A question of choice !

[Dow Dog]

I have a slight problem relating to my stock selection method which I am wrestling with.

It relates to the stock selection process by which I select stocks to trade on any particular day.

I have a "watch-list" of 15 stocks which I update every weekend based on various parameters.

My system allows for a maximum of 2 trades per stock per day over 2 trading sessions - afternoon and evening.

Therefore theoretically, I could do 30 trades a day.

In reality I aim to do a maximum of 6 - over the course of the 2 sessions ( afternoon and evening ).

I select 3 stocks per day from the 15 for trading. The selection process is based on a points scoring system from the results of paper trades of the 15 stocks. The process takes into account number of winning trades, number of losing trades, number of winner points, net profit, reward to risk ratio and ATR to price ratio - from the past 5 trading days.

My paper trades allow for plenty of slippage and throw out a winning rate of 85% once a trade is entered. Percentage of " no trades" due to entry criteria not being met is around 10%.

So, all very good or so I thought.

The problem is that I am either just plain unlucky or I am missing something because as soon as I select my 3 best performing stocks and enter real trades, I find that they invariably become losers.

Is the flaw that, I am picking the stocks which have recently performed beautifully and are now due a loser or is there a fundamental flaw in my selection technique ?

I would be grateful for suggestions.

 

[DaveGos]

Dow Dog

I had a similar problem a while back, I finally realised that I was imposing what I thought the price should do based upon it's previous performance and ignoring what it was actually doing. Now the only historic information I include in my trading strategy is the previous day's high and low and even they are secondary reference points.

 

[BBB]

As Grey1 would say: we're not trading history.

Maybe youre getting in too late. The healthy historical results could suggest the move is due for a healthier correction. What would have happened if you took the other side?

 

[Dow Dog]

I think you are missing my point ( or maybe I am ! ).

Of the 30 potential trades, I can virtually guarantee that, under my system entry criteria on any given day :-

27 will trade and of those 27 around 23 will be successful.

4 will be failures and I am trying to avoid those 4 at all costs.

But I seem to keep running into them.

Question is - how to avoid the failing 4 ??

 

[DaveJB]

How automatic is this?
If you get 4 bad trades out of 30 so what... but if you get 30 signalled and only take 4 of them, which all dive, then you aren't trading your system... you are taking system signals and then deciding which of them to trade. (Departure from system then occurs). With 85% winners you ought to be able to sort this simply by saying 'I'll trade every 5th signal' or similar, to take the trade frequency down to the required level without introducing any personal choice in to skew the system itself. Doing this you might well extend winning/losing runs, but if the underlying method is 85% good and you make 4 trades a day and lose, then either your system is pants (sounds unlikely) or what you trade involves a choice that is effectively negating it.
An example I've found with EoD P&F is that I get very similar percentages from a P&F signal when compared to the same signal backed up by P&F Rel Str - so by combining the two I reduce the number of trades whilst maintaining the 'hit rate'.
Dave

 

[Dow Dog]

Not sure I follow DaveJB

 

[Trader333]

Dow Dog,

A couple of things to I have to comment on:

I am surprised that you are complaining at only having an 85% success rate. I would just live with it as you obviously are doing extremely well and no system is ever perfect in the trading world.

Your quote:

Is the flaw that, I am picking the stocks which have recently performed beautifully and are now due a loser or is there a fundamental flaw in my selection technique ?

This statement is almost the inverse of "Gamblers Fallacy" ie Gamblers think that because they have had so many losers they must be due a win. You seem to think that because these stocks have done well they must be due to underperform. This type of thinking has been proven to be flawed.

If you are getting an 85% success rate I would be wary of trying to further improve it as there is a point at which further optimisation will prove to be counter productive.

I say Well Done for having such great success to start with.


Paul

 

[DaveJB]

Sorry,
85% winners when paper trading, 0% doing it for real - you are therefore altering the criteria for stock signal, entry, exit, or 2 or 3 of these. If your system gives 85% winners and throws up 30 trades a day, you actually trade 4, then you are filtering 30 trades down to 4 and losing, when you'd expect at least 3 to win out of the 4. Consequently either your underlying system isn't 85% effective or the way you throw 26 away is actually negating the system's effectiveness... the system picks a lot of good trades, but the system you use to pick which four to trade out of 30 is awful.
Another point is that if the average is 30 trades with 85% win rate then you expect approx 25/30 winners - to be able to identify 4/5 of the losers so unerringly is quite something! <g>
Something here just doesn't add up - you need to find out where the flaw is... choices are:- (1) your system isn't 85% effective, or (2) your method of reducing trades to manageable numbers is badly wrong and you'd be better picking the 'pick 4-5 from 30' by pure chance... in which case you'd expect to get a bad trade in line with the underlying system effectiveness, ie 6 trades with 5 winners, or (3) it's a good system on paper but when you trade for real the slippage etc is killing you.
Dave