Transcription
This strategy beats the S&P 500 with a 27% annual return, a sharp ratio of 2.4. And get this, it uses just one indicator, one. It's called the forward factor strategy, and you've almost definitely never heard of it. It trades forward volatility using a calendar spread in a way that slashes risk and draw downs. I've traded the strategy as a core part of my portfolio for years, and it's played a major role in growing that portfolio into the mid-7 figures it is now. Today, I'm giving you everything. The 19-year back tests, a real trade example, and even the calculator so you can run it yourself. And once you see how simple it is, you'll wonder why no one's talking about it.
What is the trade? Here's the setup. It's a simple rule-based trade. Open a calendar or double calendar spread using one indicator. Then hold it until right before the front contract expires. That's it. No constant tweaking, no guessing games. The edge comes from a single indicator we'll define in a minute.
This idea comes from academic research showing a persistent bias in the term structure of implied volatility. Basically, the market's forward view of volatility is often off, and that miss can be harvested with plain vanilla option structures. The paper that put this on my radar is "Term Structure Forecast of Volatility and Options Portfolio Returns." It finds that the forward implied volatility built from two nearby expirations systematically misestimates what the short-dated volatility will actually be later. When that forward V is too low relative to the front period, being long forward volatility tends to pay. The good news, you don't need exotic products to express that. A calendar spread does the job.
So the core premise is simple. If the market's forward quote for next period's volatility were perfectly fair, then when the front contract expires and the back becomes the new front, its one-period volatility should match that earlier forward quote. It often doesn't. That gap is big and persistent enough to build a one-rule strategy around it. Our one indicator will tell us when to put on the spread, which we then hold until the front expiry and close.
Let's break this down step by step. Because to understand forward volatility, we first need to understand the mechanics behind how this trade works. Let's start with an intuitive explanation of forward volatility. Imagine volatility like rainfall over time. You have a weather forecast for January. That's your front-period implied volatility. Then you have a forecast for January and February combined. That's your back-period implied volatility. Now, from those two forecasts, you can actually back out how much rain is implied just for February. That February-only number is the forward volatility from January to February.
With options, we use the same concept, but instead of rainfall, we're dealing with variance. That's volatility squared. Why variance? Because variance over non-overlapping time periods add together, just like total rainfall over separate periods. That's why we don't average implied volatilities directly. Instead, we combine variance times time. So if the near-term expiry called T1 (say, around 30 days) and the next one called T2 (around 60 days), then the forward volatility between T1 and T2 is the unique number that makes the total variance line up correctly across those two windows.
Volatility represents standard deviation per year. And because the variances over separate time windows add up, just like our rain example, we always do the math in variance space, then square root the result to get back to volatility. We'll keep this practical. One formula, then an example. Don't get scared by the math. This will be straightforward to follow, and you'll never have to calculate it yourself. There's a free Python script in the description, or if you're not into Python, a free calculator on oquants.com.
Now, let's cover the mathematical calculation. Sigma 1 = the annualized IV for the front period with time to expiry as T1 in years. We let sigma 2 = the annualized IV for the back period with time to expiry T2 in years. And this also assumes that T2 is bigger than T1. We then define the variances, which are just the square of the volatilities. The annualized forward variance for the window between T1 and T2 is (V2 * T2 - V1 * T1) / (T2 - T1). Then the forward volatility itself is simply the square root of the forward variance. So we get back to volatility space.
Here's a key detail. Time must always be expressed in years. For example, 30 days is equal to 30/365, and 60 days is equal to 60/365. Let's do an example to make this stick. Let's say the 30-day implied volatility is 45%. Let's also say that the 60-day implied volatility is 35%. Our time in years is defined as 30/365 and 60/365. We then compute the variances by squaring the volatilities. Now we calculate the forward variance, which gives us a value of 0.0425. 0425, which is our annualized forward variance measure. Then we take the square root of that number to give us 20.6%, our annualized forward volatility between the 30 and 60-day expirations. Given those two IV points, the market is implying about 20.6% volatility for the 30-day period that starts after the front expires. That means that the expected volatility of the 60-day option in 30 days, when it becomes a 30-day option, is 20.6%. 6%, as currently priced by the market. You can repeat this process for any pair of expirations. For example, 30 to 90, 60 to 90, and anything in between.
