YM just die?

[steve0617]

Quote from late apex:


So, if anything, you were overhedged by about 3 ES contracts long (15 x 0.8 = 12). Unfortunately, the fact that your long ES entry was late cannot be considered. If you go long more ES than called for by the correct OHR, you are now adding an outright long bet on ES to the long/short ES-YM spread. That's fine, as long as you're aware of the additional directional exposure risk you're taking on and accept it.

Exactly. But I was willing to be off a bit (at least today) in order to minimize the losses I was anticipating.

This was the calc that I knew existed but did not know how to create so instead of determining how many ES at the same level can be expected to offset the YM , I just threw 15 at it to try to not get my head crushed.

Thanks for the math.

 

[steve0617]

Quote from late apex:

Before I forget: let's not lose sight of DIA as a natural "perfect" hedge for YM in case of only that particular futures exchange being down, like today. Assuming that one is set up to trade stocks as well as futures. 500 sh. for each future... no regression needed. Similarly for the other e-minis and their respective ETFs.

Yes, but margin requirements have to be considered too. Not that I'm trading so big that I couldn't do it this way, but I wanted to be within the futures market rather than dropping to the ETFs.

Plus, who wants to screw with wash sales?

 

[blueskier]

Quote from late apex:

I think some will be confused by the above discussion of ES / YM hedging ratio... let me see if I can clarify.

The tick sizes are irrelevant. To compute the optimal hedge ratio (OHR) -- how many ES to go long for each YM short? -- one needs simply to consider the relative USD value of a 1% move in ES and YM, adjusted for beta.

Beta here is the coef. (slope) of a linear regression of % change in YM (dependent variable) on % change in ES (independent var.), with the intercept forced to 0. (Such a regression works well because YM and ES are highly positively correlated.) The time frame of the inputs should be similar to the expected hedge duration. So, if counteracting an exchange blackout, perhaps hourly is apppropriate; no hard rule here.

I'll let you run your own regression on your data. For this example, let's use 0.9. In this case, using today's values at 9:20 am CST:

OHR = 0.9 x (10,940 x 1% x $5) / (1,226 x 1% x $50) = 0.9 x 0.892 = 0.803 ~= 0.8 ES contracts for each YM contract.

In plain English, ES:

1) moves more in $ value than YM, for a 1% move in each, about $613 v. $547; and

2) is more volatile than YM, on average.

So, if anything, you were overhedged by about 3 ES contracts long (15 x 0.8 = 12). Unfortunately, the fact that your long ES entry was late cannot be considered. If you go long more ES than called for by the correct OHR, you are now adding an outright long bet on ES to the long/short ES-YM spread. That's fine, as long as you're aware of the additional directional exposure risk you're taking on and accept it.

No, what you are interested in for an intraday hedge is relative intraday movement. Based on last few weeks, ES ADR is about 9.4, and YM is about 90, meaning point-for-point intraday ratio is about 9.6 YM/ES. So, to hedge 15 YM you'll need:

1 ES pt x $50/pt x n ES = 9.6 YM pt x $5/pt x 15 YM
n = (9.6 x 5 x 15)/50 = 14.4 ES

But, if using ES to hedge, I wouldn't even screw around with that, and would use 15 since there are market structure reasons for expecting YM and ES to line up at 2.5 YM pts per ES tick.

Also, the statement that "ES is more volatile than YM" is irrelevant and misleading, and just an artifact of involving the absolute value of prices in your calculations (i.e., the "10940" and "1226"). It doesn't matter, for example, whether ES is at 1,000,000,000 and YM at 1000 - only what there relative intraday movement is.

 

[late apex]

Blueskier, appreciate your thoughts on proper intermarket hedging.

Quote from blueskier:

No, what you are interested in for an intraday hedge is relative intraday movement.

Agree, as long as you mean "expressed in USD terms" (or whatever the trader's base currency is).

Based on last few weeks, ES ADR is about 9.4, and YM is about 90, meaning point-for-point intraday ratio is about 9.6 YM/ES. So, to hedge 15 YM you'll need:

1 ES pt x $50/pt x n ES = 9.6 YM pt x $5/pt x 15 YM
n = (9.6 x 5 x 15)/50 = 14.4 ES

My value of 0.9 for beta was given as an example only (pls. see my 1st post above). I have not had a need to carry out such a linear regression in a while. Could 0.96 be plausible? Sure, depending on the time frame and data frequency you choose to play with.

But, if using ES to hedge, I wouldn't even screw around with that, and would use 15 since there are market structure reasons for expecting YM and ES to line up at 2.5 YM pts per ES tick.

Also, the statement that "ES is more volatile than YM" is irrelevant and misleading, and just an artifact of involving the absolute value of prices in your calculations (i.e., the "10940" and "1226"). It doesn't matter, for example, whether ES is at 1,000,000,000 and YM at 1000 - only what there relative intraday movement is.

Well, I still fully stand by that statement, "ES is more volatile than YM [intraday]." Actually, it has nothing to do with absolute price levels. Put another way, "For a 1% intraday price change in YM, one would expect, on average, a greater than 1% intraday price change in ES, over the same time period". That's something any trader can readily verify empirically, as I think you would agree. That's what I meant by that statement, no more and no less.

Not sure where you see the involvement of the absolute value of prices, as it indeed has nothing to do with that empirical observation. Perhaps you thought I meant something else?

 

[steve0617]

I am not a math guy, at least, not a stats guy.

Is there a figure I could reasonably use for myself that says, on average, if I'm trading X # of YM contracts, that if I need to hedge and assuming the ES entry price would be placed at the same price (e.g. YM on the 9:04am bar was X. The same ES price if traded on that bar was Y) that the YM entry was made, that Y # of ES contracts would approximately equal the same outcome?

e.g. like today, I was short 15 YM contracts and went long 15 ES in an attempt to alleviate as much of the YM loss as I could.

If a 1-1 ratio isn't correct *most* of the time, what would be the correct number of ES to trade when hedging YM?

I fully realize this is a floating figure that adjusts a bit all the time but *on average* what should my quantity of ES be?

 

[late apex]

Steve,

Let me take a look at the most recent data and post my results. Of course, anyone else with experience in this area is more than welcome to share.

 

[Pabst]

Wags had it right. YM typically has about 90% of the daily "dollar" range as ES.

Unfortunately if Steve had kept this ES-YM spread on a bit longer he would have done ok as this was a semi-rare day where the Dow was weak to the other indicies.

 

[G'N'R]

"semi-rare" is that like semi-pregnant?

is that like half a rare day?

or maybe half a non-rare day?

what does this mean?

either it is rare or it is not.

:^|

 

[newbunch]

Quote from G'N'R:

"semi-rare" is that like semi-pregnant?

is that like half a rare day?

or maybe half a non-rare day?

what does this mean?

either it is rare or it is not.

:^|

More like medium rare, I think.

 

[Pabst]

Quote from G'N'R:

"semi-rare" is that like semi-pregnant?

is that like half a rare day?

or maybe half a non-rare day?

what does this mean?

either it is rare or it is not.

:^|

How about "recently-rare."

i.e. The 500's are a bit more tech heavy than the Dow. Therefore it takes a relatively bullish Naz performance to make ES strong to YM. We all know how "rare" tech led strength has been this year.