Spydertrader's Jack Hershey Futures Trading Journal

[ivob]

Quote from ivob:

Sometimes it invalidates what you're thinking and sometimes it confirms.

Quote from optioncoach:

That makes it no more useful than a guess I would think.

I disagree. For example I saw a moment at the end of the day when price was at RTL. Volume very low and str/sq gave a significant signal. Especially when YM and ES don't move much like at RTL breakthru (an important moment) it seems useful.

For example if you think you have a point 3 in an uptrend but str/sq < -2 then you don't take the trade. If str/sq is neutral or > +2 you do take it. As observed so far.

I have observed several moments today when there was a sudden movement of price and str/sq did the opposite and price did not continue. If str/sq is neutral I would not pay much attention.

Of course this is from just a few days observing.

Very smart to add instruments little by little in this journal. Right now we use what we already know and we don't look too much at str/sq. This is exactly what it is for, extra information. Else we would have looked at everything at the same time and that's confusing.

regards,
Ivo

 

[Spydertrader]

04-04-2007 YM Chart

- Spydertrader

 

[makosgu]

Quote from optioncoach:

Y eah I am going to add it at the bottom and line it up with key moves and see what patterns emerge. Studying it myself is the only way to see if it is useful for me or just added noise.

Agreed. I am the same way. This is how I arrived at the +/- 2 values (ie. after figuring out what must be what and then visually calibrating)...

 

[makosgu]

Quote from dkm:

I previously had an ohlc equivalent which did show the extremes then I changed it to histogram. I am nowhere near ready to use this tool so I am putting it on the back burner. My sim results continue to be very disapppointing.

Just out of curiosity, on average, how long would you say that you typically hold a trade???

 

[ivob]

Spyder, Mak. A question about str/sq.

Do you relate bars to previous bars? I mean , is there something like "Mmm. right now str/sq is +3 and also last bar and the bar before that it was > + 2 and it doesn't go down much".

Can conclusions be drawn from str/sq values over several bars or should we just look at the "now"? So is there something like a trend in str/sq?

Also I notice when price volatility increases str/sq bar length also seems to increase.

regards.
Ivo

 

[C99]

I have a few initial questions on str/sq too if you guys don't mind. These are pointed a little more at understanding what we're looking at rather than just the raw application of it.

First, are we working on the assumption that futs always lead cash, or is there room in the analysis for the thought that cash sometimes leads? For example around 3:02-3:05 EST today the YM pushed down a few points and I had squeeze values of -2 to -3 using an offset of 66. It was like the futs tried to push cash down, couldn't and then both moved up about 15 points with very little str. In this case I would say cash was more in control.

Second question is when we see values worth noting, does it matter which one, cash or futs, brings us back to neutral? For example, in an extended move up we may see futs str, then cash follow and bring us back to neutral, str again and cash moves up again to bring us back to neutral. Close to the end of a move I've noticed futs str, but cash doesn't follow and we return to neutral with futs backing off a bit. Is this an important sequence or am I making things up in my head?

Along the same line as previous question, Often we'll get a str / squ value big enough to spark an index arb. program. Considering an arb program cares not about direction and throws an equal amount of money at both sides I've previously looked at these times as moments where the market shows it's hand. One side of the divergence, cash or futs, will have more conviction than the other and hold its ground. So in the context of this journal is it correct to view the side that stays put as the right side and the side that gives up ground to return str/sq to neutral the weaker side?

 

[dougcs]

I like others am having trouble calibrating for the value to use to "zero" the premium.

THis is a Tradestation indicator that uses a 150 bar average value. Don't know if it is any better than eyeballing the value, but I'm watching it as well as doing the intraday recalibrations.


[LegacyColorValue = true];

input: YM(c data1), INDU(c data2);
var: intrabarpersist bn(0), intrabarpersist diff(0), intrabarpersist hi(0), intrabarpersist lo(0), ma20(0), sdev(0);
var: d2(0), UBB(0), LBB(0), str1("L");

diff = ym - indu ;
ma20 = average(diff, 150);
{sdev = stddev(diff, 150);
ubb = sdev;
lbb = -sdev;
}
ubb = 2;
lbb = -2;

d2 = diff - ma20;

if time<>time[1] and bn <> barnumber then begin
bn = barnumber; hi =-9990; lo = 9999;
end;
if bn=barnumber then begin
hi = maxlist(hi, d2); lo = minlist(lo, d2);
end;

if d2 > 0 then begin
setplotcolor(1, darkgreen);
end;
if d2 < 0 then begin
setplotcolor(1, darkred);
end;

setplotwidth(1,1);

