RT Model Long Again

[Billy]

The PM MF hit the short protestion level at 9:38 am and reversed to a long position.
The new long position is protected at the-0.145% level.
Billy

[TraderD]

The PM MF hit the short protestion level at 9:38 am and reversed to a long position.
The new long position is protected at the-0.145% level.
Billy

It may not be the right way to interpret intraday MF, but it really puzzles me how a higher high in MF today coincides with an intraday double top ($47.23 & $47.25) proceeds to plunge for the rest of the day. Makes me wonder whether MF is a leading vs. lagging indicator.

Trader D

[Pascal]

It may not be the right way to interpret intraday MF, but it really puzzles me how a higher high in MF today coincides with an intraday double top ($47.23 & $47.25) proceeds to plunge for the rest of the day. Makes me wonder whether MF is a leading vs. lagging indicator.

Trader D

This ping-pong game is typical of options expirations.

I do not use the MF for day trading, because I only backtested it for position trading. Therefore there is no base to conclude that for day trading this is a lagging or a leading indicator, especially during expirations!


Pascal

[Billy]

It may not be the right way to interpret intraday MF, but it really puzzles me how a higher high in MF today coincides with an intraday double top ($47.23 & $47.25) proceeds to plunge for the rest of the day. Makes me wonder whether MF is a leading vs. lagging indicator.

Trader D

Doron,
Let’s try to interpret the models within their technical context.

GDX is trying to break the 1 ½ month downtrend channel to the upside which would be a major trend reversal. Expect fast and big up moves if and when it breaks that channel.

13884

Bearish large players are targeting a retest of the lower downtrend channel line far away in the abyss, and they are launching their sell/short programs when GDX approaches the upper downtrend line, which hurts the RT model into choppiness.

Bullish large players are accumulating for the long term here in anticipation of a reversal.
They do it when price falls back down to the support line that began early April, which helps the EOD model into staying long.

I think your portfolio is in better shape in following the EOD model when a trend is clearly trying to reverse. The RT model looks more appropriate in the middle of a trend.
Both models have edges but under different market conditions. Once again, more data and observations are needed before drawing reliable conclusions.
Billy

[davidallison@gmail.com]

Doron,
I think your portfolio is in better shape in following the EOD model when a trend is clearly trying to reverse. The RT model looks more appropriate in the middle of a trend.
Both models have edges but under different market conditions. Once again, more data and observations are needed before drawing reliable conclusions.
Billy

Billy,

I wonder whether a measure of volatility of MF might offer any clues? When you compare the last 20 days, to the previous 20 days MF there appears to be quite a difference. Perhaps an increasing volatility would point to using EOD and a low volatility to using RT. Just a thought.

Dave

[Billy]

Billy,

I wonder whether a measure of volatility of MF might offer any clues? When you compare the last 20 days, to the previous 20 days MF there appears to be quite a difference. Perhaps an increasing volatility would point to using EOD and a low volatility to using RT. Just a thought.

Dave

Dave,
The volatility factor is already included in both RT and EOD models. Backtests found thresholds of volatility for deciding to short or stay in cash after an exit from a long position. The current volatility level is very close from forbidding short positions. Volatility is also included in the porosity allowances.
The volatility is furthermore used for the optimal stop rules and LT/ST settings of the EOD robot beyond the EOD model. Ultimately, we hope to develop a RT robot with its own optimal volatility stops and LT/ST volatility settings. Only backtests wil tell us if volatility must be weighted differently in RT or EOD.
Billy

[DJones]

Pascal:

Another thought for you guys. It seems to me that the obvious value of the RT model is in knowing *when* trading is happening, whether "when" is defined by time of day, relative price, or (I think more likely) a combination of the two.

It certainly looks like big bullish money is coming in at the very end of down days, and I'm very interested in watching over the next few days (weeks?) to see if that pattern has more predictive value than, say, steady inflow throughout the day. Maybe the pattern is more predictive when missing at end of day on up days (and perhaps entering on the next day's first noticeable, whatever that is, price drop). Data, data, everywhere data.

I think we've all seen that something's going on with big up money at end of day. And it's been happening for several weeks now. Again today. Unfortunately for those of us who remain confidently long, the price increase is yet to happen, but hope springs eternal - and stops protect eternal springs...

Who are these guys? Sovereign buyers of gold? Funds? Who knows. I'm reminded of Butch Cassidy and the Sundance Kid being chased for miles and miles and days and days....Sundance: "Who ARE those guys."

And, like Sundance, we're all getting more than a little annoyed with them...Now that I think about it, I don't like the analogy at all. Let's hope our ending is better.

[davidallison@gmail.com]

Dave,
The volatility factor is already included in both RT and EOD models.

Hi Billy,

That's very interesting. I thought the volatility you are using is 'price volatility' and didn't realize you were already using 'MF volatility' It just shows how little I understand what's going on.

Dave

[Billy]

Hi Billy,

That's very interesting. I thought the volatility you are using is 'price volatility' and didn't realize you were already using 'MF volatility' It just shows how little I understand what's going on.

Dave

Dave,
I am sorry if I am misunderstood here. We are using price volatility for the model rules and not MF volatility.
MF is a “normalized” value expressed in percentage terms. This implies that the MF value is volatility-adjusted by nature. Using the volatility of the MF itself would then suppress the advantage of using a normalized value.
Billy