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)
Hi all,
I've just finished reading the coin toss thread with great interest .
Many times you see, hear and read that exits are the key and you can make money on a coin toss.
This has never been my belief and I gave a talk to the ATAA about this and how important I believe entries are.
To me there are 3 very seperate sections of a system. 1) money management (ie risk & position sizing etc). 2) Entries. 3) Exits.
I believe that entries should be the thing that thrusts you in the right direction.
To that end, I've never seen any proper testing of entries. They have always been tested as part of the whole system, so who is to say the it was the MM, entry or exit that ultimately made a winner/loser.
So I came up with the idea that the entry should be responsible for making money on better than 50% of the trades in the first 2 or 3 bars of the trade.
So to test that I do have an entry that makes money and is not just a coin toss, I force the trade to exit after 3 bars and view the results.
Note : no stops of any type are used.
Now looking at the results we can see if we have > 50% winners, what our MFE is and also very important is our MAE. With the MAE we can now see what a possible stoploss level is for this entry strategy.
So all of this tells us if we have a winning entry or it's just a coin toss.
By the way, after the 3 bars is when I really expect the exit to take control of the trade.
So, thoughts ?
Neil.
Sent from my GT-I9100T using Tapatalk 2
Can you help answer these questions from other members on NexusFi?
Yes I do this when taking a raw idea through my development process. After rough coding I will test the entries in isolation by looking at MFE:MAE at different time intervals. Seeing what happens after 1, 5, 10 or 100 bars helps expose the character of the strategy. I also like doing this as it's not deterministic, often things pop up that you were not expecting.
@NW27: This does not make sense to me. First of all 3 bars is not enough to judge an entry. Secondly what matters is the winning percentage and the ratio average win to average loss. The winning percentage alone does not mean anything.
If I had to test the quality of an entry I would select a standardized exit to judge the entry against it.
My standard exit would look like this:
-> profit target: a limit order N volatility units away from my entry price
-> stop loss: a stop market order N volatility units away from my entry price
-> time exit: a market order M bars after the entry bar
That would give me a framework for testing entries. Any position would be either exited with a predetermined profit, a predetermined loss (+ slippage) or after a specified number of bars or time period.
For example I might use 5 average true ranges as a target, 5 average true ranges as a stop loss and 50 bars as a time exit.
I can't wait to see you trade with a live account using this with no stops.
Of course, since you did not specify the size of the 3 bars I would like to suggest...
Monthly Bar Chart. LOL
That should give you plenty of room to blow up the account before the third bar is completed.
Point is, you need to at least specify the size of the bars.
Second point is, it will still blow up your account.
Trading without stops is the first sign of insanity.
BTW, the CoinToss strategy can be modified to exit after a user specified number of bars with a small modification if you want to test it.
Rejoice in the Thunderstorms of Life . . .
Knowing it's not about Clouds or Wind. . .
But Learning to Dance in the Rain ! ! !
Now that I'm at a keyboard and not trying to type on my phone, perhaps I should explain a bit more.
In the talk I presented, I started with first you need to decide on a market and time frame that you wish to trade. From here you then look at the price movements that you like (ie a trend) and how many bars they go for. In this example, I was using trades that lasted for 10 bars. From this I then suggested that the entry should be what thrusts you in the right direction for the first 20 to 30% of the targeted trade ie 3 bars.
Does this not depend on how long you expect your system to trade over? As an example, if it was a typical Al Brooks trade, 3 bars would be the end of the trade.
Hence the reason I do not apply a standard exit to the trade only the forced time exit.
You then look at the results of this data and determine if you have a good entry. For this, I use a bar graph as per below, showing the profit on each trade. From this, the Profit & MFE shows that we have a winner and the MAE shows possible stoploss levels.
Isn't all of this exactly what I have done but I have not forced a Profit target or StopLoss on the trades, so my graph shows a truer representation of what the entry is doing. Is it thrusting me in the desired direction? Is it giving me a better than 50% chance?
Using your restricted testing ie Profit target of 5 and StopLoss of 5 but still exiting on a forced 50 bars.
You would have a result that looks like this -
Which graph tells you more about how good the entry is?
Isn't any thing that happens 50 bars after the entry (whether the bars be 5tick,5min,5hours), just luck and in the lap of the gods? Surely what happens 50 bars later has nothing to do with why you entered. It is now more the Exit strategies job to maintain. Your trades must last for 100's of bars?
So Fat Tails,
I really enjoy reading your comments on the various posts and I hoped to entice you into this one. It worked
Given your comments on the coin toss entry thread, do you specifically test to see if you do have a valid entry? Is it the above method you described?
@DavidHP
I guess you didn't get the purpose of the thread. It was about not relying on a coin toss but trying to validate that we do have a good entry strategy. It is not about a complete system. The period of the bars ie 5tick, 5min, 5 day has nothing to do with it.
Yes, 3 bars can be short or long, depending on your approach, if you are a trendfollower 3 bars look short. Also there are certain types of entries, for example an aggressive entry on a volatility breakout (if you use the SuperTrend for entry timing), which will not yield good results on the first three bars, but may well do so, if you take 20 or 30 bars.
I am not a 3 bar trader.
Also what I think is important, that the profit target is hit prior to the stop loss, not when it is hit. So your approach to solely rely on a time based exit also has its limitations.
I was talking about 5 average true ranges, not 5 ticks. I am not going for a bone, I watch out for the steak. If I have selected a decent entry I want price to move in my direction more than just one bar length. Volatility as shown by the bar length is 1 average true range. That said, if you put any stop at 1 ATR of your entry you will certainly be stopped out by non-directional volatility. Starting from here my stop is at least 2.5 ATRs away and my profit target should be further away than the stop to attain a R-Multiple of least one.
I have therefore suggested to use 5 volatility measures, 3 or 4 would also have been possible.
Further I have not chosen 50 in an arbitrary way. Volatility usually expands with the square root of time. The average 4 bar true range is about twice the amount of the simple average true range. If you take the average range of 25 bars, it is about 5 times higher than the simple average true range. So in average my trade should last 25 bars, until it either hits the target or the stop. I have simply selected the double duration - that is 50 bars - for my time exit.
With a target and stop of 3 average true ranges, the time exit would be adjusted to 2 x 3 x 3 = 18 bars.
I am not pretending that this is the only way of testing entries. Your method only relies on the duration of a trade. My method also has an element, which I use for a real trades, a profit target and a stop loss. It is just an alternative approach, which I think is more realistic than a pure time exit.
I use a smaller timeframe chart for entry timing, but enter trades based on a larger timeframe trend or reversal. Yes, indeed I expect my trades to last more than 3 bars. I am not a scalper! In fact I just expect my trade to move 3 to 5 average true ranges in my favour. This is not that much. The time exit is only used in case that volatility did not pick up and price has gone nowhere.
Further I believe that an exit strategy should not be tested independently from an entry strategy. For a quick countertrade, I would not use a trailing stop, as I know that it is likely that the trend will resume and kick out my stop.
