Spydertrader's Jack Hershey Futures Trading Journal

[C99]

Futures usually lead cash as they offer quicker, leveraged exposure, but not always. So if we get a cash rally that drags the futures up you can see what you saw this afternoon. Big changes in interest rates and component stock dividends during the day can impact FV.

Like steve, I do not use this tool yet either. I do watch for one thing which Mak alluded to by calling it an arbitrage meter. The arbs don't make directional bets. They wait for the relationship to get out of line and buy one / short the other. So they throw an equal amout of money at both sides which can often be seen as a volume spike. One side has to give up ground and come back in line thus exposing themselves as the weaker side.

 

[PointOne]

Quote from callmate:

Today's chart...

Peace...

Good improvements!

2 things:

Don't let your LTL start before your RTL - it looks wrong and is wrong.

Add the major carryover channel from the previous day which influenced the entire day. (This is something I often neglect.)

 

[ericta]

Quote from makosgu:

There are a couple of internal issues with STR/SQU. This is one of the reasons why I am a staunch advocate of understanding the tools that you choose for data. If you carefully unravel what STR/SQU is, you will find that it is an arbitrage indicator. The underpinning of the futures arbitrage mechanism is the yield curve. For me, when I debriefed such areas that has been pointed out, I found that that the fair value varies slightly throughout the day. Some days virtually no variation while other days marginal to even extreme variations. On days like FOMC, the volatility spikes PRECISELY when the yield curve spikes due to FED RATE announcements which makes logical sense. The spike in the yield curve triggers the arbitrage programs to also spike as an attempt to profit from the imbalance as the yield curve moves. The math is a fair bit more complicated than PRV and requires bootstrapping of the yield curve along several cash/eurodollar/swaps datapoints. The shortcut is to simply reset your offset when necessary. My rule is to reset the STR/SQU offset whenever I see a flaw on STR/SQU. For me, a STR/SQU FLAW is when I see a STR > +3 sticking (ie. sustained for several seconds) without an ES UP TICK on a DOM CHART or a SQU < -3 sticking (ie. sustained for several seconds) without an ES DOWN TICK on a DOM CHART. I reset STR/SQU by adjusting the FAIR VALUE to NEUTRALIZE the current STR/SQU since it did not result in price CHANGING.

Regards,
MAK

Why not just smooth the base offset in 20/40 bar moving average?

something like

SS = currentBarOffset - MA(Previous 20 Bar Offset)

is above approach flawed? why is it everyone adjust it manually? am I missing anything?

BR
Eric

 

[makosgu]

Quote from ericta:

Why not just smooth the base offset in 20/40 bar moving average?

something like

SS = currentBarOffset - MA(Previous 20 Bar Offset)

is above approach flawed? why is it everyone adjust it manually? am I missing anything?

BR
Eric

Averaging a fine control tool buries the signal that you ultimately want. Consider what you are proposing carefully with respect to the information we are attempting to gather from the indicator...

MAK

 

[mephistoII]

Quote from bi9foot:

... Yes Steve, I have not relied on STR/SQU for a while because of what I perceived as it signaling the opposite direction of the next move. ..

I was discussing S/S yesterday w/ a casual observer, and he mentioned that it looked great, if you are a nineteen yr. old and can react in nanoseconds. I responded that at my age, I'm so slow, it works better as a fade signal for a quick scalp.

 

[C99]

Quote from makosgu:

Averaging a fine control tool buries the signal that you ultimately want. Consider what you are proposing carefully with respect to the information we are attempting to gather from the indicator...

MAK

He's not talking about averaging the signal but rather using an average in lieu of the offset. I don't actively use SS but have it up on my chart. Unlike most, I don't center it at 0 with a manual offset everyday. I leave the raw YM-INDU value with moving avg (BB's actually) and the MA is always within a fraction of a point of what others are using as an offset. Same exact signal +-2pts SS from the MA just not centered on zero. So to Eric I would say this approach is valid and to Mak I ask if I am missing something too?

edit- one caveat- I start the ma fresh everyday as FV changes everyday. 1 min bars,10 period, good offset value as soon as mkt syncs. And as others need to recenter through the day, this does it automatically.

