The Dumbest Question of All Time?

S

skippy

Ok, here's what I don't understand:

Price can do one of two things -- go up or go down. Which means we have a 50:50 chance of being right (or wrong) when we guess where price is going to go next (right?).

So, why not a trading strategy based on random entries, with some multiple of stop price to target price such that a 1:3 or 1:4 risk:reward ration is assured.

Yes, I am quite certain that this won't work. I just don't understand "WHY" it won't work.

Figuring that it's best to get what I can only hope will be my most idiotic post out of the way at the get-go, so that it's all uphill from here.

Thanks
 
if your risk reward ration was 1 to 3. There is more of a chance that the stop will be hit more times than the profit target.

To put it simply. You will gain more on your winners, but your winning % will be much lower, and then out of all the randomness, it will be determined if you win or not.

1 to 3 risk/reward, your system would be right about 33% of the time (not 100% sure), so I somehow believe it will all even itself out.


What if instead, you had the same random entry, but you put rules to that , to scale out of the trade once it goes in your favor and set a break-even once it comes back to your entry? Could change things significantly, even if you set a stop at break-even or one tick above break-even or something, I'd assume it would dramatically improve the success of the system immediately.

I believe managing the trade itself, is far more important then the perfect entry. Although I aim to have great entries because of my style. When I go to a longer term trading style, managing the trade is more important than where I got in at.

It's an interesting topic though, I'm sure others will have more proper ideas.


There was a very interesting trade expectancy type program that drew some great things with parameters, it could possibly help if someone knew the link.


EDIT:
http://www.agribiz.com/ngfa/CoinToss/cointossdist.html
There is a random coinflip type deal, but it wasn't what I was specifically looking for. Will try to find the other thing.
 
Using random entry... changing the size of your target doesn't make a difference.

sell target = stop loss
50% 1R gain
50% 1R loss

sell target = 2 * stop loss
33% 2R gain
66% 1R loss

sell target = 5 * stop loss
20% 5R gain
80% 1R loss

You should be able to answer your own question now.
 
Quote from jsmith:

use random entry...

sell target = stop loss
50% 1R gain
50% 1R loss

sell target = 2 * stop loss
33% 2R gain
66% 1R loss

sell target = 5 * stop loss
20% 5R gain
80% 1R loss

You should be able to answer your own question now.
More...

AKA, trade management :D

EDIT:

I found the proper link:
http://www.hquotes.com/tradehard/simulator.html

I thought it was interesting. Be even nicer to simulate what jsmith suggested.
 
I believe I have attached the correct settings, if your average winner was three times your average loser (1 to 3 risk/reward), and your system was right 30% of the time, which could be tough with being just random.


You add in proper trade management, a stop at break-even when the trade is good, and the ability to scale out of the trade, who knows what is possible?

I'm sure there are computerized systems that already do this.

I'd be interested in knowing if any trader on ET uses a completely random entry for their trade, anyone here trade this way?

How about factoring in commissions, slippage, fees, and all that as well, could make it tough.
 

Attachments

Correction to my previous post:

sell target = 5 * stop loss
20% 5R gain
80% 1R loss

Should be

sell target = 4 * stop loss
20% 4R gain
80% 1R loss

Still 0 expectancy. Without an edge, you're better off not trading.
 
skippy,

>Price can do one of two things -- go up or go down. Which means we have a 50:50 chance of being right (or wrong) when we guess where price is going to go next (right?).


Is it really just a 50/50 chance? As I don't imagine that you are
talking about the direction of the next tick after your execution, I think you could say that the chances are determined by the direction of the trend and the point within that trend at which you enter a position. Also, the length of time that you hold the position enters into play. While you may not know it at the time,
if you were to buy at or near the bottom of the next to the last
reaction in an uptrend, the chances of an advance over a given
period are, of course, 100/0.
In short, your chances of making money are, of course, greatest
if the probable direction of the price movement can be determined
through your analysis of the current trend. And if you do your
homework, I can assure you that you can arrive at a better
than 50/50 guess.

I haven't used English much over the last 40 years so what I
want to say may not be clear. But if you can find any meaning
there, I hope it helps.

Good luck.
 
1. There's commission / cost of trading.

2. Law of Large numbers. Given that you have a breakeven strategy. Can you withstand the drawdown / streak of losses...

3. Tests. How reliable is your test? Is the dataset large enough? Did you test it realistically? etc. etc.

4. Start flipping coins...
 
Quote from skippy:

Ok, here's what I don't understand:

Price can do one of two things -- go up or go down. Which means we have a 50:50 chance of being right (or wrong) when we guess where price is going to go next (right?).

So, why not a trading strategy based on random entries, with some multiple of stop price to target price such that a 1:3 or 1:4 risk:reward ration is assured.

Yes, I am quite certain that this won't work. I just don't understand "WHY" it won't work.

Figuring that it's best to get what I can only hope will be my most idiotic post out of the way at the get-go, so that it's all uphill from here.

Thanks
More...
simply put, if u go for random entries and your SL/TP is set for a 1:3 RR ratio, assuming randomness of price movements u'll hit yr SL 3 times more than your TP, hence an expectancy of 0 (less spreads, comms etc)... and u can try all the refinements u want, scale in if price in direction etc... still 0...
 
You must log in or register to reply here.
Back
Top