Where am i missing it? :(

[momothebored]

Hello

errrmmm... having some problems understanding this.

Fantastically busy these few days, so intending to start paper trading FX etc later in the week. New to FX trading, sorry if this is a silly question.

Assumptions:
Trading capital: $100,000
Methodology:
% of equity risked per trade: 2%
Minimum Reward / Risk: 2:1 (4%: 2%)
Accuracy: 50%

Let's say i trade .. 15 minute charts (random number) for 8 hours a day.
In any given day, I get 10 opportunities.


Of these 10 trades,
5 work out (50%) accuracy --> I make 4% x 5 trades x $100k= $20,000
5 end badly and trigger stop losses --> I lose 2% x 5 trades x $100k= $10,000
Net: +$10,000.. (?)


How is this possible?
The statistics of the setups and methods I put forth really aren't that difficult right?

You can't possibly be making +10% / day with a mediocre model, right?

 

[momothebored]

Please could someone help me understand this better. What am i not seeing that my estimates are way too optimistic?

Brain...itching!!

 

[Pat Riley]

Your maths is spot on. Go ahead and trade it.

 

[momothebored]

Your maths is spot on. Go ahead and trade it.

...how can that be though?

Then anyone with a lousy model like that should have made trillions of dollars after a few years :-0

warren buffett would be working at mcdonald's with his silly ROI, everybody and their dogs, cats and grandmother would all be trading and the world as we know it wouldn't exist :|

 

[Pat Riley]

I thought you were sayin that was your system? If it was just a hypothetical then there's a good reason for that.

 

[momothebored]

I thought you were sayin that was your system? If it was just a hypothetical then there's a good reason for that.

It's 100% hypothetical.. i've never made a single forex trade.. starting when i have more time end of the week.

tell me pleeeeease what's the reason?

i don't see what's so difficult about coming up with a 50% accurate model with 2:1.

 

[Pat Riley]

i don't see what's so difficult about coming up with a 50% accurate model with 2:1.

i've never made a single forex trade..

tell me pleeeeease what's the reason?

I think you've answered your own question.

Good luck to ya.

 

[momothebored]

I think you've answered your own question.

Good luck to ya.

Right. :cheesy:

 

[Pat Riley]

I'm not a forex trader nor am I a technical boy so I shouldn't have engaged with ya, but for example, take any grungy pair ya fancy. At random, nz/cad. It spent yesterday goin up a bit. It's due for a bit of breather I reckon. If you stick in a sell at market now round about 8870 with a stop at 8895 and a target at 8820, that would meet the 2:1 you're on about. I'd suggest you need to consider treating yourself to a target around the 8795 level if it's just intraday you're thinking of. That's a 3:1. But just watch what happens to that hypothetical trade. Then do that a few thousand times more and you'll get the picture I'm sure.

 

[momothebored]

I'm not a forex trader nor am I a technical boy so I shouldn't have engaged with ya, but for example, take any grungy pair ya fancy. At random, nz/cad. It spent yesterday goin up a bit. It's due for a bit of breather I reckon. If you stick in a sell at market now round about 8870 with a stop at 8895 and a target at 8820, that would meet the 2:1 you're on about. I'd suggest you need to consider treating yourself to a target around the 8795 level if it's just intraday you're thinking of. That's a 3:1. But just watch what happens to that hypothetical trade. Then do that a few thousand times more and you'll get the picture I'm sure.

I'm really not sure why i'd do that (not looking at the chart) unless it matched my setups.

You're quoting numbers that i may / may not agree with and in the end it would come down to the quality of our setups.

It sounds like you're trying very hard to articulate that 50% accuracy is unachievable?

Thanks

 

[Pat Riley]

I'm really not sure why i'd do that (not looking at the chart) unless it matched my setups.

You're quoting numbers that i may / may not agree with and in the end it would come down to the quality of our setups.

It sounds like you're trying very hard to articulate that 50% accuracy is unachievable?

I'm not what you'd call a directional trader so you should dismiss everything I say on the subject, but you're gettin fixated on accuracy. Without a clear definition o that you're in a bit of a mess. Having rigid targets via a vis yer stop is one way to go I suppose, but you'll be made of unusual stuff if once ya have a trade in play you don't start moving your stop in, or out, and moving your target in closer. All kinds of things need to considered on-the-fly and you need a game plan for every eventuality. There aren't too many fire and forget methods that'll give ya anything like the performance you think, hypothetically, should be a cinch.

 

[momothebored]

I'm not what you'd call a directional trader so you should dismiss everything I say on the subject, but you're gettin fixated on accuracy. Without a clear definition o that you're in a bit of a mess. Having rigid targets via a vis yer stop is one way to go I suppose, but you'll be made of unusual stuff if once ya have a trade in play you don't start moving your stop in, or out, and moving your target in closer. All kinds of things need to considered on-the-fly and you need a game plan for every eventuality. There aren't too many fire and forget methods that'll give ya anything like the performance you think, hypothetically, should be a cinch.