So that's how we derive forward volatility. The market's implied view of volatility for a future window of time. Now, here's where it all comes together. This one ratio not only flips losing trades into winners, it also tells you exactly when to enter. It's called the forward factor.
All right, now that we understand forward volatility, let's talk about the one indicator that drives the entire trade. The forward factor, or simply FF. The forward factor tells you how hot the front-period implied volatility is relative to the forward implied volatility by the next expiry. In short, it measures the gap between what's happening now and what the term structure says comes next. And that gap, that imbalance, is the edge we're trying to harvest.
For all our calculations and all the research we'll discuss here, we use X-earn implied volatility. That's implied volatility with the earnings implied volatility removed. Here's why. X-earn IV is the option market's implied volatility after stripping out the extra premium tied to scheduled earnings announcements. Why do we use this? Because earnings cause a temporary, ticker-specific spike in near-term IV. If we didn't remove that spike, we'd be comparing apples to oranges: a front-period IV that's juiced by earnings versus a calmer, less effective back period. By using X-earn IV, we put all expiries and all tickers on an equal playing field. So, the forward factor truly reflects term structure imbalance, not just the result of a one-off event.
Does doing this kill earnings trades? Not at all. The calendar still works through earnings when the X-earn FF is elevated. In those cases, you're trading the underlying curve misalignment, not just the earnings event. However, research from the paper shows that these calendar spreads tend to perform better when avoiding earnings altogether. For simplicity and consistency, it's easiest for most traders to avoid trading this strategy through earnings. Meaning, no earnings events should occur between your entry and the second expiration. If that holds true, the X-earn IV will simply be the same as the regular implied volatility.
The forward factor is defined as (front-period IV - forward IV between the first and second period) / forward IV between the first and second period. The forward factor measures the relative difference between the front-period IV and the forward volatility. As we defined earlier, it quantifies how the near-term implied volatility compares to the implied forward volatility over the interval T1 to T2. It's expressed as a percent difference: how much the front exceeds or lags the forward.
So, how do we interpret this? If FF is greater than zero, the front IV is greater than the forward IV. This typically happens when term structure is in backwardation, signaling near-term stress or fear. Historically, this setup has shown very strong performance with a high sharp ratio and robust returns. These are our main focus in this video. If FF is less than zero, it means the front IV is less than the forward IV. This typically occurs when the term structure is in contango, which is the default state in normal markets. These tend to be less profitable, though still solid, just not as strong as the positive FF setups.
So why does this matter? When the front period is too hot versus the forward, a long forward volatility stance tends to pay off. And the beauty is you can approximate long forward vol with plain vanilla options. To go long forward vol, we can use a calendar spread. Sell the front where IV is rich and buy the back where IV is cheaper.
So why does a positive FF capture this mispricing? In panicky markets, traders rush to buy near-term protection or speculation. This drives up the front-period IV while the back period rises much less. When you compute the forward, the market's implied vol for the next one-period window, it often comes out much lower than the front. If that gap is large, it suggests the market may be underpricing future risk once the current stress rolls off. Your calendar spread is positioned to profit if either: one, front IV deflates into expiry, where you benefit from theta and IV crush on the short leg; and or, two, the back period gains exposure to the short-term risk as it becomes the new front, where your long leg benefits. The larger the FF at entry, the bigger the misalignment you're capturing.
Let's work through a high FF example using our earlier numbers from the forward volatility example, where we have a forward V equal to 20.66%. 66%, and a front-period IV of 45%. We can then compute the forward factor by calculating the percent difference between the front IV and the forward, giving us a value of 117%. The market is implying that the next 30-day window after the front expires will run at 20.66% implied volatility, even though today's front-period implied volatility is 45%. That's steep backwardation. The front is hot and the forward looks very calm. The bottom line: a 117% forward factor signals a major term structure misalignment. That's exactly the setup where a long calendar spread is a strong bet on rising forward volatility. A positive forward factor means the front is hot. We lean into long calendars.
