Frequently Asked Questions

Equity Strategy

The strategy only goes all-in when all four factors are bullish and price volatility has been suppressed for some time.

My strategies differ significantly from how Turin has coded his. I built my system to model ideas before the Global Financial Crisis and include the dot-com era. Unfortunately, Turin’s strategies struggled during the dot-com bubble—shorting on the way up and going long on the way down. My strategies might lag somewhat because I’m targeting a longer time frame and aiming for a specific Sharpe ratio and annualized volatility, not just total return.

The recent scale-up and scale-down were due to price volatility. At the 50% allocation we saw this week, all strategies are still long but with half allocation due to volatility. The system isn’t very bearish but isn’t typically concerned about 3-5% pullbacks unless they precede larger market deterioration. It scales down to reduce notional exposure and aims to provide a smoother return curve over time.

The strategy doesn’t typically buy at market lows; it buys into strength when the market starts gaining momentum. You’ll notice long exposure increases after a bottom has been established. Buying when the day ends on lows isn’t advisable, as sellers are in control.

I haven’t tried those yet. The purpose of my current strategy was to model following all S&P sectors and equal-weight them instead of using the S&P’s weighting. You’re probably right; there’s potential with the ETFs you’ve listed. I can consider building a separate program to incorporate them.

They worked well and provided some protection. Because the equity strategy went from 62.5% to 75% after the spread gained value, when it cooled off the next day, it was a favorable trade-off.

Defense Strategy

Yes, for both Crypto and Defense, the benchmark is simply an equal-weighted buy-and-hold of the underlying traded assets, rotated quarterly.

No, it’s not a typo. Bond momentum has been down, so if I were allocated to bonds as part of a portfolio, I’d be long TMV (3x bear bonds). I’ll be releasing a blended portfolio model this week that will trade equity and defensive strategies together to demonstrate how this will work.

The Defense strategy operates without the equity allocation condition. You can think of it as an individual portfolio overlay, allowing for customization. It’s relatively uncorrelated because it’s largely a mono-strategy on uncorrelated assets. In the AHF Index, we won’t position into the Defense strategy until the equity allocation is ≤25% for three days.

The Defensive Strategy is a multi-asset, long/short trend-following system designed to systematically hedge against inflation, recession, currency, and policy risks. This strategy involves trading the following ETFs:

Gold: DGP, DZZ
Oil: UCO, SCO
USD: UUP, UDN
Bonds: TMF, TMV

Each asset is allocated a maximum of 25% of the available capital. Although allocations are equal across assets, the leverage on each ETF differs to provide blended exposure. Each asset encompasses four different trend approaches, diversified in investment timeframe and methodology. Depending on trend strength, you will see asset allocations of -25%, -18.75%, -12.5%, -6.25%, 0%, 6.25%, 12.5%, 18.75%, or 25%. When allocations are negative, you would trade the “Short ETFs.”

The strategy’s mechanics mirror our other methods. When allocations change, the model anticipates you to make a purchase the following morning. An asset will change sizing based on building or deteriorating trend strength or if daily trailing stops are triggered (these are present on every asset).

The 25% is read as ≤25%. I grouped it that way because that’s how the strategy operates. Essentially, as long as the allocation is ≤25% for three or more days, the Defense strategy becomes active. The longest streak in this state is approximately 134 days.

It depends on the asset:

Bonds: 3x leverage
Gold: 2x leverage
Oil: 2x leverage
USD: 1x (no leverage)

TLT, as represented in the Defense strategy, is still fully allocated, signaling strong confidence. There are four separate signals for each asset, so if momentum starts to stall or break down, we’ll reduce exposure accordingly.

CTA Strategy

The AHF Index is equity-heavy and only trades non-equity assets when the equity allocation is ≤25% for multiple days (think extended drawdowns in the S&P 500). The CTA model has no equity component; it’s a long/short commodity-based model.

My intention is to provide you with the components you need to build a portfolio. I didn’t include long/short equities in the CTA sleeve because shorting the equity market doesn’t perform well. I’d prefer questions like: “What does a 50% equity + 50% CTA portfolio look like?” Most institutional allocators wouldn’t put more than ~15% into CTA products.

There are a couple of reasons, mainly for efficiency:

Futures can be traded almost around the clock, allowing you to avoid the slippage from waiting for regular trading hours to get decent spreads.
Some of these ETFs don’t have short equivalents, so you’d have to short the stock, incurring short interest and reduced liquidity.