Ok thanks, makes sense.

i'll likely be sitting there staring at it anyway for a while.

 

[NVP]

Please could someone help me understand this better. What am i not seeing that my estimates are way too optimistic?

Brain...itching!!

maths is right.....but never underestimate the capacity for human screw-ups................that's what we humans do .....in all things in life.........

and trading is no exception
N

 

[billyjean]

2% risk per trade with 100k.
a random number entry, with 2:1 ratio, assumed accuracy 50%.
well, answer is clear . the main key problem is 'randomize' order.
how is it possible a random act resulted an accuracy of such number, defined 50% ???
logically it will produce a random result also.
what i see here a short term trading period with high risk percentage per trade.
2% loss probability of random act, at 50 times consecuentive loss the account would be over, is that what you think ?
is not works that way. let put in on number (this number despise any trading rule such, leverage, stop out leve, spread, commision, swap, etc.)
100.000 - 2% loss per-trade means 2000 usd. assume you put 2 lot per trade, means your stop loss are 100 pips.
at 5 consecuentive loss, we have 2000x5 =10.000 usd loss.
current equity = 90.000 usd.
yet you still used the same 2% lot size based on initial deposit, the same 2 lot for current equity 90k.
which means your next used lot are not 2% risk anymore, it get more higher risk.
how bout when you've resulted at 10 consecuentive losing trade ?
current equity at 80.000 usd, would you still used the same 2% based on 100k.
2 lot size per trade for 80k equity. would your random entry, gonna resulted 50% accuracy.

what you missed here, underestimate destructive power of losing trade.
here's a quick example, when we losing balance for amount 50%, it means a 100% profit trade need to get initial balance returned. 100k usd, loss at 50% = 50k usd, using this current equity deman 50k profit to get your total balance break even, a 100% profit!

do you see clearly now ?

 

[momothebored]

1. The 50%, or whatever number you want to substitute here is simply based on how it's performed over a long-enough time line. If you're looking for a logical basis *why* it does so, you'll never find one. If that isn't good enough for you, then a good chunk of Quant methodologies like estimate revision and the like, based on historical observations but accepted by academia and funds alike are automatically disqualified.

So, for the purpose of this question, please just accept or imagine that I have a method back-tested on a long-enough timeline to accept it works at 50% historically on the particular security.
In fact, throw all technical methods out if you're looking for a "why".


2. I'm prepared to lose 50 times in a row. The theoretical amount in question *is* just risk capital. So i'm not sure what your point is.

I'm sure statistically someone out there will be the (0.5)^50 guy to lose 50 times in a row, but then there would be the P(x)=(0.5)^50 guy who wins 50 times in a row.

Assume a normal distribution and that I can easily survive long enough for mean results of the model to come out.

"Do I see clearly now?"

Not really..
Sorry, your points aren't relevant at all.

2% risk per trade with 100k.
a random number entry, with 2:1 ratio, assumed accuracy 50%.
well, answer is clear . the main key problem is 'randomize' order.
how is it possible a random act resulted an accuracy of such number, defined 50% ???
logically it will produce a random result also.
what i see here a short term trading period with high risk percentage per trade.
2% loss probability of random act, at 50 times consecuentive loss the account would be over, is that what you think ?
is not works that way. let put in on number (this number despise any trading rule such, leverage, stop out leve, spread, commision, swap, etc.)
100.000 - 2% loss per-trade means 2000 usd. assume you put 2 lot per trade, means your stop loss are 100 pips.
at 5 consecuentive loss, we have 2000x5 =10.000 usd loss.
current equity = 90.000 usd.
yet you still used the same 2% lot size based on initial deposit, the same 2 lot for current equity 90k.
which means your next used lot are not 2% risk anymore, it get more higher risk.
how bout when you've resulted at 10 consecuentive losing trade ?
current equity at 80.000 usd, would you still used the same 2% based on 100k.
2 lot size per trade for 80k equity. would your random entry, gonna resulted 50% accuracy.

what you missed here, underestimate destructive power of losing trade.
here's a quick example, when we losing balance for amount 50%, it means a 100% profit trade need to get initial balance returned. 100k usd, loss at 50% = 50k usd, using this current equity deman 50k profit to get your total balance break even, a 100% profit!

do you see clearly now ?

 

[sunwest3d]

Also agrees, your calculations are correct...
You said ... "You can't possibly be making +10% / day with a mediocre model, right?"... and then ... "Then anyone with a lousy model"...

Being right 50% of the time consistently on Reward to Risk ratio of 2:1 is actually pretty good, if you think that is a mediocre or lousy model then think again.

It is a bit like saying at the game of head or tails (50/50): "Every time I decide to make a guess (enter the market) for head or tails it comes out twice in a row 50% of the time"... now try the experiment and tell me how you are doing after 100 attempts (Your guess must come out twice in a row or it is a fail)