 

[palinuro]

Quote from C99:

He's not talking about averaging the signal but rather using an average in lieu of the offset. I don't actively use SS but have it up on my chart. Unlike most, I don't center it at 0 with a manual offset everyday. I leave the raw YM-INDU value with moving avg (BB's actually) and the MA is always within a fraction of a point of what others are using as an offset. Same exact signal +-2pts SS from the MA just not centered on zero. So to Eric I would say this approach is valid and to Mak I ask if I am missing something too?

edit- one caveat- I start the ma fresh everyday as FV changes everyday. 1 min bars,10 period, good offset value as soon as mkt syncs. And as others need to recenter through the day, this does it automatically.

That seems a very good solution, I wish I could program it in QuoteTracker. I run an MA in the same screen and then adjust mentally or manually when it drifts too far from the zero line.

Still find Str/Squ must be interpreted carefully and in context, however.

-palinuro

 

[Pepe]

Hi,

Using a MA to automatically track the offset of STR/SQU can be done and most of the times is very helpful. But for what I see sometimes it has some disadvantages too. During periods of high volatility using a MA can cancel the signals being fired or even give wrong signals. One example is at FOMC announces, prior to that moment I change STR/SQU to a fixed value offset (this kind of cases might be solved by using a longer period MA, but I need more study on this).

Also, about the wrong signals, one needs to understand the nature and the mechanics of this tool as Mak said. Look at the formula:

“(YM – INDU) – offset”

We know that a “Stretch” is something greater than 2, and a “Squeeze” is something less than -2. But mathematically using this formula, how can a “Squeeze” (or a “Stretch”) occur?

Let’s imagine this example (forget the offset for a moment):

“Neutral”:
YM = 13900
INDU = 13900

(YM – INDU) = (13900 – 13900) = 0

“Stretch”:
YM = 13910
INDU = 13900

(YM – INDU) = (13910 – 13900) = 10

Now, what happen if the following situation occurs?

“Stretch”:
YM = 13900
INDU = 13890

(YM – INDU) = (13900 – 13890) = 10

We also have a “Stretch”, but this one is not a valid “Stretch”, because it was not done by YM moving but because INDU moved erratic. One valid “Stretch” (or “Squeeze”) signal must be fired by being YM who moves. Remember YM leads cash, so YM must move first.

This is why, I’ve programmed my STR/SQU in order to filter this situations by considering only moves initiated by YM. It gives less erratic signals.

Best Regards,

 

[PointOne]

Quote from Pepe:

Now, what happen if the following situation occurs?

“Stretch”:
YM = 13900
INDU = 13890

(YM – INDU) = (13900 – 13890) = 10

We also have a “Stretch”, but this one is not a valid “Stretch”, because it was not done by YM moving but because INDU moved erratic. One valid “Stretch” (or “Squeeze”) signal must be fired by being YM who moves. Remember YM leads cash, so YM must move first.

This is why, I’ve programmed my STR/SQU in order to filter this situations by considering only moves initiated by YM. It gives less erratic signals.

Best Regards,

I don't see how you can differentiate between them: YM NOT moving when INDU does is as strong a signal as if it moves before INDU. It's a vector: magnitude and relative direction.

 

[Pepe]

Quote from PointOne:

I don't see how you can differentiate between them: YM NOT moving when INDU does is as strong a signal as if it moves before INDU. It's a vector: magnitude and relative direction.

Hi PointOne,

For me, because I believe YM leads cash, any signal that is not originated from a move of YM I mark it as an invalid signal. So, if INDU goes down 5 pnts with YM not moving, I do not consider this a valid "Stretch".

If you monitor real time, you see sometimes INDU making some erratic moves with YM at the same value. Of course, it's not values of +/-10, but +/-2, +/-3 or more happens. Also, most of the times it is INDU that returns to the original value rather than YM that goes along to annulate the offset. If you think this in terms of the STR/SQU indicator, you will see this as a "Stretch" (or a "Squeeze") and then "Neutral". This can be interpreted as a normal valid "Stretch" signal, in reality, it was nothing.

Best Regards,