How to research and verify trading ideas

The only exit is the one given... at some point it will trade above the median of the shortest window (I think I gave 5 in the original system) and that's when you get out. And yes, this could potentially be at a very severe loss.

Every day would generate a new set of percentiles, but once you have a trade on the exit rule applies.

 

this way of trading seems very similar to that stealthtrader pdf file. im not saying that it is talontrader, but would actually like him to look at it to see if he agrees with the way ST outlined how to design a trading plan. sorry if this is off track a bit. anyone have that link?

i have it if anyone knows if it is free to post . i didnt want to post it if it isnt supposed to be

 

Post it and I'll look at it. I also know Larry Conners has some stuff that is similar to this, but I have issues (of course, I always have issues) with his research process.

 

where would you start to develop a theory like the system you just wrote about. what made you come together and say "200 day close in 50 percentile and last 15 day 10%" . just curious as to the train of thought to piece those together. why 200 and 15 rather than 300 and 10? im trying to understand how you start your process of thinking you might be on to a new edge. i read the start of the thread, but i guess i was looking for a bit more direction on this.

 

Quote from bravetom:

I never understood the *edge* thing.

Neither do 98% of the traders. They throw their money at the market, expecting to start milking it after some arbitrary period of time. Eventually, they realize it does not work that way

trading is simple indeed, probability with mental strength is all you need.

Obviously, this is not true. You need two things: serious edge (positive expectancy) and superior money management.

Everything else is distraction, smoke and mirrors...

 

No... from reading your post I wouldn't expect you to understand the *edge* thing.

I can tell you I know far more professional traders who scale out than who add to winning positions. Personally I tend to put on my position in 1 or two pieces and scale out as the position moves in my favor. I may add another (smaller) piece to trade around the main position.

This is a far better plan for consistent returns on a professional scale than saying the key is adding at breathtaking prices, based on my experience and contact with many other professional traders.

Those of us who make our livings at this tend to strive for consistency above all else.

 

In my opinion you need:

An actual edge in the market. There are many possible edges.
The personality, discipline and learned behavioral traits to apply that edge successfully.
Proper money management and risk management.
Sufficient bankroll. Sorry folks... $3K in an Ameritrade account probably ain't going to cut it. The people I have trained have usually been underwater anywhere from 10K - 50K before they started making money... and this is working 10-12 hour days and weekends... trading as a full time job... with someone reviewing your work, answering questions and guiding you every step of the process. So... if you're trying this at home it may take longer than you think. (not impossible!)

TZ I think you consistently underestimate the importance of discipline and psychology. Every professional I know emphasizes this... every single one... it's one of the very few common threads.

 

Honestly, I read the post that basically asked for a purely technical system. We trade lots of pullback systems so I just threw this together in 5 minutes, wrote 10 lines of computer code and ran it through one of our systems to check it.

it's not fantastic but it works and illustrates the concept and provides a good departure point for discussion and development if people are interested.

The question you're asking though is how did I come up with it in the very beginning, right? Let me answer that one later after I've given it some thought. I can't write a two paragraph answer to that... the short answer is after seeing a lot of patterns and market data but I can give you a better answer than that... and I think the answer to that question will significantly advance this thread.

 

I just saw this post tonight and read the follow ups. Thank you very much for posting this method. I do appreciate it.

A question from this arts and social science degree guy please. (Btw, I now have a backtesting program and have begun self study in math and statistics. It’s actually very enjoyable and satisfying at this point in my life. Previously my most demanding math question in college was attempting to determine how many full draw 5 second bong hits we could get out of an ounce of goo bud.)

You said:

“1. it is above the 50th percentile of the past 200 days' closes.”

If I understand this, we simply take .5*200 and get 100 of 200. So then I place my 200 closes from least to most and then only look at items from 101 to 200. Correct?

“2. it is below the 10th percentile of the last 15 days' closes.”

If I take .1*15 I get 1.5 and then raise it to 2 of 15 (or 199.92). But since we want the close below the 10th percentile, it must be the first number, or 194.34. If this is true, then in all cases I could simply take the lowest close of the 15 day period. Correct?

Thanks again. Great stuff.

 

great and yes you are correct in what i am asking. the step between thinking of the idea and then writing the 10 lines of code is a big one. im just speaking for myself, but if i understood where you begin your thought process im guessing it would help me/others in being able to throw some ideas into some coding and doing the same. i also am not underestimating the potential screen time needed. after reading some of the earlier stuff i was trying to think of a place to start in regards to trading a range bound area. i started looking at if X weeks close with a very similar candle body/wick size sideways, is there a test that can be run?

to be more specific, i was looking at a weekly timeframe and looking at candle bodies/wicks that closed right next to each other and didnt exceed the prior candles price range yet tested both sides. then i was looking at scaling into a day time frame and potentially looking at a doji when one side gets tested for entry. target is the other side of the range. not sure if im on the right track with that.