[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: when a strategy breaks the max drawdown/ ian



PureBytes Links

Trading Reference Links

> I agree with your "then add 30%-50%." 
> I have heard this from several other  traders as well.

This is not the right approach and will almost certainly lead to busting
the account eventually. The single DD figure from a system report is
worthless and X * worthless = WORTHLESS. 

Gary and Alex have given pieces of the right answer.

Gary: adjust the code for all backtested trades so you are risking the
same dollar amount on each trade. If you are using volatility to set the
size of your disaster stop (recommended), then contracts(shares) =
XX/volatility.

Alex: Once you have the equal dollar risk trades (raw P/L numbers for
each trade) you run them through the Monte Carlo sim. You simply shuffle
the trades into a different order and calculate the drawdown. Repeat
10,000 times. What you will end up with is two columns of numbers
similar to this.

Percent of account risked on each trade
	Probability of going broke
AA%	.1%
BB%	1%
CC%	10%
DD%	50%
EE%	99%

Then you do some soul searching about your own personal risk tolerance
and decide how much of your account to risk on each trade. Are you
willing to have a 50% risk of going broke? Then bet big. Are you more a
1% risk kind of guy? Then bet small.

Do an archive search on "bet sizing" "position sizing" and "Monte
Carlo." You will find lots of good stuff in there. Here's an old one of
mine showing how Monte Carlo works. These aren't volatility adjusted but
it demonstrates the kind of results you might get.

-- 
  Dennis

-------- Original Message --------
Subject: Oddball Monte Carlo
Date: Sat, 20 Apr 2002 12:16:26 -0700
From: DH
To: Omega List <omega-list@xxxxxxxxxx>

Mark reminded me he has the Oddball trades posted on the web. There are
366 trades of the SPX cash index - not a very realistic test but good
enough to illustrate what we are doing. He's using $250/point to
simulate the SP futures with no slip/com. I subtracted $250/RT for
slip/com and ran it through 10,000 loops of the MC. Here are the
results. Small account guys could divide the numbers by 5 to get an idea
of how much they should have in the account for each mini ES they trade,
keeping in mind that the futures probably won't do as well as the cash.

-- 
  Dennis

Probability of drawdown         
        Size of drawdown per contract   
                Min. account size per contract with $19,688 margin
0.01%   $94,618 $114,306
0.1%    $79,200 $98,888
0.2%    $73,483 $93,171
0.3%    $71,235 $90,923
0.4%    $69,673 $89,361
0.5%    $68,525 $88,213
0.6%    $67,313 $87,001
0.7%    $66,510 $86,198
0.8%    $65,923 $85,611
0.9%    $64,863 $84,551
1%      $64,275 $83,963
2%      $60,013 $79,701
3%      $57,458 $77,146
4%      $55,480 $75,168
5%      $53,903 $73,591
6%      $52,523 $72,211
7%      $51,205 $70,893
8%      $50,353 $70,041
9%      $49,240 $68,928
10%     $48,330 $68,018
11%     $47,703 $67,391
12%     $47,108 $66,796
13%     $46,480 $66,168
14%     $45,920 $65,608
15%     $45,378 $65,066
16%     $44,905 $64,593
17%     $44,493 $64,181
18%     $44,020 $63,708
19%     $43,623 $63,311
20%     $43,238 $62,926
25%     $41,078 $60,766
30%     $39,405 $59,093
35%     $37,960 $57,648
40%     $36,628 $56,316
45%     $35,390 $55,078
50%     $34,208 $53,896
55%     $33,195 $52,883
60%     $32,073 $51,761
65%     $31,038 $50,726
70%     $30,018 $49,706
75%     $29,070 $48,758
80%     $28,025 $47,713
85%     $26,840 $46,528
90%     $25,460 $45,148
95%     $23,478 $43,166
99%     $20,485 $40,173
99.9%   $17,913 $37,601
99.99%  $15,810 $35,498