I wonder if sometimes debates are not just launched for a pretext to chat . In every domain you have endlessly debates. I will take an other field to illustrate what I mean: In programing there has been a debate a long time ago between "ASM (Assembler langage) and C" then more recently between "C and VB" finally C killed ASM and VB has now huge number of programers ... because VB is just easier to grasp than C although C is much more powerful. So when something is more EASY it will attract more PEOPLE: it is the law of Evolution. Indicators are more easy to grasp than pure market's action. Everybody (today not on old days) begins with indicators like many programmers today will begin with VB (or java which is a purified C so more easy to grasp also). Some will stop with VB because they don't see the advantage to program in C++. Why because many just cope with basic programming needs like database management and not high speed 3D game which would require C++. So the people who are trading on rather high scale can think that they don't need to understand market's action (many should even wonder what's that). In fact they just don't even imagine what it is because in their heads market's action is more or less distorted random walk. On the contrary people who trades at very low scales will find market's action necessary. But market's action understanding is also useful at higher scale but it is more difficult to believe in it since it is rather esoteric at first glance (I just didn't believe in it myself when I began to trade) and so indicators will be favored. So the trend is perhaps towards large use of indicators that wouldn't mean that they are best, it is rather that they are more easy to grasp (and also to program in software packages by vendors). Then only a few "elite" traders will understand market's action like only a few assembler programers know how to program Windows in Assembler. Programming Windows in Assembler, for most newbie programmers would be considered as weird but in fact it is not so difficult and more you get a much more thorough understanding of how the windows systems works while people who program in MFC just don't sometimes know what happen underground. So people using indicators just ignore what happen underground, they don't feel the need to and they can even feel contempt with those "retarded" traders that don't want to use "modern" tools
trades from 3 Feb '03 to 28 Feb '03 criteria: long if stoch 14,1,3 fast & slow > 80 and MACDH >0.4 exit if fast stoch < 80 short if stoch 14,1,3 fast & slow < 20 and MACDH < -0.4 exit if fast stoch > 20 results: 52 trades, 41.25 pts, $4.80 round trip, net $1812.90 if no trades from 11:15 to 13:15 43 trades, 27.5 pts, net $1168.60 excel spreadsheet attached
Nice testing and some greatlooking results. When I did an automated backtest across five years, I got much lower(negative) results. But hey, the market is always changing, so who knows.... Good luck to those trading this method, and I hope it keeps working for you in the future. -blueberrycake
dkm- Should you not have deducted from each trade the bid/asked spread of $12.50? If so, then the results would be: With trading during lunch: net $1812.90-52X12.5= $1162.90 or $ 22.36 per trade With no trading during lunch: net $1168.60-43X12.5= $631.10 or $ 14.68 per trade
The above results look great. And I have a question here not a statement. In some books that I have read and from my own observation with macd and other momentum indicators, I have seen that a cross up at say 10 oclock on a 5min chart might not only look different when reviewed at 1 oclock, it may not even have happened. Now would and excel spreadsheet have the same problem?
Jem, Are you assuming that the data might have changed in some way between 10am and 1pm e.g. a data correction? Otherwise, I would expect signals to remain in place as they appeared earlier. BTW, the spreadsheet was generated simply by stepping through the 5 min bars and using closing values only.
Jem, Do you have a title and author of one of the books you read your observation in? I find your point in your last post intriguing. I think that Jack might explain this with Reletivistic and non relativistic comparisons. Excuse me Jack....I do not know you and have been following this thread. I hope it is not rude of me to try ang guess how you would deal with this issue. Please excuse me. By the way Mr Hershey....If I may follow your methodology and learn ...please allow me to. I consider it an honor to be able to read of a 48 year veteran of trading, such as yourself. Michael
If you do it this way you won't have objective results. This is because during a bar the high of the 5 min period could be much higher than the close and during this period an indicator could just move over the defined level and by the close of the bar it is back below the level. Then when you look at it on historical charts the signal didn't even take place. So except from real time testing or trading there is no valid value in those results. The trading I posted was done real time with real trading...
Alain Please correct me if I'm wrong but trading on the close of 5 min bars in real time should give the same results.