Dxtrade,

Glad you enjoyed your halloween party

The EA is not as exceptional as shown by your latest results..we are now in a very high volatility environment with streaks that allow for either the 355 pips target to be achieved or to be stopped out by Chandelier at a profit..

If you are willing to continue with your test you will see that when we get enough trades the EA results are going to be similar to those of the out of sample tests...

__________________
Somos mortales hasta el primer beso..y la segunda copa de vino

November 1st 2007

EA trade on FXDD live account.
Settings are here: http://www.forex-tsd.com/158338-post1047.html

2007.11.01 Buy GBP/JPY 239.83
2007.11.01 Closed @ 239.22 With -61

Total since october 17th, 2007: +365

__________________
Happy Trading...

"Failure is only a temporary change in direction to set you straight for your next success."
"Winning is a habit. Unfortunately, so is losing."
"It’s so simple to be wise. Just think of something stupid to say and then don’t say it."

"Say what you mean, but don't say it mean!"

Hello Simba,

I apologize for breaking this thread, but I'm a bit stuck.
This is in reference to a post much earlier in this thread regarding money management, the Kelly ratio, "right risk", etc..

I've calculated my Kelly ratio and my worst case drawdown.
I'm assuming an initial account of $1000.

I can't figure the correct "right risk"; your post uses a multiplier of 1000 and I don't know where that comes from.

I'm also not sure if the equation is:

K*(Equity/Worst DD)* multiplier
or
[(K*Equity)/Worst DD]*multiplier

Can you please indicate the correct multiplier and the proper equation to use?
Any help appreciated.

Thanks.

Ammo,

You don`t break the thread,you just add to it

I have posted several references to Kelly formula in this and other Forums,and ,so ,I am going to ask you 2 things:

1-Could you reference my post here ?I remember that I was arguing(more or less ) that the Kelly ratio is unfeasible in real trading,so there were several ways to tame it,one was to divide it by (worst loss/average loss)..and the second one to divide it by your (worst drawdown/average loss)..since it is always the worst drawdown from a consecutive series of losses that kills the account

2-What are the numbers you used to calculate your Kelly ratio?Please post them here and I will give you my opinion and,hopefully,help you

Regards

__________________
Somos mortales hasta el primer beso..y la segunda copa de vino

Thanks for your prompt reply.

Rather than try to find the post, I'll reprint a portion of it here (I saved a copy):
-----------------------------------------------------------------------
1-Kelly ratio:Your optimal bet size..Kelly % = W – [(1 – W) / R] ,Where:
W = Winning probability and R = Win/loss ratio ..so for a losing system ,or even a neutral system(ie:50% wins and 1 win loss ratio)the answer will be..

Kelly % =0.50-[(1 – 0.50) / 1]=0.50-0.50=0.00..DON`T BET,DON`T TRADE

And for a winning system,let`s say 60% wins and 2.0 win loss ratio the answer will be..

Kelly % =0.60-[(1 – 0.60) / 2]=0.6-0.2=0.4..so you "should " bet 40% of your Equity and this will get you the optimal return..Right?

Not so,you need another concept.

2-Risk of Ruin:Life is hard and then you die,and trading is a specially hard part of Life if you don`t get it right..so ,you should pay special attention to this concept..Theoretically risk of ruin is the probabilty of halving your account before doubling it,just for your info,kelly ratio has a risk of ruin of 33%..but the important thing,the concept you need to understand is more basic,and more useful

Basic definition:An example is worth a thousand words..Say We are trading on the outcome of a coin being tossed.It is a fair coin.That is 50% of the time it comes up heads, 50% of the time tails.You are trading on heads.

And I am very kind, I will pay you out twice your stake(trade) if you win, and keep your stake(trade) when you lose....You can still lose, Imagine your total capital was $100.And you bet it all on the first toss of the coin...And it came up tails...so your erroneous trade size destroyed your account on a very profitable trading system

But no-one would be so stupid to risk it all on one trade..or yes?. Right,on both counts ...and that is the point...How do you know what is the right position size – the maximum amount to trade each time?Let us look at how likely a run of tails (bad news for you) is, for say, 4 tails in a row..0.5*0.5*0.5*0.5*=0.0625..6.25%..so,if you risk 1/4 of your total cumulative equity in each coin toss your expected dismissal will be in 100/6.25=16 trades..so 1 in each group of 16 coin tosses you can expect to find a cluster of 4 consecutive tails ..again,you had a profitable system,and you ended up destroying your account..you get it ..Right?Then how to integrate the 2 concepts,in a practical manner..

