The question i asked didnt have anything to do with the original thread( got a little side tracked ), the time frame i was thinking about was say 5,10 or 15 years, even more. PeeDee.
I think you'll find the StatArb'ers out there consider maths to be a pretty important part of their trading environment. As for be *the* key, well.... there is no *one* key, even the best black boxes have a bloke sitting there with a coffee and a danish checking the lights are on in the morning.
Just my £0.02....
Dave
Just my £0.02....
Dave
I have to thank somebody on the board for this gem that I had never come across before. I would like to give them credit for this but can't remember there name.
-----
The formula for probable consecutive losing (or winning) trades based on a known percentage of success is as follows:
C = Ln(X)/-Ln(Y)
Where C = Consecutive Losing trades
X = Number of trades made
Y = Losing percentage of trades
-----
I trialled this and backtested it on a number of positions and it is very accurate. I was suprised I had never used this before or thought of it myself!!
-----
The formula for probable consecutive losing (or winning) trades based on a known percentage of success is as follows:
C = Ln(X)/-Ln(Y)
Where C = Consecutive Losing trades
X = Number of trades made
Y = Losing percentage of trades
-----
I trialled this and backtested it on a number of positions and it is very accurate. I was suprised I had never used this before or thought of it myself!!
I should credit this one to Clem on ADVFN with whom, a couple of years ago, I was discussing my problem about time being the joker in the pack. He suddenly popped up a gem.
"The predicted range is the square root of the time series"
Uh? Well it goes like this.
If the average daily range is $0.76 then the predicted weekly range is $0.76*SQRT(5)
Unfortunately it is not as simple as that but you move a lot closer to the solution using this formula.
"The predicted range is the square root of the time series"
Uh? Well it goes like this.
If the average daily range is $0.76 then the predicted weekly range is $0.76*SQRT(5)
Unfortunately it is not as simple as that but you move a lot closer to the solution using this formula.
That's just how I entered it in Excel. I believe that Excel will calculate it in order of priority therefore the answer will not be negative.
Scripophilist,
It was me that posted the formula which I had primarily used for predicting horse racing wins and losses a few years ago but it is just as applicable to trading.
Glenn,
It doesnt give a negative answer for C see attached Excel spreadsheet which does it all for you.
Cheers
Paul
It was me that posted the formula which I had primarily used for predicting horse racing wins and losses a few years ago but it is just as applicable to trading.
Glenn,
It doesnt give a negative answer for C see attached Excel spreadsheet which does it all for you.
Cheers
Paul
Trader333, Thanks for spotting this thread. I give all acknowledgement to you!!
It's worth us having a chat at some point as my original background was gambling games before I realised there was a bigger, better and fairer Casino out there.
I am on MSN / YAHOO / AOL all the time my nickname is rather suprisingly "Scripophilist" or on AOL "ScripophilistUK"
It's worth us having a chat at some point as my original background was gambling games before I realised there was a bigger, better and fairer Casino out there.
I am on MSN / YAHOO / AOL all the time my nickname is rather suprisingly "Scripophilist" or on AOL "ScripophilistUK"
Trader333
Looking at your spreadsheet, there are two hidden columns, C and D. These must contain other formulae which give the values C2 and D2 from A2 and B2.
So the formula C = Ln(X)/-Ln(Y) is not the whole picture.
The use of -Ln(Y) suggests that Y is always less than 1 in order to give a positive answer, so Y clearly isn't either of the two input values.
Glenn
Looking at your spreadsheet, there are two hidden columns, C and D. These must contain other formulae which give the values C2 and D2 from A2 and B2.
So the formula C = Ln(X)/-Ln(Y) is not the whole picture.
The use of -Ln(Y) suggests that Y is always less than 1 in order to give a positive answer, so Y clearly isn't either of the two input values.
Glenn
Cols C&D merely (D) convert integer 70 into a percentage probability (i.e. 0-1), (C) work out NOT(d). Nothing special. It would be easy to use parenthesis to clarify the equation.
