PDA

View Full Version : Why traders loose with good trading systems


inpilna
07-23-2006, 02:17 AM
Why traders loose with good trading systems...

A major misconception about a good trading system is that it should give a higher percentage of winners compared to losers, say 70% winners and 30% losers.

Amazingly very few traders understand, that the main characteristics of a great trading system is not how often it’s right, but how explosive a move is when it’s right.

Thus a system (like MTPredictor) that only gives about 38% winners and 62% losers can be considered a good profitable system. Why? Because when it’s wrong, your loss is limited to 1R (1 unit of risk), and when it’s right the average profit is about 3R.
So assuming that 2 out of 3 trades are losers, you total loss is –2R and your profit is +3R giving you total of +1R profit. If you translate it into real world numbers, when you risk 2% per trade of a $25,000 account, i.e. max. $500 per trade, it means that after each 3 trades (on the average) your equity increases by $500.

Many traders though would look at MTPredictor system and say, wow, it’s a great system if you use it in reverse! That is, when there’s a buy signal, you go short, and when there’s a sell signal, you go long. Why is it great? Because then, according to statistics, you are 62% right and only 38% wrong. It’s hard to argue with this logic, because we don’t know, how explosive is the move when you are on the right side of the trade. But proponents of this anti MTP system say that even assuming that you take profits at 1R, and loses at 1R you would comfortably make some money: +2R +(-1R)=+1R. So you get the same +1R profits with a lot less stress.

This puts a profitable system developer in a difficult situation.

- First of all, he has to provide users with average win/loss ratio. It’s probably can be done by using regular back testing. In MTPredictor case we know it’s about 38/62.

- Secondly, he has to show how explosive on the average the move is when the system is right. And not the maximum move, but the move based on standard management technique. This aspect is not well studied with MTPredictor and the developer admits that no back testing was done in this respect. But some of us know from practice, it’s approximately 3R.

- Thirdly, you can’t get discouraged after 2-3-4 losing trades and stop using the system, you have to take all the trades every time when a trade is triggered.

- Based on the 3 previous facts, the most important message the MTPredictor developer has to deliver to MTPredictor users (assuming they follow position sizing and honor stops): if you use this system, your loses are limited to 1R. When you have a winning trade, you HAVE to let it run using standard management technique to at lease above +2R and higher. If you take profits on a winning trade at +2R or earlier – you GUARANTEED to loose even using a proven profitable system like MTPredictor.

All of us remember fiasco with Surgical Strike approach, you take profits too soon – eventually you lose or break even. It just shows you that even experienced traders forget about this simple issue: with a system with percentage of winners below 50% you HAVE to let the profits run if you want to win with this system.

So, why do traders loose with good trading systems? Just think about it: you had 2 losses in a row (we know it can be 5-6 losses in a row), finally you have a break: a winning trade; you’re making 1R, then 1.5R and finally you take your profit too soon at 2R being anxious to protect the profit remembering your 2 previous losses. That’s why it’s so important for the developer to assure users that the system is tested, you should trust the system and you should expect to have 2 losers out of 3 trades, and that you absolutely have no choice but to (by using management technique) ride your winner to a minimum of above 2R and preferably to at least 3R.

CapTech
07-23-2006, 12:46 PM
I did something similar back when I first started trading MTPredictor. I traded two separate accounts, one taking the signals produced by MTP in the prevailing direction and the other account took a opposite signal to the one produced by MTP. Each signal either with or against only risked 1R and winners were allowed to run their conventional course and managed accordingly. I however trade MTP signals differently. I don't wait for a reversal bar to enter on a tick above. I wait for a typical WPT and enter at it's base with a 3 to 4 tick risk only off that bar as long as order flow and arb relation indications are in my favor.

MTP does a good job at showing Turing points at WPTs so the few I miss going with the signals and not waiting for a reversal bar do play out because my !R unit is only 3 to 4 ticks. I trade in size and can't afford a 1R unit provided by a reversal bar setup range. I did this experiment looking at QM, ER2, ES and YM on a 2, 3, 5, 15 and 60 min basis. Taking both signals for and against. I can tell you this was a very trying and busy period. I don't know if anyone has done a study on just taking all MTP signals with 1R management on all the symbols above and all of the periods. You will find some scary results with losses mounting before the percentile profit run appears. Just a thought that maybe 4 users could do the above experiment taking one symbol and taking all signals with proper filtration and producing a daily report showing the results. You will of course manage the trades using Steve's trading course parameters along with his money management ideas. All !, :) and whether you are a expert/advanced trader and should have not been surprised and anticipated the move...aside. Just a raw report showing net profits/losses on the day per symbol/periods. My thoughts are my own and don't reflect on any others and their efforts.....Jim

Steve Griffiths
07-23-2006, 03:18 PM
Hi Everybody,

A very interesting thread here................. may I add in a few comments that updates some of the points here, in particular how (as outlined in my daily report), we are now using the larger degree (15min) STF (strong trend filter) and the $TRIN, both of which did not feature in these original results. Not only that, but we are also using the new ATR (average true range) for the protective stops, particularly on days when a strong larger degree trend is in force.

All of this means that the % of winners should be higher, but the main change should be an increase in the size of the winning trades, particularly as we now have far move very larger (almost 10xR) winners now with the new ATR. As such, this will all help to increase profitability.

lastly, for those of you who have the patience and discipline to follow my lead of only trading when everything all comes together to paint the same picture (as outlined in my daily reports), you should have even better results.

As such, all of this should help improve these base results.

But, Yes, I agree with the jist of the first post that everybody must accept that a successful system will not have a high % of winners because it is the size of the winners (in relation to the losses) that is important - this is the strength of MTPredictor - its ability to keep the losses small when they unfold. This allows the profits (when they come through) to be large. This is what makes you the money over time............. There are just so many amateur traders that simply do not understand this vital concept of successful trading


Steve

Steve Griffiths
07-24-2006, 07:13 AM
Hi Everybody,

Friday was a pefect example of what I was saying in my last post. Here you can see how the new (larger degree) STF and $TRIN helped avoid the raw buy set-ups on Friday (which woudl have otherwise been loses). All of which helps to improved the basic set-ups when looked at just on their own :)

Thanks

Steve