3-Practical "RIGHT Answer":100k equity,60%W,R=2...Remember..your Kelly% was 0.4..Now..
A)Test your system on at least 5 years of data,if mechanical,and see the maximal drawdown and the % drawdown for 1 lot(or 0.1,0.001..your standard position size)..
For example let`s say the maximal dd was 10k and the relative dd was only 5%(because at that moment your cumulative equity was 200k from your starting 100k)..
TAKE THE HIGHEST OF THE 2..why,because you have a system that can cumulatively lose 10k on 1 lot,the 5%(instead of 10%) was due to having started in a "lucky" period..
B)If your system is manual check your trade statistics and obtain your cumulative dd,if not enough history,then do a Montecarlo simulation with your trade parameters and take the worst "Montecarlo drawdown"

let`s presume your case is A..then you do the following formula..

RIGHT RISK=(KELLY Ratio*Equity/worst dd)*1000$=(0.4*100k/10k)*1000$=4k$..this the risk you can allow in your trades that will guarantee both near maximal profitability and very long survival probability..so..the Right answer to how many lots you should short will depend on were you would locate your stop or exit the trade..imagine stops at 34 pips and 6 pips slippage..so your "1 lot risk"would be 40 pips=400$..then your position size should be

RIGHT POSITION=RIGHT RISK/1 LOT RISK=4K$/400$=10 LOTS
-----------------------------------------------------------------------

Numbers I used for calculations:
(from manual backtesting)

Total wins: 240
Total losses: 94
Average win: .0047
Average loss: .0024
Worst drawdown (from highest EQ point to lowest): .0137
$1000 account

My calculations:
Winning probability: 71%
W/L Ratio: 1.95
K ratio: .56

Thanks again for your help.

QUOTE "Numbers I used for calculations:
(from manual backtesting)

Total wins: 240
Total losses: 94
Average win: .0047
Average loss: .0024
Worst drawdown (from highest EQ point to lowest): .0137
$1000 account

My calculations:
Winning probability: 71%
W/L Ratio: 1.95
K ratio: .56

Thanks again for your help."

Hi Ammo,

1-Kelly ratio:Your "optimal" bet size..Kelly % = W – [(1 – W) / R]=0.71-(0.29/1.95)=0.56 as calculated by you is ok

2-worst drawdown/average loss=0.0137/0.0024=5.71

3-safekelly=0.56/5.71=0.098...so,you could risk a max of 98$ for the initial trade with an equity of 1k$..and,as long as you keep on the same track(9.8% max risk per trade,and,some kind of results)that`s ok.

4-MORE IMPORTANT:Your numbers are too good,and they are from manual testing,I am not disputing them,I just advice you to recheck them,manual testing is prone to many biases when estimating the results of a system..as they are ... the probability of having 5 consecutive losses is 0.29*0.29*0.29*0.29 *0.29=0.00205..so,you can expect,on average,to have 5 consecutive losses every 488 trades..and you can expect that this will set you back to 0.902*0.902*0.902 *0.902*0.902=0.597...59.7% of the equity before the streak...can you stand it,an expected 40% drawdown every 488 trades?Depends on your trading frequency and your tolerance of risk...ANYWAY..your numbers are too good...71%Winning trades and 1.95 average win/average loss is exceptional..please recheck them.

Regards..and hope this helped

__________________
Somos mortales hasta el primer beso..y la segunda copa de vino

Thanks much for your help.

I agree, it looks awfully good, maybe too good.
It's an extremely simple system so I think results are a close approximation, I also verified them in Excel.

Forward testing will tell.

Your point #4 puts things into perspective.

Thanks again.

Hi Santiago,thanks for your help with the break-out strategies,They have ben unbelievably profitable..I would like to comment one sad issue I had with Saxobank,They issued a requote 8 hours after the fact,the trade was closed and confirmed 8 hours before their PHONE CALL asking me to accept the requote or else they will nullify the trade,since their requote(35 points in Spanish index Ibex) was only for 70 Euros,after haggling with them I accepted,but I feel scammed by them..what do you think?Should I pursue the issue legally?After their requote,instead of winning 415 Euros I won 345,but I feel like they took advantage of me

Hi Faust,

Actually this is higly unusual,a requote a few seconds after the order is issued is "acceptable",as long as it is not the usual thing..a requote of a "8 hours ago closed trade" is unbelievable..can you post your "fill and time" records?35 points in Spanish IBEX is a LOT..like having a 30 pips requote on EURJPY..I will not accept them,or,if forced,I would change broker ASAP..In any case,since Saxobank is actually a regulated European bank,I would contact their customer service department and,if no solution was found,I will issue a complaint against them at the appropiate regulatory authority..

Additionally,feel free to comment on this issue here or by PM,I will do my best to help you...if I can

__________________
Somos mortales hasta el primer beso..y la segunda copa de vino

Thank you for your reply Santiago. This is what happened: At 09.03 spanish time(CET) I sold two CFD of Ibex index that I have bought yesterday. The sell was confirmed at 15937.5 .Then,after around 6 hours,+-to them 15:50 spanish time I receive Saxobank's call, saying that there has been a mistake in the information of the ibex, and that are going to change the sale from 15937.5 to 15902, and that or he(she) was accepting or they were cancelling the trade

Attached Images

Ibex Saxo.GIF (75.2 KB, 679 views)