Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
Currently I use ninjatrader and I am satisfied with it. But it does not really seem to do the sort of analysis I want.
Ninja seems to work like this.
You set up a strategy with conditions. (example only)
1. MACD crosses above zero line
2. ADX greater than 25
3. VOL this bar is greater than previous bar
Next you set Profit target and stops
1. target 8 ticks
2. Stop 4 ticks
Now you backtest your little strategy to find out how much money it made. But this is not what I want.
I want to examine from my data pool what happens to price when
1. MACD crosses above zero line
2. ADX greater than 25
3. VOL this bar is greater than previous bar
(Individually and in all combinations)
I do not want to see if this strategy is profitable. What I want to see is what occurs to price when these conditions are present. Adding in a target or stop strategy forces me to include a factor into the analysis with no reasoning behind it. The strategy analysis procedure seems to not be in the correct order. I want to be able to perform the analysis in a more scientific way.
Any suggestions? Thanks
Can you help answer these questions from other members on NexusFi?
"What I want to see is what occurs to price when these conditions are present."
This does not sound like a scientific formulation of the problem. If you wish an anwser, you need to be more specific and explain, what exactly you intend to do.
Okay. First I want to isolate signals from concerns about stops or targets. Stops and targets can be determined later after we determine the probabilities for our setups. Additionally our setups usually have more than one condition, so we want to isolate each condition and examine how it relates to price. Then we want to examine each combination of conditions to see how they relate to price movement.
Why? First the success rate of a strategy can be alerted by a vast amount my changing the stops and targets. We can not know the correct stops and targets until we know the probabilities (or statistical breakdown) of the outcomes of the strategy. Including them in the analysis before we know the probabilities makes no sense.
Many people promote strategies, or indicators, or setups with rather vague or anecdotal probabilistic references.
"Go long when A and B happen and you have a 80% of having a good trade". Oh yeah. Prove it.
Lets load them in and look at the results. Lets see real data.
Many things, and I think trading strategy analysis is one of them, do not lend themselves well to intuitive or instinctual answers. Empirically obtained or logically derived answers are usually better.
So if I my plan is to go long when the ADX is over 25, DM + is above DM -, and MACD has crossed zero within the last 4 bars; I want to know that there is a 70% chance of an 8 tick up move, a 5% chance of a 10 tick down move.
I want to know as exactly as I can what might happen based on what has happened. So I can make the best judgement of when to enter and where to place those targets and stops.
This seems very scientific. I want to be able to gather clearer and more discrete evidence. And I find it hard to do when everything is jumbled together to early.
I don't understand what this seems to be a difficult question. All I want to do is find out how often particular setups occur, and what has happened when they have occurred in the past. I have come to believe that ninja trader might not be the correct tool for this. I just want to know what the correct tool might be?
Many people enter a trade based on some technical signal. They also enter that trade based upon some expected risk reward calculation. That might be a 5 tick stop and a 10 tick target. This is picked because of some rule of thumb about making the reward at least twice the risk. But if you don't know how often your signal actually reach the 10 tick target and 5 tick stop, these selections are just without sufficient rational basis. You might as well make a 20 tick target.
For most of the basic indicators this should be readily known information. I really should not even have to figure it out.
I just want to be able to make sound statistical and probabilistic trading decisions. How can you do that with out the numbers. When you play poker you can calculate the odds because you have actual numbers to work with. I am just trying to figure out how to get actual numbers.
Thanks. I don't mind being wrong. With my limited background in both mathematics and trading, I would be surprised if I wasn't. Your reply puts me on the correct track. No blame. Just gratitude.
Cheers
@mcteague: Thank you for your friendly reply, much appreciated.
I think that there are a few interesting threads that cover the subject. Determining the winning probability and the average winning and losing trade is just the start. Once you have a proftable setup, it is all about position sizing. Too small positions will not lead to a satisfactory compounded growth rate, too large positions will deplete your account. Here are two links that cover the subject.
I'd like to start a discussion on Risk of Ruin. The general concept of a Risk of Ruin analysis is to determine if you will go bust over a statistical number of trades. For example you may have a probability edge of 65% but after factoring commissions, …
I generated the following trials using a trading simulator with the following parameters:
10000 trials, each trial having 386 trades, with 1 to 2 Risk/Reward ratio.
Variable winning percentage.
The number of trials is set at 10000. This seemed like a …
I am trying to develop a trading methodology that uses a point count system. The point count is used in the game of Bridge to help players determine the strength of their hands. It is a bit artificial, but has great value in simplifying evaluation. (4 points for ACE, 3 for king, 2 for queen, 1 for Jack).
What you can do in trading is take the indicators you use and assign a point value for specific conditions. Lets use MACD as an example as everyone is familiar with it. There are specific conditions we might want to assign values to. This can be applied to any similar indicator
(Example only)
Just crossed zero line 4 points
Just crossed signal line 3 points
Above zero line 2 points
Above signal line 2 points
increasing divergence 1 point
increasing convergence -1 point
Below signal line -2 points
Below zero line -2 points
Just crossed below signal line -3 points
Just crossed below zero line -4 points
You do something similar for all the indicators you use. Now you have a total numerical value based on your indicators. Then you figure out how many points you would want to take a particular action. Enter long or short, How many contracts. etc. You also might want to have firm rules not based on the total, but on one particular indicator. Like never go long with a negative MACD score.
But basically you add up the numbers and act accordingly.
So this is just a particular method for determining "Trading Rules". I think this is a probably a good method for beginning and intermediate traders. Not complete and has some limitations. But it forces you to look at all your indicators and eliminates fear and greed responses.
I should say that indicator could include more than just "indicators". For example I subtract 2 points for a lot of recent candles with big wicks and tails. To Volatile.
So the reason I am interested in what happens to price when particular conditions or events occur is so that I can assign correct point values to those events. I am just trying to develop this point count system
The other problem I am trying to work on relates to ATR and stop levels. I know that more advanced traders recommend setting stops based on swing pivots rather than ATR. And I don't use ATR for stops. Except that I theorize that any stop (or target) inside the ATR must be statistically unreliable. If your stop is inside the ATR you will hit it just by noise fluctuation. This is a problem for me when the ATR is high, say over 2 points, because I would take 2 points in the ES every time if I could. The problem is that ATR has no actual relationship to what a setup will do. It is just completely arbitrary to use 1 atr as stop and 2 atr as target. But it does tell me, I think, what the noise level is. And that should be useful. There is a relationship which must be considered
Anyway I just wanted to explain what it is I am actually thinking about and trying to do.
