The book I was refering to is The Complete Arbitrage Deskbook by Stephanne Reverre. It's heavy on math and covers a lot of situations that are only practical for institutional trading. There was an article in an issue of stocks and commodities magazine last spring dealing with pair trading. It was an excerpt from this book. There is one chapter on statistical arbitrage, which includes pair trading. I think it's a worthwhile reference book for any spread trader. I bought it at Amazon.
spread trading
- Thread starter lescor
- Start date
The real question is, can you use options to do riskless arbitrage. Spread trading where you are hoping for a percieved relationship to come back into line is far from riskless, and many small gains can be wiped out by one bad trade, such as the long side of the spread reporting company specific bad news.
Good topic. Hope we get some ideas going.
Good topic. Hope we get some ideas going.
Ron,
I agree we are talking about different things here and I agree that atm options do contain the most amount of time value, but they also have the largest gamma. (theta pays for gamma, right?) And if I want to trade less delta than the stock, I will have to move away from deep in the money options (otherwise just trade the stock) and toward strikes near the stocks current price. And a simple way to do this is just go for the atm options (that too, is where most of the volume transacts and makes it less likely that I will be arbed).
I will, however, dust off all my old option books, and see where the sweet spot is for option purchasing, given perfect information of the impending underlying move. (I think I remember that a forward atm option purchase is the way to maximize your gains)
Also, you must have a superior model that finds, as stock777 suggested, riskless arbitrage. If you do, shhhhhh. But it seems to me you are doing covered writes with a deep in the money option instead of the stock?
And yea, pairs trading is speculative; but no risk=3% per annum returns. But actually, using options to pair trade stocks, your max risk is clearly defined up front by the amount one pays in premium.
I agree we are talking about different things here and I agree that atm options do contain the most amount of time value, but they also have the largest gamma. (theta pays for gamma, right?) And if I want to trade less delta than the stock, I will have to move away from deep in the money options (otherwise just trade the stock) and toward strikes near the stocks current price. And a simple way to do this is just go for the atm options (that too, is where most of the volume transacts and makes it less likely that I will be arbed).
I will, however, dust off all my old option books, and see where the sweet spot is for option purchasing, given perfect information of the impending underlying move. (I think I remember that a forward atm option purchase is the way to maximize your gains)
Also, you must have a superior model that finds, as stock777 suggested, riskless arbitrage. If you do, shhhhhh. But it seems to me you are doing covered writes with a deep in the money option instead of the stock?
And yea, pairs trading is speculative; but no risk=3% per annum returns. But actually, using options to pair trade stocks, your max risk is clearly defined up front by the amount one pays in premium.
Index arbitrage is basically risk free, your only risk is that settlement doesn't take place as it should, which is miniscule. Tough to do as an individual trader though. This is why I am anxious for ssf's to start trading. If you can lock in a spread there, and it's more than your cost of carry, it's a guaranteed profit. I'm sure things will be kept very tight through automated trading, but there might be some inefficiencies there.
Corey
Corey
Corey:
I'm facing the same dilemma as you are concerning risk control. With summer around the corner and the expectation of low volume choppy days, I figured this might be a viable strategy so I have been working on this strategy with one pair that appears to trade somewhat consistently with each other. My conclusions/questions are the same as yours (sorry I don't have any answers for you).
On choppy days, this strategy seems to work very well but on trending days it seems that I could lose a weeks worth of profit in one day by averaging into a spread that is going against me. I feel like I'm trading off of stochastic bands i.e.: telling the stock what it SHOULD be doing, and basically fighting the trend. What's more is that when a sector starts to trend (either up or down) I am always buying the "weaker" stock and shorting the "stronger" stock, which is completely counter-intuitive to me. If a sector starts to move up, I want to be long the strongest stock and then I look for the weakest stock so that if the sector turns around I know which one to short.
As you pointed out, the key probably lies in the subjective view of the trader (as with all strategies of trading) and a pure mechanical approach based on fixed parameters would be doomed to failure (especially for me as I do not want to have any positions overnight, and therefore must exit by the end of the day). What I have noticed though is that if I approach it as a one sided trade first, and base everything on tape reading I can be profitable by taking some quick scalps when the spread widens a little too much on one stock.
