee and Cocoa pulling back. Turns out it was Bitcoin and the Stock market pulling back, Cocoa had one of the biggest gains. It was then the Chinese stock market (China Enterprises contract) that finished carrying the portfolio to make up for Bitcoin and Index losses for the day. That position has been sitting there since day1 and then unexpectedly kicked into gear. Should never trade off predictions & expectations.. CAD short rolling today to March, most of these short currencies are paying a credit of $USD for the price difference. Short RS (Canola) exit rule triggered due to volatility breakout against me. 3rd position triggered for Micro Bitcoin, not very confident about this one at the current price but I already sold my discretionary BTC bet at around $100k - and I've learnt not to try to predict.
There is a prob;em with your journal... It's fake. It is Monte Carlo BS. Why don't you come back with a real money trial you POS?