September was all in the first 10 days, 1%/day is all you need right?
August - after 3 stale months finally had a good one, my best systematic month actually. For some reason August is always good, last year came off a DD.
Can't control the market, can only control myself so I just spent the flat months working on improvements so that the system would be ready when the trends came. Made a bunch of changes, too many to list.
Also some new dynamic reports that help me figure out performance attribution for date ranges without needing to manually work with data.
Productivity is higher than ever. My need for a rest is also. Next week is always going to be the week I stop working on the system and take some time off.. but next week keeps getting moved to the week after!
3
29
1,391
Going back a lot of years, hard to fit them all into just 9..
All PC "IBM Compatible" as they were called back in the 90s.
1
2
474
If you're benchmarking against S&P, then investing in stocks is what you'd need to do to track it and SocGen is not where you'd look for managers to give you that beta.
CTAs are selling a lego block for portfolio building, and part of the reason CAGR is so low is that allocators keep rewarding funds that are low vol with more allocations. So the vol and return keeps getting lower and lower.
But it's not just vol, CTAs do things in a way to maximise capacity. Which means stock indices actually get an over-allocation because they're extremely liquid. As do fixed income and currencies. But trend tends to buy indices too late, sells too late, and only gets a portion of the middle. Indices don't produce enough convexity or outliers to make up for that when compared to something like commodities. If they have non-trend strategies they can get a bit more out of indices.
When I say indices, it's global - Europe, Asia, ASX, everywhere that has a futures contract. So comparing just the S&P/Qs might look better than those if using recency bias in a backtest.
And this is just my retail trader/allocator opinion of course, and I'm less an allocator and more a trader these days:
x.com/Dan_Trend/status/19952…
Trend-following funds have delivered mediocre returns, but why?
Societe Generale, the French bank, collects performance data on CTAs (Commodity Trading Advisors). See next post for the link to the live data.
The most interesting data is the SG Trend Index, which is the average performance of the top 10 largest pure trend-following CTA funds.
From January 2000 to August 2026, the SG Trend Index returned 336% in total, or 5.7%/year. The period between 2000 and 2009 was particularly beneficial for trend-followers, as equities underperformed and commodities rallied.
And yet, trend-following CTAs still underperformed the S&P 500. Those from the industry like to talk about negative correlations and positive skew, but overall performance still matters.
For reference, a simple buy/hold portfolio (TiltFolio Balanced) delivered total returns of 660% or 7.4% over the same period. And this period included the worst drawdown for TiltFolio Balanced ever, i.e. the 2022 bond bear market.
On the TiltFolio blog, I've published some theories as to why trend-following funds haven't delivered higher returns. My main hunch is that they underinvested in stocks.
4
1
43
10,465
Some old discussion
A few thoughts with trader friends about managed futures funds, taking some guesses at why many (and the CTA indexes) underperform expectations.
Maybe some thoughts are academically "wrong", but critical thinking is valuable.
Looking at Nilsson index: where is the Cocoa effect?
6
576
Some commodities vs stocks arguments. We all have our biases though, I think they're both good to trade but in different ways.
Some downsides to that argument:
1. You can't leverage heavily into stocks, because they all fall together. Commodities have lower correlations so you can take on more risk at once.
2. You qualified the stock selection in order to see those results, out of thousands to choose from it's not trivial to get the selection right. For commodity futures you can realistically take every trade and not be forced to choose from thousands.
3. Usually limited to bull markets. Even if you want to short, there's many problems shorting stocks. Futures have equal longs and shorts with no limitations. Some cluster, but usually there's at least some good markets somewhere.
As an example TTF natural gas might have only gone up 150% since March, but I didn't commit the full notional value to the position, just a fraction of that as margin. Cocoa is a more extreme example in 2023-2024. Sure stocks can use margin too, but doing that on multiple positions is like the same bet repeated as soon as the index takes a downturn.
While some commodities (eg Lumber) fell to lows, it didn't affect those that were in a bullish phase. If these positions were stocks, they'd mostly lose at the same time.
What you say is true if you only want to pick one trade. It's also true about exponential potential, but so is buying lottery tickets.
If you trade more than one thing at once though, sticking to just stocks is severely limiting yourself.
There's some value in specialisation, but it's good to know the trade-offs.
