- Floor traders are who execute transactions from the floor of an exchange and they trade exclusively for their own accounts.
- This report analyzed weekly call options contracts purchased by floor traders in August 2021. We found that roughly ⅓ of the weekly calls purchased were out of money calls (OTM) with a bias in underlying stock sectors.
- By monitoring the day to day option price fluctuation, more than 50% of these OTM calls would yield a positive return under favorable conditions.
- Lastly, another inferred bias in option pricing from these floor trades might be useful for your learning option contract selection.
- This report is the first of many to come that focused on floor trades data collected in Unusual Whales.
Traders who execute transactions from the floor of an exchange (in-person) and exclusively for their own accounts, are known as floor traders. Typically in movies, they are the ones shouting with enthusiastic hand gestures from the floor of the stock exchange. Floor trading has become increasingly rare as the majority of trading has moved online. In fact, Chicago Mercantile Exchange (CME) announced the closure of its physical trading pit in May 2021. However, this doesn’t mean floor traders and their accounts will cease to exist overnight.
Using Unusual Whales Flow, we are able to monitor trade sentiments flagged with the “floor” tag. Given the fact that these professional traders are operating exclusively from their own accounts, we want to investigate how profitable their trades were and to discover some useful patterns from floor trades that retail investors can learn from.
About the Data
- Data for this analysis was confined to option contracts purchased in August 2021.
- Price was the nbbo recorded on the day. And informed returns which are expressed as percentages.
- Buy or sell estimations on contracts were made by comparing the contact’s execution price to the bid and ask price recorded in nbbo.
This report examines weekly call option contracts purchased with underlying stocks (excluding ETFs) with market capitalizations less than $150 billion dollars. Floor trade data in 2021. To ensure market capitalization did not introduce selection bias in terms of contract expiration, Table 1 summarized the days till expiry (DTE) by market capitalization. The DTE distribution was comparable for both groups, with the large cap group skewed a bit towards shorter DTE.
Table 1: Floor trades days till expiry (DTE) summary by market capitalization
In the August floor trades data, nearly 300 weekly calls bought are OTM (out of the money) and 600 were ITM (in the money). Let’s examine the OTM calls.
More weekly OTM calls placed in technology and consumer cyclicals
Figure 1 shows the distribution of floor trades in each sector as percentages of counts per contract DTE type. We can see some preference towards floor trades in Technology (Blue) and Consumer Cyclical (Orange) sectors. One could argue the observed bias was due to the availability of weekly contracts in these two particular sectors. For example, $TSLA (Consumer Cyclical) would have weekly options to choose from compared to $AXTA (Basic Materials).
Figure 1: Normalized weekly OTM calls sector distributions by contract DTE
Figure 2 expanded the analysis to calls with DTE between 14 to 60 days. Yet again we saw that contracts in Technology (Blue) and to a lesser extent Consumer Cyclical (Orange) were preferred up to 40 DTE.
Figure 2: Normalized OTM (14 to 60 days) calls sector distributions by contract DTE
Figure 3 shows how the floor traded weekly ITM calls distributed by sectors. In contrast to weekly OTM calls, these weekly ITM call purchases leaned towards Consumer Defensive and Financial Services sectors, suggesting the skew on Technology (Blue) and Consumer Cyclical (Orange) was associated with the strike price rather than time or option availability.
Figure 3: Normalized weekly ITM calls sector distributions by contract DTE
Performance of weekly OTM calls
There’s the perception that floor traders have some sort of advantage in identifying stocks with explosive growth potential. One thing we thought to look at was the strikes of the weekly OTM calls made by floor traders. Were they betting far out of the money?
We looked at the option “moneyness” when purchased by the floor traders: how far out of the money (or in the money) the strike price is as a percent of the underlying price. An October $90 call with underlying price at $80 would have an OTM moneyness ~11% since($90-$80)/$90. Table 2 summarized the moneyness of weekly OTM calls purchased by floor traders. Although very far OTM calls did exist, the mean value suggested the majority of weekly calls were centered around 12% OTM.
Table 2: Floor trades OTM weekly calls moneyness summary
All the weekly calls were grouped into 4 categories based on Table 2: 5% (or less) OTM, 5-10% OTM, 10-20% OTM, and 20% (and beyond) OTM calls. Figure 4 illustrates the percentage of total counts for each category. We saw that 20% and beyond “wild calls” were in the minority (less than 20%), while the majority of the calls were placed < 10% OTM with the 5% OTM dominating.
Figure 4: Distribution of weekly OTM call moneyness
OTM calls are constituted purely of extrinsic values (in contrast to the ITM calls that contain intrinsic value) and are therefore cheaper compared to ITM calls. From this perspective, OTM calls can give traders serious leverage if the underlying stock indeed moves in favor. The further out of the money, the cheaper the option contract costs and the more one stands to gain from this feature. On the other hand, the further out of the money, the less likely the option will expire in the money as reflected by options delta.
Perhaps the observed dominance of close OTM calls was biased by the nature of the DTE range: floor traders were buying contracts with DTE less than 14 days, so it makes sense for them to focus on 5% OTM calls that have higher delta, thus higher chance to expire in the money and profit.
