Traders' Roundtable

Hi

If you have a technical system, it is very enticing and seemingly logical to use your brain to alter your trades when something happens in the real world. For example, if there is a crisis in the middle east or an explosion at an oil refinery, it is tempting to go long oil or take off your profit target to let your wins run. Do any of you have any good thoughts on overlaying macro/news influences on your technical system?

Thanks

Rod

For most people, the vast majority, this will be extremely dangerous.

There are always "fundamental" reasons to do most anything. Most people will use fundamentals or news at the wrong time to do the wrong thing.

I believe that discretionary overlays work for those who have already developed a disciplined approach to the markets when used very sparingly. For those who are not already successful, the concept can be very dangerous.

The problem is that our brains don't work as well when dealing with uncertainty and risk as we'd like to think. In general, human judgement is not very good for these sorts of decisions

- Forum Mgmnt

Thanks c.f.

The issue I think is - if one has developed a good understanding for a market or two (e.g. stock indices and oil) then one feels, probably rightly, that they can outperform a mechanical system by considering more factors than only historical prices.

The problem, I think, is when relationships between different inputs change (e.g. in mid to late 2005 stock prices were often moving inversely to oil prices on a day to day basis, but now they are not so much) the trader may not know what to do, as the cause and effect relationships he believed in are no longer as probable. Not knowing what to do causes stress.

I think a discretionary component to any system is a healthy thing, whether used for better fills, or as a signal or market filter of some sort.

BUT a key to using this type of strategy is to track and measure each discrete portion. Track and measure the results of your purely systematic approach, and track and measure separately the effect of your discretionary trading. This is critical to understand the performance of your strategy over time.

As with any system component, you should have a plan in place for the results you expect vs. reality. If you expect to better your fills by 10%, then you should meet or exceed this target. And if your fundamental overlay stragety is intended to improve your results by some percent, then be sure that is in fact happening. If you don't track this you will never be able to improve over time.

You may find that you have a knack, and an understanding of the market. Or you may find that you don't, and you should discontinue the discretionary componenet of your strategy. This is a key learning point that until you have truly convinced yourself one way or the other, you will always be second guessing your approach.

I am mostly systematic, but I have found a couple of things that are difficult to code. One of them is whether an instrument is liquid enough right now to enter a trade with my desired position size. Yes I have a mechanical list of tradable instruments using market cap, ATR, ADV, price, etc. but when it comes to an actual entry I "eyeball" a 1 minute chart to see if it is orderly, liquid, and other "fuzzy" logic things that I can't quantify.

If it's OK then great, the rest of my system is (mostly) mechanical. I have some discretion with certain stops for profitable positions (especially on the NYSE), but apart from that it's all based on pure price action.

I agree with Tim that tracking the actual effect of your "discretionary" input is vital to ensure you are not having a negative effect. I also agree with c.f. that discretion is not generally a good idea - especially if you are not a full-time trader and don't have the hundreds of trades experience to actually be able to "out-perform" a computer program - you may find your discretion just adds a random component to your trading based on your personality and how you feel at the time, which is probably not the results you are looking for.

Paul King