Trading Myths and Some Forex Math | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Twelve widely believed forex myths 121 replies

Forex Myths 29 replies

Ridiculous Trading Myths 323 replies

Dispelling Some Myths on Trading Practices 23 replies

Trading Myths 35 replies

Attachments: Trading Myths and Some Forex Math
Exit Attachments

Trading Myths and Some Forex Math

  • #1
  • First Post: Edited Sep 26, 2018 2:49am Aug 25, 2018 1:03pm | Edited Sep 26, 2018 2:49am
I'd like to use this thread to discuss some trading myths and forex maths/statistics that will destroy some concepts that "the gurus" preach day in and day out. Please discuss these myths here with us and back up your words with evidence.

The two topics I willl start with are

  1. the myth of risk:reward ratios and
  2. trading with moving averages.

Happy discussion!

= TOC ========================================================

Trading myths:

  1. The myth of risk:reward ratios
  2. Further myths (VEEFX)
  3. Every indicator is unique: Use multiple indicators to complement with each other

FX math:

  1. How does an average candlestick look like?
  2. The relation between price and volume

General view:

  1. Binary nature of Currency Markets (VEEFX)

=============================================================

  • #2
  • Edited 2:47pm Aug 25, 2018 2:07pm | Edited 2:47pm

=== The myth of risk:reward ratios ===


I guess everybody who starts studying the markets may have heard one platitude:

"Only open trades with high rewards and low risks!"


What does it mean? It means that you should only enter a trade where your predefined prize target is more distant to your entry than your exit level. Sounds great, doesn't it? I mean, you will always win more than you lose. Not bad?!

Hell no! Let's analyze it.

Although this may be not the full truth, let's consider that the market is random, i.e. you have no idea on where/when to enter and which direction your trade should have (buy or sell). Thus, we consider that at each point in time there will be always a 50 % chance that the next prize will be above or below our current prize (or that there is an equal chance for the next candlestick to be either green or red). In such a condition prize can be thought of as a random walk. And as for random walks like this there's always and equal chance to either reach point A above the current prize or reaching point B below the current prize as long as A and B are equi-distant to our current prize.

However, the risk:reward platitude suggests to use a high reward to risk ratio (RR), e.g. 2:1, 3:1, etc. We will now just observe what happens if we use a 2:1 RR. Nevertheless, the same logic is applicable to other RRs, too.

Below you can see a picture of a potential buy trade you could enter:

Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 15 KB


The left part, a), shows your setup. Assume that you'd enter at the open of the candle marked with an arrow. The bold horizontal (black) lines show your entry prize, the exit prize SL, and the target prize TP (which is twice as distant from the entry then the exit). Additionally I've drawn a thin horizontal line (i) which represents an intermediate line which is as distant from the open prize than SL.

Since we now know these levels that lie inbetween the next higher/lower level, we can model our trade in a discrete manner: Every time the prize hits/breaks one line, it has a 50 % chance of reaching the next higher or next lower level. This is shown in b).

When we enter the trade we have an equal chance of either reaching i or SL. If we have reached SL our trade was lost, if the prize has reached i it could either go up to TP (our target) or retrace back to the entry level. Each and every situation that can happen as long as our trade survives can be seen in the decision tree b). Once we know the prize history, we can calculate its probability by multiplying all (0.5) probabilities along the edges from our entry (uppermost point in the decision tree) towards the outcome.

We have already observed that (with a 2:1 reward to risk ratio) we have a chance of 50 % (= 0.5) for losing the trade (prize goes straight forward to the SL level). However, our first time we could win (prize reaches TP without retracing towards a lower level; the only "TP" in the tree) has a probabilty of 0.5 * 0.5 = 0.25 = 25 %. Although there are further possible scenarios (e.g. reaching SL or TP after retracing a few times) the ratio of the probabilities for a win/loss won't change anymore.

What have we learned so far? (in regards of our 2:1 RR setup?)

  1. With a RR of 2:1 you can win 2 units of money while you only risk 1 unit of your money each trade.
  2. With a RR of 2:1 you will lose twice (2 times) as much as you will win.

So, there seems to be an inherent trade-off: If you change the RR from 1:1 to x:1 (where x > 1) you will win more when you win but you will lose more often. A high RR will never give you an advantage (or an edge) in trading!

And from this considerations we can also see that a simple grid trading strategy where you buy/sell/close your trade everytime prize crosses a level will never be a successful system alone.