Let me show you why that isolates the forward slice. A calendar spread, whether long or short, is fundamentally a bet on forward volatility between two expirations. A long calendar, where we sell the near period and buy the far period, is long forward vol. A short calendar, where we buy the near period and sell the far period, is short forward vol. Let's review why this is by revisiting our variance intuition. Remember, variance is volatility squared and adds through time. The back-period variance is equal to the front-window variance plus the forward-window variance. Whereas the front-period variance is just equal to the front-window variance. By selling the front and buying the back, you're effectively removing the front slice and keeping exposure to only the forward slice. That's what makes it a long forward volatility position. Flip it, and you're short that forward slice.
Let's explore the long calendar spread that is long forward volatility. Again, the position is: sell a near expiry (for example, 30 days) and buy a further expiry (for example, 60 days). The view is that the market is underestimating volatility in the next window. We want forward volatility to rise. So, what helps forward volatility? If front IV drops, this raises our forward volatility. The front cools off. If back IV goes up, that also raises our forward volatility. Future risk is repriced higher. Either way, we gain on the long calendar. What hurts is if the front IV goes up. This lowers the forward vol. Signals more near-term stress. Back IV going down also hurts the long calendar. This also lowers our forward vol, with the future risk again repriced lower. Either way, the long calendar loses.
Let's explore the short calendar spread that is short forward volatility. So again, the position is to buy the near expiry and sell the further expiry. The view is that the market is overstating volatility in the next window. So forward volatility should fall. Again, what helps forward volatility to fall is if front IV goes up; that lowers our forward volatility. Or if back IV goes down; again, lowers our forward volatility. What hurts is anything that makes forward volatility go up. So front IV going down is going to raise our forward volatility, and back IV going up is also going to raise our forward volatility. Either way, the calendar spread loses if that happens.
Both long and short calendars are path-dependent. Large price moves can push your P&L outside the tent, and outside that zone, your Greek exposures flatten, meaning your sensitivity to forward volatility fades. However, the main lever you're turning is still forward volatility. We want it to go up for a long calendar, and we want it to go down for a short calendar. That diminishing exposure outside the tent is simply a risk we accept when trading vanilla listed options instead of OTC products that trade forward vol or forward variance directly.
Okay, now that we know all about forward volatility, forward factors, and calendar spreads, let's hop into the real backtest results and data setup. We didn't just take the paper's word for it. We tested the idea with real investable spreads under realistic trading frictions. We rebuilt the forward volatility logic using options you can actually trade, then layered in slippage, commissions, and capacity limits so the results reflect what a live trader could actually see.
We ran the tests on two families of time spreads. The first is the long at-the-money call calendar. Buy the further-dated at-the-money call, sell the near-dated at-the-money call. Same strike, different expiries. This is the clean, classic calendar spread. We also explored the long 35-delta double calendar. We sell strikes off the front period's options chain around plus 35 delta for calls and negative 35 delta for puts. We accepted a plus or minus 5 delta difference if needed to hit the listed strikes at those same strikes by the back period options. Effectively, you're running two calendars at the wings: short front, long back at the 35-delta call side, and short front, long back at the negative 35-delta put side. This structure taps into skew. You're typically selling richer front-period skew and owning relatively cheaper back-period skew while widening the profit tent versus a single at-the-money calendar.
For both structures, we tested 3 DTE combinations, each with a small buffer so we don't skip trades due to listing granularity. We tested the 30/60, the 30/90 DTE, and the 60/90 DTE. Again, we allow a plus or minus 5 DTE buffer around each target. This means anything between 25 and 35 would count as 30. That keeps the data set realistic when the chain lists 29 or 32 days instead of a perfect 30 days. We only traded names with a 20-day average option volume above 10,000 contracts per day. This keeps fills plausible and capacity reasonable. Tests were run since 2007, covering nearly 19 years of options history. Plenty of cycles, panics, low-vol regimes, enough to make distributional statements with confidence. Slippage and commissions applied to every trade. We capped the number of simultaneous spreads per symbol based on open interest and trade data. So the backtest can't unrealistically soak up the book. This avoids the paper-only performance that vanishes when you size up. We entered when the target DTE and, for double calendar, the target deltas line up within the allowed buffers, and we enter the spread all at once. At exit, we close the entire position as a spread on the day the front contract expires, practically just before close.