Right now, my entries are based on two different setups (still looking for more) with the condition that the spread is approaching a short-term extreme level (I do not use a "set" value to determine when that is, as I'd rather go by "feel"). The first is when one of the stocks gaps up/down (in the appropriate direction) or has an unusually wide spread that I can take advantage of. If I can get executed, I look to exit that position with a quick 10-20 cent scalp. If it looks like it's going to take a few minutes I then focus on the other stock to make sure I can still hedge myself. If the original position turns sour I first try to get out of that position. If I can't without giving up too much, I'll take the second trade on the other stock.
The second setup is when I have a large bid/offer to lean on in my favor on one of the stocks. If I'm looking to buy stock A and short stock B, and stock A has 20,000 for sale on a tight spread, I will short stock B and look to make a profit on just that leg, unless I'm forced to take the other side if others start taking the offer.
My exits are solely based on the tape right now as far as taking a loss is concerned. If I'm losing on the spread, I try to evaluate the trade in terms of the present situation. If I'm long stock A and the tape looks weak, I will exit immediately and look to re-enter at a better price. The same is true on the other side of the spread. The problem arises when the tape really doesn't tell you anything, but the two stocks keep drifting slowly in the wrong direction. In that case I will only average down in the spread once, and keep an ultimate tight stop about 5 cents away. But to be honest, it seems like whenever that scenario happens, I tend to get stopped out, so I'm really not confident that that is the best solution.
So far these two methods have been somewhat profitable for me whereas just selecting values to enter the spread has not. Unfortunately it requires much more attention and energy on my part and I wonder what the opportunity costs are as it's hard to keep track of other potential trades. I also wonder if I have put myself at a disadvantage by trying this strategy on a pure intraday basis as opposed to a longer-term basis.
Hopefully others will shed some light on this based on their own experiences.
Darren
I'm facing the same dilemma as you are concerning risk control. With summer around the corner and the expectation of low volume choppy days, I figured this might be a viable strategy so I have been working on this strategy with one pair that appears to trade somewhat consistently with each other. My conclusions/questions are the same as yours (sorry I don't have any answers for you).
On choppy days, this strategy seems to work very well but on trending days it seems that I could lose a weeks worth of profit in one day by averaging into a spread that is going against me. I feel like I'm trading off of stochastic bands i.e.: telling the stock what it SHOULD be doing, and basically fighting the trend. What's more is that when a sector starts to trend (either up or down) I am always buying the "weaker" stock and shorting the "stronger" stock, which is completely counter-intuitive to me. If a sector starts to move up, I want to be long the strongest stock and then I look for the weakest stock so that if the sector turns around I know which one to short.
As you pointed out, the key probably lies in the subjective view of the trader (as with all strategies of trading) and a pure mechanical approach based on fixed parameters would be doomed to failure (especially for me as I do not want to have any positions overnight, and therefore must exit by the end of the day). What I have noticed though is that if I approach it as a one sided trade first, and base everything on tape reading I can be profitable by taking some quick scalps when the spread widens a little too much on one stock.
Right now, my entries are based on two different setups (still looking for more) with the condition that the spread is approaching a short-term extreme level (I do not use a "set" value to determine when that is, as I'd rather go by "feel"). The first is when one of the stocks gaps up/down (in the appropriate direction) or has an unusually wide spread that I can take advantage of. If I can get executed, I look to exit that position with a quick 10-20 cent scalp. If it looks like it's going to take a few minutes I then focus on the other stock to make sure I can still hedge myself. If the original position turns sour I first try to get out of that position. If I can't without giving up too much, I'll take the second trade on the other stock.
The second setup is when I have a large bid/offer to lean on in my favor on one of the stocks. If I'm looking to buy stock A and short stock B, and stock A has 20,000 for sale on a tight spread, I will short stock B and look to make a profit on just that leg, unless I'm forced to take the other side if others start taking the offer.
My exits are solely based on the tape right now as far as taking a loss is concerned. If I'm losing on the spread, I try to evaluate the trade in terms of the present situation. If I'm long stock A and the tape looks weak, I will exit immediately and look to re-enter at a better price. The same is true on the other side of the spread. The problem arises when the tape really doesn't tell you anything, but the two stocks keep drifting slowly in the wrong direction. In that case I will only average down in the spread once, and keep an ultimate tight stop about 5 cents away. But to be honest, it seems like whenever that scenario happens, I tend to get stopped out, so I'm really not confident that that is the best solution.
So far these two methods have been somewhat profitable for me whereas just selecting values to enter the spread has not. Unfortunately it requires much more attention and energy on my part and I wonder what the opportunity costs are as it's hard to keep track of other potential trades. I also wonder if I have put myself at a disadvantage by trying this strategy on a pure intraday basis as opposed to a longer-term basis.