if time>=0830 then begin
if d2>= ubb then begin
setplotwidth(1,power(d2,0.6)); setplotcolor(1,green);
alert(symbol+" +2 Str / Squ");
end;
if d2 <= lbb then begin
setplotwidth(1,power(absvalue(d2),0.6));
setplotcolor(1,red);
alert(symbol + " -2Str / Squ");
end;
end;
plot1(d2,"Dif");
plot2(0);
plot3(ubb,"UBB");plot4(lbb, "LBB");
plot5(hi,"Hi");plot6(lo,"LO");
{
if lastbaronchart then begin
str1 = "avgPrem= "+NumToStr(ma20, 2);
value11 = Text_New(Date, time-5, h+1, str1);
value1 = Text_SetColor(value11, cyan);
end;
}

If you know how to read this code, you can see that I'm playing around with using standard deviations instead of the +/- 2 value.
Too soon to see if it helps.

Doug

 

[Lightbody]

Quote from Bearbelly:

No way to do it that I know of. I have been trying to get someone to code an excel app to do it but so far no takers. Maks excel works but I would rather have the histogram.

Thanks

 

[Spydertrader]

Quote from ivob:

Do you relate bars to previous bars?

Other than to insure I do not need to 'reset' the 'offset' value, no.

Quote from ivob:

Can conclusions be drawn from str/sq values over several bars or should we just look at the "now"? So is there something like a trend in str/sq?

I look at STR / SQU only in the now.

Quote from ivob:

Also I notice when price volatility increases str/sq bar length also seems to increase.

A very nice thing to notice.

Quote from C99:

First, are we working on the assumption that futs always lead cash, or is there room in the analysis for the thought that cash sometimes leads?

The futs leads the cash until you see an inversion. On February 27, 2007, we witnessed the first inversion in a very long time.

Quote from C99:

Second question is when we see values worth noting, does it matter which one, cash or futs, brings us back to neutral?

The value of STR/SQU is what matters, as well as, at what point in time the market generates the value.

Quote from C99:

Is this an important sequence or am I making things up in my head?

Neither, determining who pushed or pulled whom doesn't assist the trader in generating profits.

Quote from C99:

Along the same line as previous question, Often we'll get a str / squ value big enough to spark an index arb. program. Considering an arb program cares not about direction and throws an equal amount of money at both sides I've previously looked at these times as moments where the market shows it's hand. One side of the divergence, cash or futs, will have more conviction than the other and hold its ground. So in the context of this journal is it correct to view the side that stays put as the right side and the side that gives up ground to return str/sq to neutral the weaker side?

Another tool (not yet discussed) provides the signals for "which side is weaker" with respect to the market and not STR / SQU.

All signals generated with respect to how one 'sees' the market exist within a binary paradigm. In other words, we see them, or we do not see them. Once a trader views the market generating a signal, one must move immediately to action. Often, other signals create the sufficiency needed to act, and we never arrive at a STR /SQU observation. On other occasions, STR / SQU (and the tools used previously) fails to provide any signal (for either continuation or change), and as a result, we need to look to additional tools for our sufficient data sets. Currently, we do not yet have the 'Fine Resolution' tools available.

For now, monitor STR / SQU in an effort to learn the behavior at your individual Resolution Level Monitoring (Forest or Tree). Note the changes in price which occur at those 'action points' as well. Remember, STR / SQU does not represent the 'silver bullet' of indicators, or the 'end all be all' of signals. In addition, STR / SQU does not represent a 'confirming' indicator. STR / SQU simply brings the trader a step closer (with 3 steps left to go) to the point at which change occurs in the market.

Think back to when we only had ES Price Channels and Volume Gaussians to determine a signal for change or continuation. At that time, many individuals posted how they often found it difficult to 'see' the FTT until a bar or two later. Adding the YM as a leading indicator of a change signal on the ES decreased the time a trader needed to 'see' the signal for change. As a result, many began to 'see' change occur closer to the actual bar the FTT formed. Now, with the addition of STR /SQU, we start to have the ability to capture Intra-bar Gaussian shifts. Although STR / SQU permits the trader an opportunity to see change signals develop within the bar itself, it does not provide pinpoint accuracy.

As we add additional tools (later this year), we move closer to 'seeing' the change signal develop even earlier than we do with STR / SQU.

Hopefully, the above information provides some additional clarity.

- Spydertrader

 

[The Swordsman]

Spydertrader,

If you dont mind, i have a quick question for you concerning bars that take out both a new high and new low to the previous bar. Are there any rules associated with them? Do they signal anything with any consistency?

4/4 ES bar at 15:30 is an example. Thanks!