So why two structures? The at-the-money calendar is the cleanest proxy for being long forward volatility at the center. The 35-delta double calendar adds skew harvesting and a wider payoff geometry, which can improve win rate and tighten the link between returns and our indicator when the forward slice is mispriced. Bottom line, this is not a toy simulation. It's a large, liquid, multi-decade data set of actual calendar spreads traded with buffers, frictions, and capacity constraints, and managed with mechanical entry and exit rules that match how you'd run the strategy in the real world.
For both data sets, we have 124,000 observations for the 30-to-60 DTE pair, 95,000 observations for the 30-to-90 DTE pair, and 88,000 observations for the 60-to-90 D pair. That's more than 300,000 fully specified, fully costed spread instances. Enough to make distributional statements with decent confidence. We can see from the return distributions that in all cases, just blindly trading all of these positions results in a negative mean return. That's an important baseline. The calendar is not a free lunch. If you don't condition on the term structure, transaction costs and slippage will grind you down if you fire every time. We can see that the 30/90 DTE spread has the best negative return, and that all distributions are positively skewed. That means mostly losers, but some larger winners. We also notice that the mean return of the double calendar is better than its at-the-money call counterpart, and its win rate is also higher. That tracks with expectations. We are selling skew in the front as well as expanding our profit area by making the profit tent of the calendar wider. This, of course, comes at the cost of more potential commissions and slippage due to more contracts being traded. The worst possible return is the debit paid, meaning 100%. But we will see that some are larger than 100%. And this is due to slippage and commissions. In practice, wide markets near expiry and paying spreads twice (entry and exit) can push measured losses a few points past 100%. This is a measurement artifact, not that the broker took more than your debit, but slippage and commissions turn the roundtrip cost into something a tad larger than the entry debit in some cases.
Looking at the returns grouped by quarter and plotted over time, the strategy looks relatively consistent with no obvious seasonality. Again, the 30/90 being the one that looks like it has the best return, although still mean negative. That's encouraging. The edge, if any, shouldn't depend on a particular period or quarter. Term structure dislocations occur throughout the cycle.
This is the hinge. Above simple FF cutoffs, returns flip positive. I'll give you the exact thresholds and a dead-simple rule. Looking at the scatter plot with a regression line of the calendar spread return versus the forward factor, we see a clear relationship. The higher the forward factor, the better the long calendar spread return. Visibly, we can see from the graph that a factor at or above around 0.1 to 0.2 is when returns start becoming positive. We can also see that this relationship is stronger with the double calendar. More evidence that the double calendar may be better to trade the strategy, along with the previously established better negative mean return. This is the signature we want. Conditioning on forward factor reframes the trade from "let's open calendar spreads and hope" to "harvest a repeatable misalignment." The fact that the double calendar slopes more steeply in the scatter signifies that the relationship between its payoff and forward factor is slightly stronger.
Looking at a plot of return on the y-axis and forward factor deciles on the x-axis, we can clearly see the relationship: that higher positive forward factors lead to better calendar spread returns. For the long at-the-money call calendar, we see that for the 30/60 pair, an FF above 0.14 leads to positive returns. For the 30/90 pair, an FF above 0.03 leads to positive returns. And for the 60/90 pair, an FF above 0.41 leads to positive returns. For the long 35-delta double calendar, we see for the 30/60 pair, an FF above 0.11 leads to positive returns. For the 30/90 pair, an FF above 0.01 leads to positive returns. And for the 60/90 pair, an FF above 0.14 leads to positive returns.
Let's create a model from these values and see how this changes our outcome. The simplest model is only to take trades when the FF is above the threshold we just saw for each DTE pairing and structure. Everything else stays the same. For the long at-the-money call calendar, taking trades only above the FF cutoff: the 30/60 now has a positive 9.2% mean return, a win rate of 47%, and an average of 50 trades per month. The 30/90 now has a positive 4% mean return, a win rate of 50%, and an average of 138 trades per month. The 60/90 now has a positive 39.7% mean return with a win rate of 49.5%, but an average of only four trades per month. For the long double calendar, taking only trades that are above the FF cutoff: the 30/60 now has a positive 7.9% mean return with a win rate of 54.6%, and an average of 62 trades per month. The 30/90 now has a positive 3.8%. 8% mean return with a win rate of 56.7%, and an average of 166 trades per month. The 60/90 now has a positive 16.9% mean return with a win rate of 51.1%, and an average of 36 trades per month.
