Know your share....

[Cigar]

After reading some of Trader Vic's book last night, I decided to check some numbers on KGF, who have enjoyed a nice medium term uptrend.

I decided to count the number of continuous updays and continuous downdays in 3 medium length uptrends. Note.. UPTRENDS only

So I counted how many times the price moved up 1 day, or 2days, or 3 days etc. Likewise I counted down days.

Updays
1 day upmove 37% of moves were 1 day
2 day upmoves 23% ... 2 days
etc
etc

Using the figures, I can see 85% of moves last no more tha 3 days. Conversely, 75% of downmoves are 2 days or less.

So you can form an idea, if you have bought, how may probable days it will move up, or if you are trying to time a buy on a dip, the probable number of down days until it is likely to bottom.

Of course, this would work on whatever time scale you choose, minutes, hours, days, weeks.

 

[ChartMan]

spot on !just what we need to know.

Martin

 

[Dr Iraj]

You have a good point here. Knowing your share gives confidence in trading be it weeks/days or hours. I however have not yet been to able to find a systematic approach to capitalize on this issue. I just run an analysis on BLM using My net model to count the number of the days the stock moved up before profit takers moved in. I had a similar result as your own. I also run the same model on a small cap stock (EEN)the result was totally different.

Makes you wonder about Human's pattern of behaviour.!!!!

[This message has been edited by Dr Iraj (edited 13-01-2001).]

 

[Cigar]

It was interesting to note most long runs (6/7/8 days)(and there were very few) occurred right at the top of the trends, confirming what is usually said.. the most dangerous parts of trends are the 1st and final 10%

 

[titus-uk]

Excellent observations all. I guess it all goes back to what Dr Iraj saying that stocks are like people. However, knowing which ones have some predictability makes life a hell of a lot easier!!
Mark

 

[CHARTY]

I have used similar considerations in my model. I knew that someone else may have fought about similar, but did not know who.
Personally, I found that this is a very good way to know your share. From a probability point of view, if a stock has gone up 4 dys on a row only 0.5% of the time, then it is very much likely it will fall, if it is currently entering its 5th day after 4 days up.
It is not too difficult to calculate this, particularly if you are a futures trader, and, therefore, stick to few selected securities.
What I am still working on, though, is whether we should take a very large time span, like 5 years, which would give us more samples, and, therefore, from a probability point of view, make our chances more accurate, or whether we should only stick to the last major trend.
See the Nasdaq. Do we want to go back 5 years to calculate the up and down periods, in which case you will find 3 days up is not so unlickely, or do we want to take only the current trend, initiated in March last year, in whihc case 3 days up is relatively unlikely?
Personally I am only taking the major trend.
However, I found that there is an interesting lesson/use of this indicator in SOME markets.
Would anyone guees which?

------------------
The above does not represent financial advice.