How about floor traded OTM calls with DTE ranging from 14 to 180 days? We explore these in Figure 5. A similar pattern was observed: the majority of the calls were placed < 10% OTM with the most at 5% OTM. It appears the professionals didn’t opt for crazy far out of the money bets. Assuming floor traders have a directional bias on these weekly OTM calls, a bigger delta means a bigger rise in the option price per 1 point of move of underlying stock towards the strike. Wait, there is more! Option delta is not stationary and option pricing itself is a convex curve in response to the underlying price movement. The option price doesn't just increase by the fixed amount of delta. Gamma, which can be thought of as the “optionality” we are buying, dictates how sensitive that option delta can change in response to the change of the underlying stock. The option gamma rises to its max when the underlying stock price moves to the strike price (becoming ATM). In this sense, a close OTM call will have a bigger delta and gamma compared to the far OTM ones, therefore a bigger price jump per 1 point of move of underlying stock.
Figure 5: Distribution of weekly OTM call moneyness(14 to 180 days)
One challenge with analyzing floor trades is the lack of information on trade exits. For example, after the identification of a call option purchased by a floor trader, there is no concrete information on when the position has closed. However, we can infer trade entry and exit by tracking the change in options volume and open interest.
To estimate all possible scenarios, a “moving daily return” was calculated for each trade. Simply, the concept was to monitor the option price each day until it reached expiration since the contract purchase date, thus allowing all possible exits to happen. And for each day, we calculated the daily return by giving the floor trader the daily high price. By constructing this “all is possible'' model, Figure 6 illustrates the returns of different OTM weekly calls purchased by floor traders. The red dot shows where the group mean was and a t-test compares each group’s mean return to that of the 5% OTM group.
Figure 6: Weekly OTM calls and their moving daily returns
Not surprisingly, 5% OTM calls had a positive mean, and all other groups had a significantly lower mean (also in negative) in comparison, explaining why the majority of the bets made by floor traders were at this 5% OTM group. This return pattern was also observed even when looking at longer expirations (from 14 to 180 days) in Figure 7 where the 5% OTM group outperformed further OTM groups. This is consistent with their larger delta, a higher probability to expire in the money.
Figure 7: OTM calls (from 14 to 180 days) and their moving daily returns
Buying far OTM calls could yield higher returns on the off chance you’re right. However, large portfolio managers may opt for less risky moves and thus prefer these closer OTM calls(assuming when they have a directional bias). From the selling side of these OTM calls, usually market makers, the option pricing convexity (gamma) limits their profitability. Since the premium collected equals their maximum profit, they make money through time decay but will lose money if the underlying moves unfavourably due to the negative gamma and therefore the dynamic hedging involved.
To get an idea of how often these weekly calls win, we turned to the moving daily return distributions as the area under the curve would indicate a rough estimate on the probability of winning (Figure 8). The vertical black line indicates 50% returns. An additional filter added was the DTE (shown by color) so we could quickly examine if a particular DTE had better performance over others simply by checking the presence of “fat tails” that lay to the right.
Figure 8: Distribution of the moving daily returns for weekly OTM calls by DTEs
Knowing the 5% OTM calls were the winners, we could see most of the DTE had most of their returns distributed between 0 and 50%. Although some DTE might look better than others, we could not identify a specific DTE that consistently outperformed the rest across different OTM moneyness groups.
Interestingly, out of the 280 OTM weekly calls purchased by floor traders, 85% of them had max returns above zero, and 58.21% of them had their max return above 50%!
Figure 9: Open trades by floor traders (weekly OTM calls) and the maximum possible return
To get an even clearer picture, floor bought OTM weekly calls were further filtered to identify open trades which must have been opening positions only by gauging the size, open interest and volume. Out of the 280 OTM weekly calls bought, 83 of them were definitely bought to open and their max returns were shown in Figure 9. In short, more than 52% of them had positive returns and roughly 25.6% had max returns over 50%.
Examine the weekly ITM calls calls purchased by floor traders and their performances
Now that we have seen the OTM weekly calls purchased by floor traders, time to check those ITM weekly calls. Since we already had the sector preference of ITM calls in Figure 3, we will dive into the ITM moneyness in Table 3 below.
Table 3: Floor trades ITM weekly calls moneyness summary
From Figure 10 we could see better that the ITM weekly (all defined by DTE less than 14 days for this section) calls purchased by floor traders were “deeper” in the money, judging by moneyness, compared to the OTM calls purchased. Most of the ITM calls were clustered in the 5 to 20% segments. And again, we could see the reason why the professionals tend to do so in Figure 11, where 5 to 20% ITM groups beat the 5% ITM calls.
Figure 10: Distribution of weekly ITM call moneyness
The initial observation makes sense since the deeper in the money, the safer a bet should be. However, it’s interesting to note that the mean values (red dot) of all moving daily returns were negative. This implied that ITM calls were outperformed by the OTM calls. How could this happen? It might have to do with the reasons for purchasing ITM calls, a deeper dive into ITM calls can be found in our other post.