C=(ln(x)/(-ln
)
C=(ln(x)/(-ln
)
Scrip
"Cols C&D merely (D) convert integer 70 into a percentage probability (i.e. 0-1), "
So there is a formula missing to calculate that probability.
"(C) work out NOT(d). "
Don't undesrtand what you mean by this.
Using the Input values on the spreadsheet, x=400 and y=70,
The Natural Log of 400 is 5.99146
the negative Natural Log of 70 is -4.24849
Therefore C works out at -1.4102563, which is nonsense.
Incidentally, has anyone tried entering other figures e.g. 10 total and 5 winners ?
This gives 45 consecutive losing trades ! Interesting.
Glenn
"Cols C&D merely (D) convert integer 70 into a percentage probability (i.e. 0-1), "
So there is a formula missing to calculate that probability.
"(C) work out NOT(d). "
Don't undesrtand what you mean by this.
Using the Input values on the spreadsheet, x=400 and y=70,
The Natural Log of 400 is 5.99146
the negative Natural Log of 70 is -4.24849
Therefore C works out at -1.4102563, which is nonsense.
Incidentally, has anyone tried entering other figures e.g. 10 total and 5 winners ?
This gives 45 consecutive losing trades ! Interesting.
Glenn
Glenn,
If you highlight the whole spreadsheet and then adjust one column width a small amount then columns C and D will appear and are only used to convert percentage wins to a convertable formula between 0 and 1. The reason they are hidden is because there is no input to be placed in them and to prevent confusion for users and that is all.
As said previously the formula does not return a negative result because the - LN is always calculating a number between 0 and 1 which is a converted percentage. If it was greater than 1 then as you say it would give a negative number.
Paul
PS I take your point on only 10 trades. It is a while since I did this, in fact many years ago and I think it requires a minimum of 100 trades to be truly effective and also bear in mind that statistical analysis only works effectively on larger numbers. I found it useful for what I did and only put it on the boards to help people as I have nothing to sell.
If you highlight the whole spreadsheet and then adjust one column width a small amount then columns C and D will appear and are only used to convert percentage wins to a convertable formula between 0 and 1. The reason they are hidden is because there is no input to be placed in them and to prevent confusion for users and that is all.
As said previously the formula does not return a negative result because the - LN is always calculating a number between 0 and 1 which is a converted percentage. If it was greater than 1 then as you say it would give a negative number.
Paul
PS I take your point on only 10 trades. It is a while since I did this, in fact many years ago and I think it requires a minimum of 100 trades to be truly effective and also bear in mind that statistical analysis only works effectively on larger numbers. I found it useful for what I did and only put it on the boards to help people as I have nothing to sell.
Last edited:
All I was pointing out was that the formula published by Scrip was not the whole formula.
e.g. To calculate Consecutive Losing Trades, the formula is
C = Ln(A2) / -Ln((100-B2)/100)
"As said previously the formula does not return a negative result because the - LN is always calculating a number between 0 and 1 which is a converted percentage"
Erm, you didn't say this previously and anyway what you say is not the case. The reason for using -Ln is that Column D2 (1-C2) creates a negative result, which has to be reversed to positive in order to give a sensible answer.
And for Scrip, the Priority method of dealing with a formula does not work as you think for division, no matter where you put the brackets.
Without intent I've probably got up everyones noses now, but I think it's important to be clear on the facts. No offence intended.
Glenn
e.g. To calculate Consecutive Losing Trades, the formula is
C = Ln(A2) / -Ln((100-B2)/100)
"As said previously the formula does not return a negative result because the - LN is always calculating a number between 0 and 1 which is a converted percentage"
Erm, you didn't say this previously
and anyway what you say is not the case. The reason for using -Ln is that Column D2 (1-C2) creates a negative result, which has to be reversed to positive in order to give a sensible answer.And for Scrip, the Priority method of dealing with a formula does not work as you think for division, no matter where you put the brackets.
Without intent I've probably got up everyones noses now, but I think it's important to be clear on the facts. No offence intended.
Glenn