1
6
720
Competitions look useful to build a record, but being based on % not Sortino they can psychologically push you into leverage/risk you wouldn't normally take.
USIC has a +$1m division which can reduce the desire to go nuts for your competitors, but % still not the ideal metric.
That is the problem with these challenges. They really have very high survivorship bias, and beginners get the impression that 80% after a few months is the norm. They do not see the volatility that is used to produce the results and the fact that these people usually blow the account just a few months later.
2
9
1,203
August - after 3 stale months finally had a good one, my best systematic month actually. For some reason August is always good, last year came off a DD.
Can't control the market, can only control myself so I just spent the flat months working on improvements so that the system would be ready when the trends came. Made a bunch of changes, too many to list.
Also some new dynamic reports that help me figure out performance attribution for date ranges without needing to manually work with data.
Productivity is higher than ever. My need for a rest is also. Next week is always going to be the week I stop working on the system and take some time off.. but next week keeps getting moved to the week after!
July was a pretty quiet month in the first half, then things got a bit wild towards the end!
1Y return increased even though July is down, because my July 2025 was much worse and now drops off.
Challenging period for futures net, though some longs did ok. Total futures still down, held up by ETFs and equities.
Energy trading (refiner stocks) has its best month I think. Helped by a big earnings surprise on $PBF
Went live with new overnight 2 day MR strategies that sit in-between multi-day MR and 1-day MOC. Unfortunately they fully loaded up too early in the Nasdaq downtrend and had to book a loss the day before the reversion happened. It's the right exit plan on average, but unlucky time to add them.
14
52
4,427
Haha finally something a trend follower wants to see.
Usually it's "Coffee is so expensive it can't possibly go up any more", and I only just entered the trade...
Don't tell me stocks are expensive though 😆
1
9
782
Trading futures but worried about 1,000 barrels of crude being delivered to your doorstep?
Depending on your broker, usually you have to actually elect to take physical delivery intentionally. If you don't they will usually liquidate your position before you get to any date like first notice that would put the broker at risk. E.g. IBKR starts giving warnings of increasing intensity as you get closer.
The first notice dates and contract specs are all available from the various exchanges websites, e.g. SGX, CME, Eurex etc. But rarely do we need to check.
IBKR publishes dates when they're going to start warning or liquidating you: interactivebrokers.com/en/tr…
You can infer whether it is based on a calendar day of the month offset, or a trading days or calendar days offset from the last trading date. If you trade individual contracts you only need to calculate this once then save it for all future use for that market.
If you are trading from a continuous back-adjusted series, you might arrange the roll when the highest volume or OI flips between contracts. If you want to avoid the front contract for various reasons you might even be rolling before that. That volume flip happens because everyone else is rolling before the delivery window. You can automate this, or figure out an offset rule.
Often overlooked is that there's no reason you have to trade the front contract, it might be more liquid sometimes but there are disadvantages of being in the front like more contango, a trend that's less supply driven and less smooth, and sometimes higher margin use (eg VIX). there's so many choices on how to roll that it can be overwhelming though. You don't even have to roll and can just trade individual contracts if you like.
This is all for commodities, financials are quite different and usually focussed in the front month.
I assume you track FND to avoid physical delivery? I haven’t found an automatic way to get that information. How do you do it? Calculate it yourself?
1
11
1,077
In the systematic world, pyramiding can come from multiple systems firing on the same market. E.g. you may have an old Trend position that pulled back (keep stops wide), then you get a reversion entry.. then maybe something else eg Term Structure, Seasonality, Skew, COT, Fundamentals etc.
The position gets bigger, looks riskier, but it has a higher expectancy for all of those trades to win even individually because of all the other signals agreeing. That makes it worth the risk of being concentrated.
On a Sharpe ratio basis I haven't been able to make pyramiding with a single trend system look better (i.e. stronger signal + more time = add to position). But adding due to different uncorrelated signals does increase Sharpe and other metrics too. Some metrics can be improved (eg total profit per trade), especially pyramiding into stuff like mean reversion.. but have to be careful not to become a martingale trader.
E.g. my Rubber position is nowhere near my best trend, but it did enough different things in the lifecycle of the full trade to allow heavy sizing (I'm on the hook for a silly number of metric tonnes of rubber). So it's now in the top 5 profitable markets purely because it became a concentrated bet.