Figure 11: Weekly ITM calls and their moving daily returns
Figure 12: Distribution of the moving daily returns for weekly ITM calls by DTE
By checking the distribution of moving daily returns of those weekly ITM calls in Figure 12, we saw most of the ITM calls did not have any “long tails” (i.e., barely had any area to the right of the vertical 50% line). The distribution of these “all possible” returns were somewhat symmetrical to the left of zero. That is, if you are after “big gains”, ITM calls will not help you much.
Figure 12 also suggested that the floor traders were probably betting heavily, in volume, on the ITM calls to make consistent small gains (the area under the curve between 0 and 50%). By filtering the max daily returns in these 600 ITM weekly calls, ~72% had a positive max return (compared to the ~85% observed for OTM weekly calls), but only 10% ITM weekly had over 50% max returns (compared to 58.21% in OTM).
We should keep in mind that “purchased” does not always mean opening a position. To get a cleaner comparison, an open trade filter was applied, similar to the same way as done in Figure 9, to view the maximum returns for open trades of the weekly ITM calls in Figure 13. Though the pattern remained very similar to what we saw in Figure 11, the mean value did improve. Out of the 295 weekly ITM open trades, 40% of them had positive max returns and 5.7% were over 50%. From this we can conclude that floor traders did better in their weekly OTM calls than their ITM calls, at least for August.
Figure 13: Open trades and their maximum possible return
Inferred bias in weekly OTM call picking by price
So far this report has shown that weekly OTM calls purchased by floor traders do well. Next, we’ll try to figure out criteria that help floor traders place these bets. Since we saw they like to pick calls just ~5% OTM, one thing that might be interesting to look at is the option chain prices.
We know that option chains are priced differently even when the underlying securities are priced similarly. To illustrate the idea, Table 4 displayed parts of $IRM and $XM option chains. Both stocks were trading similarly when analysed but their option prices (same expiry dates), especially for OTM calls, were very different. Option pricing is largely estimated by theoretical pricing models (such as the Black-Scholes Model or the Binomial Model). Essentially, these models assign probabilities to a series of future expiry prices. With all else being equal (ex. interest rate, time to expiration and strike prices), the difference between the observed option prices $IRM and $XM is a reflection of their volatility. Generally speaking, higher volatility contributes to a higher chance the option will expire in the money, thus more expensive option contracts. The so-called “buy low and sell high” in traders eyes is really “buy underpriced volatility and sell overpriced volatility”.
Unfortunately, no one has a crystal ball to predict future volatility of the underlying. With that being said, could option chain pricing act as good criteria to buy cheap options? When we express the ask price of call options as percentages of the underlying price (Table 4. Column 3) and check the strike moneyness (Column 5), we observed ~10% OTM, $IRM OTM call was 0.3% of its underlying but $XM was priced at ~7% of its underlying. Due to the convexity of option pricing, if both of them moved to at-the-money (assuming nothing else changes), the price jump between $IRM (0.3% to 3.6%) and $XM (7% ~14%) would be hugely different.
Table 4: OTM chain pricing difference
This feature of price distribution (Table 4, Column 3) could be modelled by borrowing the concept of kurtosis, which describes the “tailedness” of a distribution. In Figure 14 the simplified illustration shows that the blue distribution (leptokurtic, “skinny”) has longer tails than the red ones. The resulting effect is that for the x amount of change to the right, the change in value d1 is much bigger than d2 (and vice versa).
Figure 14: Oversimplified illustration of kurtosis and its implications
Now if we imagine from the deepest ITM call to the furthest OTM call (Table 4 column 3), the option chain prices as percentages of underlying for each ticker could be used to model their distribution would be something similar to the “half distribution” in Figure 14. The main idea here is to figure out whether floor traders were buying cheap OTM calls like IRM over $XM, where the price of the option drops faster as the chain moves out of the money. Option chains with a distribution like $IRM, with a steep price drop in their OTM arm, could be measured with a positive kurtosis value bigger than 3 (so-called “leptokurtic”).
For all 280 OTM weekly calls purchased by floor traders in August, we went back to fetch the options chain data of each trade on the execution date and examined their options chain price distribution (for those who had more than 5 chains, 210 passed this filter). We found on average, the kurtosis value was ~4.86, meaning their OTM weekly calls picked were highly leptokurtic. This means when the underlying moves in their favour and the call option becomes in the money, the price jump will resemble the magnitude of “d1”.
Table 5: Floor trades OTM weekly option chain price distribution kurtosis value
This report is the first of many to focus on tracking floor trades. Our goal was to detect interesting patterns or strategies employed by professional options traders and to share those with retail investors. . We found that floor traders had preferred sectors for OTM and ITM options. More interestingly, we found that floor traders do not bet on crazy far OTM calls. Instead, most of the purchased weekly calls are ~10% ITM, with about a 40% chance to turn a profit. And when it comes to OTM calls, nearly half of these trades end up making gains and roughly a third of them can hit 50% and above. We cannot know exactly the criteria they used to place each bet, but using the Unusual Whales Flow tool to monitor floor trades, especially those open trade OTM calls, could be a way to help retail players make better decisions.