A few important reads here. Number one, filtering flips the mean return sign for every bucket. This is exactly what we hoped for. Two, the 60/90 at-the-money call calendar bucket looks massive on mean return but very sparse on trade count. And three, the double calendars generally have slightly higher win rates at comparable thresholds. All of these models look good, but to get something more tradable, I would like to get this number to a point where the average of trades per month is around 20. Since even with a big portfolio, this will be plenty of trades, and we will likely increase our return by reducing the number of trades for all except the long at-the-money call calendar, where we actually need to raise our number of trades per month by lowering the threshold. This will likely reduce our return but decrease our variance and give us more trades.
For the long at-the-money call calendar, adjusting the FF threshold to around 20 trades per month, we get the following results. The 30/60 with an adjusted FF of 0.23. 23 now has a 13.9% mean return, a win rate of 47.8%, and an average of 20 trades per month. The 30/90 with an adjusted FF of 0.23 now has a 9.3% mean return, a win rate of 50.3%, and an average of 19 trades per month. The 60/90 with an adjusted FF of 0.2 now has a 24.7% mean return, a win rate of 45.4%, and an average of 21 trades per month. For the long double calendar, adjusting the FF threshold to around 20 trades per month, we get the following results. The 30/60 with an adjusted FF of 0.23 now has a 14.2% mean return, a win rate of 56%, and an average of 20 trades per month. The 30/90 with an adjusted FF of 0.23 now has a 10.4% mean return, a win rate of 57.9%, and an average of 19 trades per month. The 60/90 with an adjusted FF of 0.2. 2 now has a 22.6% mean return, a win rate of 55.8%, and an average of 21 trades per month.
We can clearly see that using the forward factor is a predictive signal of long calendar spread returns, with all time frames and positions being profitable. We also notice that the 60/90 seems to look best. Plotting the modeled mean returns grouped quarterly, we see a much more consistently profitable strategy than just trading them blindly. We also see that the 60/90 time frame still looks best.
Now that we've established profitability with a solid number of opportunities, we can move to sizing the trades and backtesting the results. Let's look at the Kelly fraction for each strategy combination. Note that Kelly is typically much larger than what we actually want to trade. In practice, we prefer fractional Kelly to manage drawdown convexity and parameter uncertainty. For the long at-the-money call calendar, the 30/60 has a Kelly fraction of 16.1%. The 30/90 has a Kelly fraction of 20.1%, and the 60/90 has a Kelly fraction of 18.4%. For the long double calendar, the 30/60 has a Kelly fraction of 25.7%. The 30/90 has a Kelly fraction of 31.5%, and the 60/90 has a Kelly fraction of 29.1%.
Based on the Kelly fraction and models above, it seems like the double calendar is the better, more profitable structure. Further, it looks like the 30/90 is the best time frame based on the Kelly fraction. We can now backtest these strategies to see how they would have performed historically. The backtest starts with $100,000 and allocates positions based on the highest forward factor first until the portfolio is full. Once again, these positions are held until right before the front contract expires, then closed as a spread. Make sure to check out the link below where you can download a full document with the backtest results along with the script to find these trades yourself. The portfolio construction details matter. Allocating to the highest FF concentrates capital into the most mispriced names. Closing as a spread just before expiry avoids pin risk and other expiry headaches.
Now the results. Let's start with the full Kelly results. For the long at-the-money call calendar: the 30-to-60 DTE pair achieves a 21.5% CAGR and a 1.58 Sharpe. The 30/90 DTE pair achieves a 22.6% CAGR and a 1.93 Sharpe. The 60/90 DTE pair achieves a 28% CAGR and a 1.72 Sharpe. Second, let's look at the double calendar results. The 30/60 pair achieves a 21.9% CAGR and a 2.11 Sharpe. The 30/90 pair achieves a 21.5% CAGR and a 1.97 Sharpe. And finally, the 60/90 pair achieves a 27% CAGR and 2.27 Sharpe. At full Kelly sizing, both structures perform well across maturities. The 60/90-day window clearly leads in CAGR for both, while the double calendar posts slightly higher Sharpe ratios overall.
