10 items for inclusion in automated trading code-(2)How to evaluate market direction

[Charlton]

This is the second of the threads around the list provided by Iraj about what to include in automated trading code

1) how to reduce risk using diversification
2) how to evaluate the market direction
3) if market direction in doubt then how to hedge
4) how to scale in /out in real time to reduce risk
5) pos sizing relative to capital
6) different pos sizing during oscillation and trend mode
7) exit /stop loss during osc/ trend modes
8) real time risk manager
9) execution of the trades
10) dynamic profit making based on Money management

In the discussion on the previous item in the list Glenn discussed how market fundamentals, news events and current affairs need to taken into account in determining market direction. He also said that these things cannot be incorporated into automated code and require discretion.

What we can do, however, is incorporate the results or this discretionary analysis into automated code i.e. once we have established whether the market is long or short, likely to be highly volatile or not etc we can switch on the most appropriate code to trade that type of market.

I think therefore that the most fruitful direction of discussion for this item in the list would be the various ways in which we can determine market direction and the factors that influence it. We might wish to concentrate primarily on the direction of the US equity market and it's main determinants.

Charlton

 

[Glenn]

What we can do, however, is incorporate the results or this discretionary analysis into automated code i.e. once we have established whether the market is long or short, likely to be highly volatile or not etc we can switch on the most appropriate code to trade that type of market.

Yes - a good point.

Glenn

 

[lote_tree]

Guys

Just thought I'd throw a few ideas in the pot as far as quantifying market sentiment. One resource that is useful for me and which is easily accessible, quantifiable and gives an excellent idea of market sentiment is the CFTC website. There you will find reports that show the commitment of traders on the long or short side in many of the major futures/options markets.

Obviously when looking to “code the quantification of market sentiment” for the Macci strategy it would be a good idea to analyse the commitment of traders in the DOW, S&P 500 and NASDAQ futures markets. I look at the reports which combine the options and futures contracts.

Here is the link to the report which combines the options and futures contracts:

CFTC Commitments of Traders Long Report - CME (Combined)

Here is the link to the report for the futures contracts only:

CFTC Commitments of Traders Long Report - CME (Futures Only)

Just scroll down to the bottom and you’ll find information concerning all the stock indexes.

Cheers
Naeem

 

[dcraig1]

Where to start on this one ? There are so many things that might be useful intrday.

1. The behaviour of the European markets before the US open. For example if the DAX is screamingly bearish, it may be ill advised to take any long position(s) in the US session. How to quantify this ? Perhaps percent change since the open on the DAX combined with some measure of how it got there ie how smooth is the trend for the session - perhaps sharpe ratio or fractal dimension. Some historical statistical analysis might be needed.

2. Analysis of the market delta - ratio of traded volume at bid and at ask. Possibly for the preceding European session as well as US session. Again need historical stats.

3. Order book analysis for ES. If size is consistently on one side of the book, the market nearly always travels in that direction until the book behaviour changes. Processing book data is not so easy to do because it is extremely "noisy".

4. Advance/decline ratio for NYSE.

5. VIX ????


.
.
.
.
.

 

[leovirgo]

I'd like to add just one thing... price action.

Cycles alone do not tell the whole story. There must be analysis of how price is behaving at certain cycle nodes/phases e.g, if it is OB level, is price higher than that of previous cycle high? If cycles start to turn up from o/s, then is this low lower than previous one? That's how I come to distinguish between oscillating mode and trend modes. In trend modes, cycle high is almost always higher than previous cycle high, then forget o/b and just go long for a trend. I used to short markets at o/b points but it is easier to go long in a trend even if it is o/b technically but prices are higher than previous highs. Grey1 used to say exhaustion points are time to exit a position rather than entry points- of course there are exceptions.

 

[Trader333]

In my view one of the key reasons why intra-day trading is so difficult is that each bar contains less information the shorter the time frame. For each daily bar then all traders that are intra-day will have settled and exited which removes the temporary manipulations that obviously do occur. I have found that EOD technical analysis and especially price action becomes more reliable and when I have factored in weekly and monthly analysis it gets even better.

Of course the converse view is that more risk is involved because any moves are larger and this is true but by position sizing correctly this rarely becomes an issue for me.


Paul