But the big question is:

How can we overcome the trade-off between the won amount (pips/money) and the winning percentage?

Is it only possible with an edge that shifts the winning probability in our favor?

Any answers/discussions/contributions are very welcome.

  • #3
  • Aug 25, 2018 2:35pm Aug 25, 2018 2:35pm
It seems you are only considering the risk amount, rather than also taking into account how risky the trade in itself actually is.
  • #4
  • Aug 25, 2018 2:43pm Aug 25, 2018 2:43pm
Quoting Bushy
rather than also taking into account how risky the trade in itself actually is.
Maybe I don't get what you mean, But I guess that this is already incorporated in the assumption of the random market, isn't it?
  • #5
  • Aug 25, 2018 3:22pm Aug 25, 2018 3:22pm
The market isn't random.
  • #6
  • Aug 25, 2018 3:29pm Aug 25, 2018 3:29pm
Quoting AntiCre
I'd like to use this thread to discuss some trading myths and forex maths/statistics that will destroy some concepts that "the gurus" preach day in and day out. Please discuss these myths here with us and back up your words with evidence. The two topics I willl start with are the myth of risk:reward ratios and trading with moving averages. Happy discussion!
ma's, bb breakouts and other concepts that have "the trend" and or a "drift" works, even if a trend can determinated, defined after the fact.
i'm not so smart and i'm not in statistics, in fact i;m very dump in this area and math.
Attached File(s)
File Type: pdf portofolio.pdf   911 KB | 1,225 downloads
Price itself it's an indicator
  • #7
  • Aug 25, 2018 7:03pm Aug 25, 2018 7:03pm
Quoting Bushy
The market isn't random.
I think so, too. Otherwise nobody would make money (imho). However, I've never seen a explanatory proof (thus, the assumption of 50 % which may be even true most of the time although sometimes the market can be more deterministic). Maybe you may show one.
  • #8
  • Aug 25, 2018 7:35pm Aug 25, 2018 7:35pm
It is a myth limited ONLY to your own personal understanding on the merits of RRR and the forex PA process. It's akin to saying indicators don't work becos we don't understand how it work and know how to make it work.
Hehe
  • #9
  • Aug 26, 2018 3:50am Aug 26, 2018 3:50am
Quoting genghistar
It is a myth limited ONLY to your own personal understanding on the merits of RRR and the forex PA process. It's akin to saying indicators don't work becos we don't understand how it work and know how to make it work. Hehe
Well, I haven't talked about indicators - just RR. And yes, you're correct, I can't see any merits in RRR alone. I mean, it's obvious that my RRR interpretation will be different if you have a strong edge (I even mentioned it in the 2nd post). But maybe you may shed some more light on RRR and PA ...
  • #10
  • Aug 26, 2018 3:54am Aug 26, 2018 3:54am
Quoting m3rlin
{quote} ma's, bb breakouts and other concepts that have "the trend" and or a "drift" works, even if a trend can determinated, defined after the fact. i'm not so smart and i'm not in statistics, in fact i;m very dump in this area and math. {image}
Hi m3erlin,

I don't get that bold part. Do you mean that it works without knowing the trend while you trade? How? And what do the backtests exactly show (general outline of the strategy/with(out) MM/indicators)?
  • #11
  • Aug 26, 2018 4:11am Aug 26, 2018 4:11am
Quoting AntiCre
{quote} I think so, too. Otherwise nobody would make money (imho). However, I've never seen a explanatory proof (thus, the assumption of 50 % which may be even true most of the time although sometimes the market can be more deterministic). Maybe you may show one.

Chart snapshot... random?
Attached Image
  • #12
  • Aug 26, 2018 4:31am Aug 26, 2018 4:31am
Quoting Bushy
{quote} Chart snapshot... random? {image}
I don't know the answer. Do you?

Counter question: Here is a tick chart. Random or not? (*)

Attached Image


----

(*) Have you answered "not random"? Be honest! I have to admit that this chart has been created using a normal distribution and thus it is random. Trends can be generated by random processes, too.
  • #13
  • Aug 26, 2018 4:47am Aug 26, 2018 4:47am
I'd say they are definitely not random, they head in a clear direction.
  • #14
  • Aug 26, 2018 4:52am Aug 26, 2018 4:52am
Quoting Bushy
I'd say they are definitely not random, they head in a clear direction.
That's not a proof. Open Excel, generate random numbers and calculate the cumulative sum. Then plot it. You will find long trends within your random data!

