Multi-Pivot Clusters For December 29, 2011

[Billy]

12076

12074

12075

[Pierre Brodeur]

Billy,
I use the multi-pivots methodology that you invented on a daily basis in order to determine the most likely very short term support and resistance level for each stock and ETF that I follow in the Canadian market. This supplements my intermediary term P&F approach which also identifies R & S levels. I focus only on the 1st support and resistance cluster and determine the median. I do not look at the second cluster levels.

From day to day, the pivot clusters move up or down by a certain amount depending on the movement of price and indirectly because of the new daily pivots. I suspect that some calculation of ATR also has an influence on the movement but probably more on the expansion and/or contraction from top to bottom of the cluster.

I call resistance and support the median of the cluster: I find that there is some value in knowing that:
R & S are moving toward each other or are moving away from each other OR
R is stable while S has been increasing since yesterday
and so on.

My question is: Have you given any though to this part of your cluster model? If so, what are the trading decisions that you would make given your analysis of such R and S movements? or am I losing my time on this aspect of your model?

[Timothy Clontz]

Is it possible that our stop targets are also becoming targets for HFTs, and is there a way to determine such an obstacle in the future? I've noticed lately that the reversals after a stop are getting eerily routine (or at least it seems to be).

Perhaps we should work out a way to determine if or when that occurs in the future so that we can incorporate that in our model adjustments.

[nickola.pazderic]

Tim,

I follow top traders at Slope of Hope. It is easy to do. Simply sign up, etc...

The top traders are routinely buying at the end of a down day and selling at the close of an up day. A simple reversal strategy. See the EOD MF from yesterday to see it is not only a few traders. This is a major pattern.

Will this pattern persist into the future indefinitely? Probably not. What I see is that the top traders have a nose for the goal, even if the goal is being shaken around.

It makes me wince to see a top trader indicate a long/short GDX or IWM position when the robot dictates the opposite. Are these folks better programmed than the bots? No. But are the more flexible? Probably.

Does this situation make me doubt the robots? No. But, I sure don't like to be hemmed into bad trades.

What do I do?

E.g., today I went short at the opening and then reversed the GDX and made back 1/2 the losses. Now I''m short again. I also am long TNA until the IWM short price is reached. I call this style: "trading into a robot position."

Perhaps others will have a laugh at this. I don't know.

But I do know that I like to have the robot directions and prices in place and trade my way into them.

This has worked the best for me during the choppy market over the past couple of months.

[Pascal]

Is it possible that our stop targets are also becoming targets for HFTs, and is there a way to determine such an obstacle in the future? I've noticed lately that the reversals after a stop are getting eerily routine (or at least it seems to be).

Perhaps we should work out a way to determine if or when that occurs in the future so that we can incorporate that in our model adjustments.

Billy is taking time off the board until next year.
I do not think that our stop targets are becoming targets of HFTs.
Last stop was overshot by some distance. Hence, we were not the target.


Pascal

[Billy]

Billy,
I use the multi-pivots methodology that you invented on a daily basis in order to determine the most likely very short term support and resistance level for each stock and ETF that I follow in the Canadian market. This supplements my intermediary term P&F approach which also identifies R & S levels. I focus only on the 1st support and resistance cluster and determine the median. I do not look at the second cluster levels.

From day to day, the pivot clusters move up or down by a certain amount depending on the movement of price and indirectly because of the new daily pivots. I suspect that some calculation of ATR also has an influence on the movement but probably more on the expansion and/or contraction from top to bottom of the cluster.

I call resistance and support the median of the cluster: I find that there is some value in knowing that:
R & S are moving toward each other or are moving away from each other OR
R is stable while S has been increasing since yesterday
and so on.

My question is: Have you given any though to this part of your cluster model? If so, what are the trading decisions that you would make given your analysis of such R and S movements? or am I losing my time on this aspect of your model?

Pierre,
These expansion/contraction phenomenon’s are mostly due to volatility (ATR) changes from day to day and to the daily levels. Then you also have the calendar effect of new (weekly, monthly, etc.) levels.
I am not sure if using the median of the clusters as S/R is very helpful without some backtesting. From our own research, it is most judicious to use the strongest floor levels as S/R reference. They gave the best performance when used as targets, although the robot performs better with its trailing stops rules for exit and never exits a trade on target.
But in my discretionary trading for position-sizing I use a lot these strongest S/R floor levels as targets or entry for scaling in and out of positions. That’s the way market makers do it in a sound “normal” market environment. The top senior market makers are always in charge of the longer timeframes inventories and size-in or out at the corresponding levels. They actually distribute to junior traders the daily/weekly/monthly… quotas to trade at the respective daily/weekly/monthly floor levels. Nowadays, all this process is very much optimized with algorithms and HFT. I doubt the median is an important number in their programs.
I also have profitable experience relying on the cluster strengths ratios to decide more or less aggressiveness in my daily trading plan. I am eager to read the conclusions from Paul’s research in progress on this topic.
Billy

[Pierre Brodeur]

Pierre,
These expansion/contraction phenomenon’s are mostly due to volatility (ATR) changes from day to day and to the daily levels. Then you also have the calendar effect of new (weekly, monthly, etc.) levels.
I am not sure if using the median of the clusters as S/R is very helpful without some backtesting. ...
Billy

I will adjust my process accordingly
Pierre