It’s a four-part trend-following system with diversity in both assets and trend signals:

Pure momentum/slope of price
New highs and lows
Moving average crossovers
True strength index

All four are equally weighted, binary long/short signals, blended into an overall allocation (10% per asset).

The returns are based on the leverage from allocating to futures products, as several don’t have ETFs or have highly illiquid ETFs.

Long silver, flat on gold.

AHF Index

Up to 16% per component.

In 2022, there were 105 days with zero allocation. Additionally, the defense portfolio activates when allocations are below 25% for three days. This strategy effectively trend-follows risk-off assets, which tend to perform well during significant market drawdowns (observable in 2008, 2020, 2022). I built the defense strategy separately and found that overlaying it on top of the Equity strategy was highly beneficial.

Gold: DGP, DZZ
Oil: UCO, SCO
USD: UUP, UDN
Bonds: TMF, TMV

The strategy never shorts. Each component in the strategy has a trailing stop, tuned individually and looking as far back as possible (in some cases only to 1998, in others to 1970). Because the components are decoupled (they don’t rely on each other for signal generation), they tend to operate independently, which helps control drawdown even further. Therefore, risk exposure is basically uncorrelated.

Greedy Strategy

No specific rules like that—it’s back to a full portfolio on Monday. That’s why I like layering the two strategies: 50% Equity / 50% Greedy. You always have an allocation unless it’s truly risk-off.

Yes, that’s correct. The Greedy strategy is a derivative of the Equity strategy. While the Equity strategy uses a volatility-scaled approach to sizing, the Greedy strategy simply trades in and out at 100%. This leads to higher annualized volatility and larger drawdowns but significantly boosts returns. Blending Equity and Greedy together results in a very attractive portfolio.

The “Greedy” algorithm goes fully long any time our aggregate positioning from the Equity model increases by ≥12.5%. Likewise, it goes fully flat on any reduction in positioning of ≤-12.5%. The returns are quite impressive, and this is still based on SPXL with no other fundamental changes to the model.

My focus across most strategies has been to find uncorrelated and/or low-volatility returns to layer against our default allocation of being “long the world.” This strategy increases volatility significantly. I heard a great quote recently: everyone calls this game “investing,” but it’s really “saving.” The key is to save more over a long enough time horizon and have cash when you need it. The trick is not to capitulate.

I’m considering that this strategy might pair well with buying call spreads. The average daily return of this strategy is 0.15%, and trades last about six days. You could effectively buy a call spread with the bought leg at the money and the sold leg at 0.9% away. You might roll the spread every five days until the signal reduces. I haven’t backtested this concept, so please take it with caution.

Here’s how it works: Suppose we’re at a 37.5% allocation, and the model is 50/50. You split your portfolio into two 50% segments:

First segment: Long SPXL at 18.75% (which is 37.5% of this half)
Second segment: Long SPXL at 50% (full allocation of this half)

For a $100K portfolio, it would look like:

Second segment: $50K × 1.0 = $50,000
First segment: $50K × 0.375 = $18,750

Allocation and Position Sizing

25% SPXL (75% SPY notional), 16% TMF (48% TLT notional), 8% UUP.

Scaling works as follows:

Scale Down: Base the reduction amount on the allocated capital.
Scale Up: Base the percentage on the full balance.

It equates to $37.5K (37.5% of $100K), giving you a little over $100K in notional exposure to SPY.

Good question! My labels refer to the non-levered assets. The model is built on TMF. So it’s 16% of the overall portfolio in TMF, which equates to 48% notional exposure to TLT.

Execution and Rebalancing

I didn’t really test different execution strategies. The platform works by:

Grabbing end-of-day data around the futures market close.
Recomputing strategies to determine if positions will change.
Assuming you know the position changes today and execute at the market open tomorrow morning.
Accounting for some slippage (about 0.05% per trade) for non-levered S&P-based trading.

Entries and exits are treated the same. If you can trade futures, you can reduce slippage and the close-to-open gap while also gaining tax advantages, as the equity strategy tends to trade around 70 times a year.

The position should be reduced by one-quarter. We don’t rebalance on position changes.

That’s a straightforward question! We rebalance quarterly, so you don’t need to adjust until the end of the quarter (which is somewhat arbitrary—see Corey Hoffstein’s “Rebalance Timing Luck” paper).

