COT myth?

[cbpetro]

Hello,

I've heard much mentioned about the us of COT data as an indicator of future price action. People quote COT figures claiming "Commercials adding Longs", interest increasing, etc. as an arguement to short or long a particular pair. Looking at some of the charts presented it does look intriguing.

Like all traders I'm looking for an edge so I decided to investigate the correlation between the COT data and the following weeks price. As I understand the COT data is provided every Friday for the positions held the previous Tuesday by the Commercial, Non-Commercials, and Small Traders. The arguement is that the position held by the Commercials or "smart money" will indicate the future direction.

For now I looked at the GBP/USD pair and using about 560 weeks of data I compared the following weeks difference between open and close price against the overall position held by the Commercials. The results were less than satisfactory. It appears the position held by the "Commercials" were only correct a little over 50% of the time in determining following weeks price. Here's some of the work i did:

COT Study for GBPUSD

I also compared the weekly pivot, high/low average and the numbers where only slightly better. I also compared the delta between previous weeks position, I evaluated the prices only when positions increased a certain % and again the results again were only about 50%.

I'm inclined to say the COT data lags so much or only provides a snapshot of the current positions that it can not effectively be used as an indicator of future price.

Any thoughts?

cb

[united46]

CB

Nice work! My thoughts are:

1. As with many indicators, this may be a better indicator when commercial, large speculators positions are at extremes rather than each week.

2. In contradiction to point 1, I would never be too sure about using an indicator where the volume traded represents less than 3% of the overall market ( though it is increasing).

3. In contradiction to point 2, if you find that your getting positive returns just over 50% of the time, you've hit the jackpot. Keep those losers smaller than winners and you make money.

I think most people who look at/use cot data would confirm this is only as a small part of their systems and as a possible heads up.

Kind Regards..

UNited

[cbpetro]

Hello,

United, you have made some excellent points and I will continue my analysis of the COT data with them in mind.

Thank you for the input.

cb

[jasperforex]

Latest IMM data for week ended last Tuesday showed a long euro position of over 44000 contracts which, for those that don't follow it, is very high. This information came out Friday evening after the close and was summarised Sunday night on a few of the news wires. What happened first thing this week? Euro lost over a hundred points. Ok you can put it down to a fall in commodities oil etc but the point is that the IMM data wouldn't have had you go long euro, only short.
Sure it doesn't always work immediately, but it shows you where the danger areas are and what potential there is should a shock happen. I spent some time on it during a fx course so I know the basics, but there is so much more to learn and you should speak to a futures expert if you really want to know what its all about.

[benkorea]

For a long time, I wanted to study/search for the effects of the COT on the spot FX market. You did a nive job. Interesting!

There is an expert about the COT in moneytec, IRIS I think, I wish he can join the discussion.

Thanks anyway!

[Gamma_Jammer]

United 46 pretty much nailed the main points. It's a proxy - nothing more. IMHO anyone who tells you this market drives fx probably lives in Chicago :-)

It doesn't really factor in any flow from pension funds (but probably does catch hedge fund activity to a certain extent). But it's the only volume and positioning data widely available so might as well get what you can out of it.

[mikead_61]

united made following points:

1. As with many indicators, this may be a better indicator when commercial, large speculators positions are at extremes rather than each week.

and

3. In contradiction to point 2, if you find that your getting positive returns just over 50% of the time, you've hit the jackpot. Keep those losers smaller than winners and you make money.

i analysed the data back and forth and found:

1. total pips profit over 559 trades = 2738, i.e. 5 pips per trade. not my idea of fun if you are holding position open for a week

2. if you apply SL, situation improves a little, if you apply SL and filter by swing size you optimise at SL 300 and position swing >=3000. you get about 300 trades and total profit of 3000 pips = 10 pips/trade- still not very lucrative.

so the conclusion here is that even you have > 50% chance of winning it's not that easy to make a trading system out of it

the problem is in this system even before you make profit, position often goes far the other way

3. Now if you use SL of 300 and ABS(position swing) > 25000 then you have much better results - average profit per trade ~110 pips. But you have only 22 trades over 10 years - will not make you a millionare soon.

conclusion here is that you can use it for all currencies (about 10 trades/year) as an additional system to gen a few pips here and there - but not as a main system (what i'm getting to is that one can trade several systems at the same time)

further ideas:

a. why held a week? maybe results will be better if held only first 1,2,3... days?

b. additional filter? i.g. only go long if big trend (like visible on a daily chart) is up?

regards

[Roger Damon]

I would agree that the COT is best used as a filtering tool. The ways that I use it are basically to determine both trending and non-trending conditions, and then when trendingto have my trend following system generate signals in alignment with the COT and my counter-trend systems generate signals against it. Not perfect mind you, but it has kept me out of some really bad trades and allowed me to take advantage of some really good ones.

For those that haven't seen this done before, here's a chart showing when the COT Non-Commercials all bought US Dollar futures against the Yen and then the yen took a dive. I myself didn't actually catch this trade, but it shows you how the data can be useful.

http://www.moneytec.com/forums/attac...1&d=1156881533

Happy trading.