I think a real point people are missing with pyramiding is the fact that the best trades usually go without offering much of an entry, either you are in or often missing.
Furthermore the best trades don’t have one leg but often 3-5, they don’t just pop and die.
Pyramiding builds upon that understanding to exponentially size into an idea that presents those top odds.
It maximizes the overall expectancy of the trade and the linked low variance/volatility along the path of those best opportunities. Measured as total expected return vs volatility to achieve the target.
It maximizes the certainty about winrate at the cost of the buffer by using tight structural builds along the path and by maintaining a strong average set by the original anchor.
A balance between win rate off of the average, using the buffer, RR along the path both on the overall trade and the singular adds along the path that combine into one bigger trade.
The overall cost is win rate in the end, but one that lowers it marginally vs the gains it generates on the winners.
All in all the idea being being big on the best ideas with the most anticipated alpha rather than missing, too small and having little chance to pay for mistakes.
Trading is a game of breakeven for 80% of trades with 15% paying bills and perhaps 1-5% actually creating the Wizards… Make sure to find a way to create the hockey stick chart.
4
2
39
12,858
Mean reversion in my futures system is so different to stocks, it takes time and is not driven by daily noise.
Something satisfying about being rewarded for patience..
Had a mean reversion entry the day after declaring the trend dead. A 3rd position may be scary to take, but more likely to work out than not. Bounce took a few days but all 3 positions still alive so far.
This screenshot will kill it though? 🤔
2
27
1,399
AI won't touch my trading rules, but it's not bad for reporting.
Haven't checked market attribution for a while so this was interesting. I always felt Fixed Income was the worst but FI Short Term actually became profitable, and Currencies are now the worst.
For back-adjusted contracts the attribution is slightly wrong and makes some things (Live Hogs) look worse than they are. It's hard to get right because I have changed roll-rules after the trades, but it's close enough to get a general idea.
I've defined an "R" just for privacy reasons. It's not a stop loss or max loss, it's just a fixed unit close to my volatility sizing variable that's equal for comparison purposes on all trades.
1
18
787
Russell 2000 stocks didn't look so hot prior to 2025 but it's the best performing universe this year. #diversification.
Been seeing a few stocks like this that I've never heard of, do something enough times you make your own luck ;)
I had one of my best weeks this year. As luck should have it, I bought $ETON just before it jumped 44% yesterday.
11
897
Dan retweeted
While puts also cover gaps while the market is closed, stops are fully exposed to those gaps.
The place I found stops useful - intraday execution for a daily bar EOD trader. Similar same place in the toolbox as limit orders.
1
1
4
347
July was a pretty quiet month in the first half, then things got a bit wild towards the end!
1Y return increased even though July is down, because my July 2025 was much worse and now drops off.
Challenging period for futures net, though some longs did ok. Total futures still down, held up by ETFs and equities.
Energy trading (refiner stocks) has its best month I think. Helped by a big earnings surprise on $PBF
Went live with new overnight 2 day MR strategies that sit in-between multi-day MR and 1-day MOC. Unfortunately they fully loaded up too early in the Nasdaq downtrend and had to book a loss the day before the reversion happened. It's the right exit plan on average, but unlucky time to add them.
Ended financial year with a return that myself and the Australian Tax Office will be very happy with 😁 Negative June though, so I lost my 10 month green streak..
This is also a year since I switched from pure Trend Following to Multi-strategy, and 8 months since adding stocks. Diversifying in a single consolidated account was one of the best decisions of my trading career.
Long term commodity strategies entered pullbacks/whipsaw mode. Diversification cushioned the losses with futures shorts, and non-futures shorter term strategies filling the potholes.
Futures: Trend & Carry were the worst, MR & Season breakeven, and Skew slightly positive.
Equities: ETF strategies outperformed the backtest, VIX, sectors, country ETFs etc. Also been a great period for Russell 2000 smaller stocks compared to the usual large cap dominance. Stock Shorts were crappy as usual for me but they're not huge.
This is one of those periods where I'd be way ahead if I only traded stocks, but I prefer a diverse portfolio that's viable in all environments.
Have improved my execution & portfolio management software a lot, as well as data collection so I'm looking forward to the next month.
3
33
4,658