Hopefully others will shed some light on this based on their own experiences.
Darren
I have tried to do covered call writes using deep in the money calls instead of buying the stock, and, believe me I have never made money with that strategy.
Either the stock goes way up and I have to buy to close my short call, or the stock drops so much I lose money on the long call.
Believe me, I do not use that strategy anymore!
Either the stock goes way up and I have to buy to close my short call, or the stock drops so much I lose money on the long call.
Believe me, I do not use that strategy anymore!
Interesting discussion. I trade this method, using only stocks, almost exclusively because it fits my personality and my dogmatic insistence on risk control. Over the years, some key concepts have emerged:
1. The selection of instruments- whether done systematically or by tape-reading/feel- is the most crucial aspect for long-term success. It is best to test the pairs to include at least one raging bull and one viscious bear- some pairs do well in trendless markets, then fall apart during strong trends. I have narrowed down my trade selection process to 43 pairs, and the best single idea is set up for the next trading day.
2. Seems crucial (at least in my experience) to have a dynamic volatility component in your analysis. The May issue of TASC has a system using historical volatility. I programmed this into my Excel system to compare, and got almost exactly the same signals, though my system was a bit more responsive (also, historical volatility-10-period, 30-period, etc. has to be programmed in frequently for this system to work).
3. Risk can be further controlled by entering spreads in the direction of your signal (this is a filter I use 100%). In other words, if a spread is targeted to enter based on a 2.2 standard deviation variance from mean, I do not enter that spread until it trades at +2.2 the next morning. Exit is a +1 standard deviation further excursion (in my system).
4. Most of the spreads are closed within two days. With the pairs that I use, I found out that holding periods over two days had a higher probability of failure, or at the least, tying up capital while better opportunities were missed.
5. In cases where more than one pair are active, the pair must be in different sectors. I also like them to be in non-correlated sectors. Obviously, this is just another way to help control risk and directional bias.
6. I amplify the position with options only within two weeks of expiration and only if I can put it on at a credit. Doesn't happen often, and the only rationale (for me and the time frame I trade) is to try to capture future volatility emerging from low volatility periods.
These are just a few of the concepts I use-don't know what others do but this works for me. I really enjoy this board though rarely have time to post (and heading out the door tomorrow for two months of travel!). Kudos to the people who put this site together. Good luck in your trading.
1. The selection of instruments- whether done systematically or by tape-reading/feel- is the most crucial aspect for long-term success. It is best to test the pairs to include at least one raging bull and one viscious bear- some pairs do well in trendless markets, then fall apart during strong trends. I have narrowed down my trade selection process to 43 pairs, and the best single idea is set up for the next trading day.
2. Seems crucial (at least in my experience) to have a dynamic volatility component in your analysis. The May issue of TASC has a system using historical volatility. I programmed this into my Excel system to compare, and got almost exactly the same signals, though my system was a bit more responsive (also, historical volatility-10-period, 30-period, etc. has to be programmed in frequently for this system to work).
3. Risk can be further controlled by entering spreads in the direction of your signal (this is a filter I use 100%). In other words, if a spread is targeted to enter based on a 2.2 standard deviation variance from mean, I do not enter that spread until it trades at +2.2 the next morning. Exit is a +1 standard deviation further excursion (in my system).
4. Most of the spreads are closed within two days. With the pairs that I use, I found out that holding periods over two days had a higher probability of failure, or at the least, tying up capital while better opportunities were missed.
5. In cases where more than one pair are active, the pair must be in different sectors. I also like them to be in non-correlated sectors. Obviously, this is just another way to help control risk and directional bias.
6. I amplify the position with options only within two weeks of expiration and only if I can put it on at a credit. Doesn't happen often, and the only rationale (for me and the time frame I trade) is to try to capture future volatility emerging from low volatility periods.
These are just a few of the concepts I use-don't know what others do but this works for me. I really enjoy this board though rarely have time to post (and heading out the door tomorrow for two months of travel!). Kudos to the people who put this site together. Good luck in your trading.
Lescor,
Okay, I was wondering if there was another one. I've got the book you refer to, and I think it's a good one. However, I thought the part on stat arb, and on pairs trading in particular was relatively incomplete, compared to the rest of the book. Still, it is definately a good book for any trader's bookshelf.