Please don't misunderstand me: I'm sure some edges can be find within FX data. But maybe these edges can even be found under a random regime. I don't know it and nobody could ever proof to me that market isn't random. With random I mean that nobody can really predict the future prize path with 100 % accuracy because nobody knows the intentions of all market's participants.
  • #15
  • Aug 26, 2018 4:55am Aug 26, 2018 4:55am
Quoting AntiCre
{quote} Hi m3erlin, I don't get that bold part. Do you mean that it works without knowing the trend while you trade? How? And what do the backtests exactly show (general outline of the strategy/with(out) MM/indicators)?
a trend can be determinated after the fact. some trends are false trends (even if they persist for 1 year or more) they don't have fundamental, economical and geopolitic basic (the trend in that asset is made only by people behaviour: they only speculate and make decissions and they exprim they decissions taking long and shorts possitions) this process can be found in every market from currencies to stock market.

for example: if a company is listed on stock market even if the company is not profitable the price of shares can go up and up and because the stock price increase will atract new buyers (from small retail traders to some hedge funds and others) of course others speculators will go short but the thing is that beccause the stock is accting so good the company can borrow money at cheap rates and in fact the stock will go higher and after that the company can borrow even more cheap that before, is a self reinforcing proccess that can go on and on for a very long time creating trends and in some cases "bubbles" that in the end will blow up slowly or fast.

the pdf in my post is a long term bb breakout system applied to daily data on dif. asset classes like stocks, commodities and forex. (it was designed to test the hypothesis that publicly know approches are profitable in long term if proper risk and protfolio rules are build)
Price itself it's an indicator
  • #16
  • Aug 26, 2018 5:05am Aug 26, 2018 5:05am
Quoting AntiCre
{quote}With random I mean that nobody can really predict the future prize path with 100 % accuracy because nobody knows the intentions of all market's participants.
That my friend is the art of trading
  • #17
  • Aug 26, 2018 5:22am Aug 26, 2018 5:22am
Quoting AntiCre
{quote} That's not a proof. Open Excel, generate random numbers and calculate the cumulative sum. Then plot it. You will find long trends within your random data! Please don't misunderstand me: I'm sure some edges can be find within FX data. But maybe these edges can even be found under a random regime. I don't know it and nobody could ever proof to me that market isn't random. With random I mean that nobody can really predict the future prize path with 100 % accuracy because nobody knows the intentions of all market's participants.
i will not debate if markets are random or not, if they are efficent or not what i'm saying is that in market is some decision making process (good or bad, based on good information or not) by people or algoritimc one and this in effect can shape the form of things and when the thing have a new form, a different one same process start again.

another thing: speculation, trading is discretionary (even at funds like Renaissance they first started to buid/design profitable system/approches to trade markets and even if they deisgned 100 profitable systems in the end the people will select just a number of them to use them in they fund)
hope make sense.

ps: my market knowledges and of the world in general, are very limited and in most of the time they are false and wrong so better don't height to much what i'm writing.
Price itself it's an indicator
  • #18
  • Aug 26, 2018 6:03am Aug 26, 2018 6:03am
Quoting m3rlin
ps: my market knowledges and of the world in general, are very limited and in most of the time they are false and wrong so better don't height to much what i'm writing.
That's true for me, too. Just like to present what I've seen/found so far. Maybe it may be helpful for anyone.
  • #19
  • Aug 26, 2018 6:41am Aug 26, 2018 6:41am
Quoting AntiCre
{quote} That's not a proof. Open Excel, generate random numbers and calculate the cumulative sum. Then plot it. You will find long trends within your random data! Please don't misunderstand me: I'm sure some edges can be find within FX data. But maybe these edges can even be found under a random regime. I don't know it and nobody could ever proof to me that market isn't random. With random I mean that nobody can really predict the future prize path with 100 % accuracy because nobody knows the intentions of all market's participants.
Dear, it is possible to predict with 100% accuracy, but it may not be in all 100 cases.
  • #20
  • Aug 26, 2018 7:24am Aug 26, 2018 7:24am
Quoting sol4live
{quote} Dear, it is possible to predict with 100% accuracy, but it may not be in all 100 cases.
Evidence, please.
0 traders viewing now