YM just die?
- Thread starter steve0617
- Start date
OK, I've run a statistical analysis on optimally hedging YM exposure with an offsetting ES position, on an intraday basis.
Time frame: last 10 trading days, 2/22 - 3/7/05, RTH.
Frequency: 1) hourly bars; 2) daily bars.
Here's a snapshot (done with SnagIt) of the relevant area of the Excel spreadsheet:
The results are not that different over this particular time frame, whether using hourly or daily data. However, my recommendation for "emergency" intraday hedging (like today) would always be to use hourly results. So, that means going 0.83 ES contracts long (short) for each YM contract short (long).
Steve, if your standard YM trade size is a 15-lot, 15 x 0.83 = 12.5. So you'd want to hedge with either 12 or 13 ES contracts, your choice. That's the "pure", optimal hedge itself. Anything beyond that would be taking a separate, directional bet, per your trading plan, using ES as a proxy for the unavailable YM.
Of course, this works in reverse as well. Should the CME GLOBEX go down, um, in flames and stay down for a while, a YM position could be used -- if one wants to stick with futures -- to hedge, e.g., ES exposure, with a reverse 1 / 0.83 = 1.2 ratio (1.2 YM for each ES).
I'll gladly answer any questions about the above, if any.
Time frame: last 10 trading days, 2/22 - 3/7/05, RTH.
Frequency: 1) hourly bars; 2) daily bars.
Here's a snapshot (done with SnagIt) of the relevant area of the Excel spreadsheet:
The results are not that different over this particular time frame, whether using hourly or daily data. However, my recommendation for "emergency" intraday hedging (like today) would always be to use hourly results. So, that means going 0.83 ES contracts long (short) for each YM contract short (long).
Steve, if your standard YM trade size is a 15-lot, 15 x 0.83 = 12.5. So you'd want to hedge with either 12 or 13 ES contracts, your choice. That's the "pure", optimal hedge itself. Anything beyond that would be taking a separate, directional bet, per your trading plan, using ES as a proxy for the unavailable YM.
Of course, this works in reverse as well. Should the CME GLOBEX go down, um, in flames and stay down for a while, a YM position could be used -- if one wants to stick with futures -- to hedge, e.g., ES exposure, with a reverse 1 / 0.83 = 1.2 ratio (1.2 YM for each ES).
I'll gladly answer any questions about the above, if any.
I assume you don't mean Washington Area Girls Soccer League? 
All of these calculations about the right hedge parameters are a good first attempt but they will get you in trouble in real life.
In order to do the calc correctly, you have to know the underlying distributions of each instead of doing blind statistical analysis. Options market makers run into this all the time in order to correctly be delta neutral, and that case it is _much_ easier than this case.
nitro
In order to do the calc correctly, you have to know the underlying distributions of each instead of doing blind statistical analysis. Options market makers run into this all the time in order to correctly be delta neutral, and that case it is _much_ easier than this case.
nitro
Nitro, I agree that's an excellent and valid point in general. (Sadly, so much of applied finance is distribution-blind or, worse yet, assumes normality.)
However, I am implicitly assuming that the distribution functions of YM and ES are not dissimilar. Outside of extreme outlier events, the consensus of both traders and quants is that that's a reasonable assumption. It's worthwhile to note here that the intraday correlation between YM and ES is typically above 0.90. Over the last 2 weeks it's running at 0.96, based on hourly data.
So, in this case, at the end of the day, I am quite comfortable putting money on the line and relying on the results of a relatively simple, first order hedge ratio calculation like that above.
Appreciate your comment, as always. As it happens, over time, you're one of a short list of people on ET I've learned more than a few things from... bet you didn't know.
late_apex,
I don't think it is a terrible assumption, but as long as one is aware of these assumptions that is what matters most. If you lose and you don't know why, that is what kills you. If you lose and you have at least some idea where you went wrong, then all you got was a lesson.
Good lessons are rarely free, at least the ones handed out by the market.
Thanks for the kind words.
nitro
I don't think it is a terrible assumption, but as long as one is aware of these assumptions that is what matters most. If you lose and you don't know why, that is what kills you. If you lose and you have at least some idea where you went wrong, then all you got was a lesson.
Good lessons are rarely free, at least the ones handed out by the market.
Thanks for the kind words.
nitro
Quote from late apex:
Nitro, I agree that's an excellent and valid point in general. (Sadly, so much of applied finance is distribution-blind or, worse yet, assumes normality.)
However, I am implicitly assuming that the distribution functions of YM and ES are not dissimilar. Outside of extreme outlier events, the consensus of both traders and quants is that that's a reasonable assumption. It's worthwhile to note here that the intraday correlation between YM and ES is typically above 0.90. Over the last 2 weeks it's running at 0.96, based on hourly data.
So, in this case, at the end of the day, I am quite comfortable putting money on the line and relying on the results of a relatively simple, first order hedge ratio calculation like that above.
Appreciate your comment, as always. As it happens, over time, you're one of a short list of people on ET I've learned more than a few things from... bet you didn't know.
Bear made a flash video on what to do if the YM/CBOT goes down and your in a trade.
Here is the link:
www.puretick.com/video/hedge/
CajunSniper / Puretick Administrator-Trader
Here is the link:
www.puretick.com/video/hedge/
CajunSniper / Puretick Administrator-Trader