Now let's look at the half-Kelly results. First, the long at-the-money call calendar: the 30/60 pair achieves a 20.5% CAGR and a 1.92 Sharpe. The 30/90 pair achieves a 21.9% CAGR and a 2.06 Sharpe. The 60/90 pair achieves a 27.8% CAGR and a 1.97 Sharpe. Second, looking at the long double calendar, the 30/60 achieves a 21.5% CAGR and a 2.07 Sharpe. The 30/90 achieves a 21.4% CAGR and a 2.04 Sharpe. And the 60/90 achieves a 27% CAGR and 2.38 Sharpe. At half-Kelly, performance stays remarkably close to full-Kelly levels while meaningfully improving Sharpe ratios, showing that reducing position size smooths the equity curve and boosts efficiency. The 60/90 DTE pair continues to dominate.
Finally, let's look at the quarter-Kelly results. First, the long at-the-money call calendar: the 30/60 pair achieves a 16.9% CAGR and 2.37 Sharpe. The 30/90 pair achieves a 20% CAGR and 2.64 Sharpe. The 60/90 achieves a 26.7% CAGR and 2.4 Sharpe. Second, let's look at the long double calendar: the 30/60 achieves a 20% CAGR and 2.31 Sharpe. The 30/90 achieves a 20.3% CAGR and 2.34 Sharpe. And finally, the 60/90 achieves a 26.5% CAGR and 2.42 Sharpe. At quarter-Kelly sizing, Sharpe ratios are the highest across the board. CAGR barely declines relative to full sizing, but the stability and efficiency of returns improve substantially. This highlights the practical takeaway. Fractional Kelly sizing, especially around a quarter or less, delivers the best trade-off between growth and smoothness.
The headline is straightforward. Both structures work across all three DTE pairings, and the differences in long-run performance are modest. The 60-to-90-day window consistently shows the strongest blend of CAGR and Sharpe, with the 30/60 window being the most sensitive to structure choice. That pattern reinforces the central point of this video. The edge lives in the forward factor, not in some ultra-precise choice of spread or deep pairing. The real advantage is in its ability to identify cheap, meaning cheap forward V, calendar spreads.
You'll also notice something important about sizing. Moving down from full Kelly to half Kelly, then to quarter Kelly, barely dents CAGR, but Sharpe improves meaningfully. That's because smaller bets let you run more concurrent positions, meaning better diversification. Smooths the equity curve and reduces drawdown convexity. The practical takeaway: favor fractional Kelly (quarter or less) and spread risk across names rather than leaning hard into a few.
Between the two structures, the at-the-money call calendar is the simplest to enter, manage, and scale. If you want the cleanest forward vol proxy with minimal moving parts, start here. The double calendar often posts a slightly higher risk-adjusted profile in some windows because you're selling richer front-period skew and owning relatively cheaper back-period skew, which can widen the profit tent. The trade-off is a touch more complexity, execution risks, and more costs. Bottom line: for execution simplicity and cost reduction, just pick the call calendar, since differences between the structures are marginal, and it has simpler execution and better costs.
We know this works. Let's theorize why this works so we have confidence in the edge, and then let's go through some easy-to-follow rules for actually running this strategy live. This edge shows up when the term structure imbalance becomes extreme. That's exactly what our forward factor is designed to detect. Let's break down why this works, where it comes from, and why it's tradable in practice.
Very few participants explicitly trade forward volatility. Can you think of many retail traders or outlets that are advocating trading forward volatility? Also, most of the flow in options markets is concentrated in the near-term expiries, driven by hedging, short-dated speculation, and event trading. The crowding can push near-term implied volatility too hot while leaving the forward slice, the next period implied by the term structure, underpriced. In periods of backwardation, localized panic, sector shocks, or heavy headlines, traders tend to pile into near-term protection or speculation. They bid up the front-period contracts but ignore the fair relationship to the next expiry. That leaves the back period relatively calm, and the forward implied volatility, the bridge between the two, gets compressed to levels that don't make sense once the front rolls off.