My intuition suggests performance would improve slightly over the long term. You’d reduce slippage and react to down moves faster by avoiding overnight moves.

No, I simply roll about a week before expiration to the back month. Platforms like Tasty allow you to do this seamlessly.

Futures and ETFs

There are a couple of reasons, mainly for efficiency:

Some of these ETFs don’t have short equivalents, so you’d have to short the stock, incurring short interest and reduced liquidity.
Futures can be traded almost around the clock, allowing you to avoid the slippage from waiting for regular trading hours to get decent spreads.

Yes, GLL is acceptable and has a 93% positive correlation to /GC, which is the basis of the underlying model.

No, I simply roll about a week before expiration to the back month. Platforms like Tasty allow you to do this seamlessly.

Crypto Strategy

Yes, for both Crypto and Defense, the benchmark is simply an equal-weighted buy-and-hold of the underlying traded assets, rotated quarterly.

My overall thesis is achieving similar returns with less volatility. While there’s some tax alpha loss, if this is just a sleeve in your overall portfolio and you can rebalance somewhat, rotating that capital into less volatile allocations will reap long-term benefits.

Strategy Mechanics and Concepts

It’s a four-part trend-following system with diversity in both assets and trend signals:

True strength index
Pure momentum/slope of price
New highs and lows
Moving average crossovers

The position should be reduced by one-quarter. We don’t rebalance on position changes.

Scaling works as follows:

Scale Down: Base the reduction amount on the allocated capital.
Scale Up: Base the percentage on the full balance.

Here’s how it works: Suppose we’re at a 37.5% allocation, and the model is 50/50. You split your portfolio into two 50% segments:

First segment: Long SPXL at 18.75% (which is 37.5% of this half)
Second segment: Long SPXL at 50% (full allocation of this half)

For a $100K portfolio, it would look like:

Second segment: $50K × 1.0 = $50,000
First segment: $50K × 0.375 = $18,750

The strategies don’t use any logic like “I was stopped out; should I re-enter now?” Positioning is purely a reaction to underlying factor strength (or lack thereof), so it re-enters as soon as strength returns.

When I refer to the market being “extended,” I’m describing a condition where the market has been in a sustained upward trend for a considerable period without significant pullbacks or corrections. This prolonged rise leads to strong bullish signals in our underlying factors, such as momentum and trend strength indicators.

In our strategies, these factors are designed to capture and respond to persistent market movements. As the market continues its upward trajectory, these indicators become deeply entrenched in bullish territory. This means that minor fluctuations or short-term declines aren’t sufficient to alter the overall positive signals generated by the models.

Because the bullish momentum is strong and well-established, it requires a more substantial shift—like significant selling pressure or a notable increase in market volatility—to reverse these signals. Only such pronounced changes would indicate a potential trend reversal or heightened risk, prompting the strategies to scale back on risk exposure.

Portfolio Construction

Yes, they definitely are! The only one that might be challenging for smaller accounts is the CTA strategy, which I haven’t automated yet.

My main reservation about providing Nasdaq (NQ) exposure by default is that it could be overfitting to the best-performing sector historically, introducing sector concentration risk. The S&P 500 already performs some sector rotation for us through its rebalancing and sector inclusion.

Market Outlook and Conditions

Let’s discuss performance first. Here are the 9-day, 21-day, and 100-day distribution curves for SPY returns when our equity model is ≤25%: it’s “left-tailed.” Conversely, for ≥75% allocation, it’s “right-tailed.”

Regarding allocation churn (averaged since 2005), if you were to ask me—not as financial advice—I’d echo what Joel said: a large portion of your portfolio should probably be parked in cash. At a 25% allocation, if trading SPXL, you’re still notionally long 75% of the market, so you shouldn’t experience FOMO if it continues higher. The 25% allocation would likely adjust quickly if that happens.

Risk Management

They worked well and provided some protection. Because the equity strategy went from 62.5% to 75% after the spread gained value, when it cooled off the next day, it was a favorable trade-off.

Performance Metrics and Benchmarks

Yes, for both Crypto and Defense, the benchmark is simply an equal-weighted buy-and-hold of the underlying traded assets, rotated quarterly.

The returns are based on the leverage from allocating to futures products, as several don’t have ETFs or have highly illiquid ETFs.