Hi guys,
My question concerns market cap and the risk of holding a short position in equities/ETFs in AMEX, NYSE, and NASDAQ across all sectors/industries (excluding healthcare) overnight. If you want the TL&DR, please skip to the end.
I conducted a study using historical end of day stock data from eoddata.com about the risks of shorting stocks over night. I want my trading system to avoid any overnight jumps of 2.8x, that is, where the price increases by a factor of 2.8 overnight. Because I did not have historical market cap data for many stocks, I use daily value of the product (price * trading volume), and I also look at the most recent quantity (the previous day) of price * volume. Basically, if the average amount traded (price * volume) is less than 2.5 million USD, or if the recent (1-day average) value of (price * volume) is less than 30 million USD, then it is impossible for my system to trade the stock. I increase these thresholds if the stock is in the healthcare sector due to increased probability of buyouts/announcements and faster motion. With these thresholds, I do not see any >2.8x overnight blowups in stocks from 2007 onward, regardless of how my system places it's trades (I understand there has been inflation since then, and I will adjust the thresholds accordingly as time moves forward).
Even with this study, I want to be sure that the risk of a >2.8x blowup is extremely low. Assuming that my trading signal conditions do not affect the likelihood of blowup, I am looking for a conditions that enables less than 1/10,000 probability per year of a >2.8x overnight blowup if I trade
a) ~100 positions per year
b) stocks with average holding time of 7 days
c) stocks across ETFs/stocks (including leveraged ETFs) in AMEX, NASDAQ, and NYSE across ALL sectors/industries except for those healthcare.
It is possible that my historical data set does not contain certain stocks that would not be filtered out by the above conditions, so I would like to include market cap to be safe (which I don't have access to for historical data as mentioned above). I am thinking that not trading stocks if they have market caps smaller than 100 million USD in today's dollars would be a good cutoff (for healthcare stocks, maybe 2 billion USD), but I am not sure. What do you think is a good cutoff for market cap in addition to the conditions I mentioned above. That is, if any of the 3 conditions involving market cap, average price * volume, and recent price * volume are satisfied, the stock will not be traded.
TL&DR:
I can tolerate overnight jumps of up to 2.8x in a stock while holding a short position in it (i.e. the price increases by a multiple of 2.8x overnight)- any jumps more than this are unacceptable. In today's dollars, what is a reasonable condition involving market cap and/or the product of price * trading volume (recent averages/values of these quantities) that filters out riskier stocks and prevents these overnight jumps from happening? For example, a no-trade condition for a stock could be if it satisfies any of the following conditions: current market cap less than 100 million USD, or 20-day moving average price * volume value of less than 2.5 million USD, or recent 1-day moving average value of price * volume less than 30 million. Assuming that my trading signal conditions do not affect the likelihood of blowup, I am looking for a conditions that enable less than 1/10,000 probability per year of a >2.8x overnight blowup if I trade
a) ~100 positions per year
b) stocks with average holding time of 7 days
c) stocks across ETFs/stocks (including leveraged ETFs) in AMEX, NASDAQ, and NYSE across ALL sectors/industries except for those healthcare.
Thanks in advance,
DU$T
My question concerns market cap and the risk of holding a short position in equities/ETFs in AMEX, NYSE, and NASDAQ across all sectors/industries (excluding healthcare) overnight. If you want the TL&DR, please skip to the end.
I conducted a study using historical end of day stock data from eoddata.com about the risks of shorting stocks over night. I want my trading system to avoid any overnight jumps of 2.8x, that is, where the price increases by a factor of 2.8 overnight. Because I did not have historical market cap data for many stocks, I use daily value of the product (price * trading volume), and I also look at the most recent quantity (the previous day) of price * volume. Basically, if the average amount traded (price * volume) is less than 2.5 million USD, or if the recent (1-day average) value of (price * volume) is less than 30 million USD, then it is impossible for my system to trade the stock. I increase these thresholds if the stock is in the healthcare sector due to increased probability of buyouts/announcements and faster motion. With these thresholds, I do not see any >2.8x overnight blowups in stocks from 2007 onward, regardless of how my system places it's trades (I understand there has been inflation since then, and I will adjust the thresholds accordingly as time moves forward).
Even with this study, I want to be sure that the risk of a >2.8x blowup is extremely low. Assuming that my trading signal conditions do not affect the likelihood of blowup, I am looking for a conditions that enables less than 1/10,000 probability per year of a >2.8x overnight blowup if I trade
a) ~100 positions per year
b) stocks with average holding time of 7 days
c) stocks across ETFs/stocks (including leveraged ETFs) in AMEX, NASDAQ, and NYSE across ALL sectors/industries except for those healthcare.
It is possible that my historical data set does not contain certain stocks that would not be filtered out by the above conditions, so I would like to include market cap to be safe (which I don't have access to for historical data as mentioned above). I am thinking that not trading stocks if they have market caps smaller than 100 million USD in today's dollars would be a good cutoff (for healthcare stocks, maybe 2 billion USD), but I am not sure. What do you think is a good cutoff for market cap in addition to the conditions I mentioned above. That is, if any of the 3 conditions involving market cap, average price * volume, and recent price * volume are satisfied, the stock will not be traded.
TL&DR:
I can tolerate overnight jumps of up to 2.8x in a stock while holding a short position in it (i.e. the price increases by a multiple of 2.8x overnight)- any jumps more than this are unacceptable. In today's dollars, what is a reasonable condition involving market cap and/or the product of price * trading volume (recent averages/values of these quantities) that filters out riskier stocks and prevents these overnight jumps from happening? For example, a no-trade condition for a stock could be if it satisfies any of the following conditions: current market cap less than 100 million USD, or 20-day moving average price * volume value of less than 2.5 million USD, or recent 1-day moving average value of price * volume less than 30 million. Assuming that my trading signal conditions do not affect the likelihood of blowup, I am looking for a conditions that enable less than 1/10,000 probability per year of a >2.8x overnight blowup if I trade
a) ~100 positions per year
b) stocks with average holding time of 7 days
c) stocks across ETFs/stocks (including leveraged ETFs) in AMEX, NASDAQ, and NYSE across ALL sectors/industries except for those healthcare.
Thanks in advance,
DU$T