High forward factor readings often show up in mid-liquidity single names, the ones large institutional funds will tend to skip. With sensible liquidity filters and position caps, these pockets remain harvestable for the retail trader. They're big enough for retail and midsize accounts, but too small to move the needle for large volatility funds, which keeps this edge alive. When the front-period IV is bid beyond what the forward implies is reasonable, or conversely, when it's lagging too far below, time spreads naturally serve as curve balancers, helping bring the term structure back towards fair value.
It's not about one magical DTE pair. We tested 30/60, 30/90, and 60/90. And while 60/90 often looks best, the forward factor is a general detector of term structure dislocations wherever they appear. If there's a real mismatch between front implied volatility and forward implied volatility, the trade exists whether it's 7/14, 10/30, 12/21, 20/35, or anything in between and beyond. The general recommendation from experience and testing is if the forward factor is above 0.2 when measured using X-earn implied volatilities or avoiding earnings, the setup is typically tradable. Obviously, the higher the better.
Here's the key takeaway. Forward factor isn't tied to a particular calendar template. It's tied to a mispricing structure, meaning calendar versus double calendar or versus diagonal. Strikes (at-the-money versus the 35-delta) and maturities are just tools to monetize that same imbalance. Size smaller and trade more names. Size at fractional Kelly, typically quarter or less, materially improves Sharpe ratios while only modestly reducing the long-term return. A practical sizing guideline is to cap each position at around 2 to 8% of portfolio equity, with 4% being a good default. It's also good to spread risk across many uncorrelated tickers.
Here is a simple playbook to follow. One, we want to find liquid tickers with elevated near-term implied volatility and term structure in backwardation. This will lead to high forward factors. Compute the forward factor between your chosen expirations using implied volatilities or avoiding earnings altogether. If your forward factor reads at 0.2 or higher, we can go long the calendar or double calendar, where we're selling the front and buying the back. We want to size this with around 2 to 8% of portfolio capital, with 4% being a good default. Stick to quarter Kelly or less. We want to fill positions starting from the highest absolute forward factor readings downward, staying within our overall portfolio caps. Close as a spread just before the front expiry, then redeploy into fresh signals. The forward factor is simple, intuitive, and rooted in the time additivity of variance. When the short-term contract is bid far beyond what the longer-dated contract implies for the next period, you're being paid to act as the term structure balancer. If you'd like this process computed and screened for you automatically with a live screener, daily alerts, and historical forward factors, you can find it all at Oakquants.
All right, that's enough theory. Let's walk through an actual trade so you can see exactly how to find, validate, and place one of these setups today. Let's take a look at AES, which tops the list on the Oakquant screener for forward factors. Alternatively, you could find similar setups using your brokerage platform. Simply screen for tickers with very high near-term implied volatility versus much lower implied volatility in further expirations. That means we're looking for a backwardated term structure.
Now, we always want to check for earnings. In this case, earnings are estimated for October 30th. We could trade through it using X-earn volatilities, but to keep this example simple and practical, let's avoid earnings altogether. To do that, we'll choose expirations that end before earnings. We'll look to sell the October 17th, 10 DTE, 14.5 strike at-the-money call at an implied volatility of 61.97%, and we'll look to buy the October 24th, 17 DTE, 14.5 strike at-the-money call at an implied volatility of 52.11%. Plugging these values into the calculator, either the downloadable Python script or the free forward factor calculator on Oakquants, we find a forward volatility of 33.37% and a forward factor of around 86%. That's well above our 20% threshold, making it a strong trade candidate. To put on the trade, we sell that front month October 17th, 14.5 strike call for 51, and we buy the October 24th, 14.5 strike call for 61. That's a net debit of just 10. This illustrates the real advantage of using forward factor. It helps us find extremely cheap calendar spreads where we're effectively buying forward volatility at a big discount. Now, we simply hold until that October 17th expiry date and close the spread just before the front month expires. We'll have to see how this one plays out, but setups like this with high forward factor and low entry cost are the bread and butter of this strategy. If you haven't already, make sure you either download the script below or check out Oakquants.com to access the live screener, calculators, and community. Thanks for watching.