|
PureBytes Links
Trading Reference Links
|
Frans,
You've got it correct as far as an individual trade. For a system, you
aggregate all the MAE's of the individual trades in a graph (or table) that
shows the MAE's for winning trades and the MAE's for losing trades. If the
trading rule can discern between winners and losers, you should see
different distributions. Typically, the winning trades have an MAE beyond
which they usually won't go. In other words, winning trades only go wrong
just so far.
This was first published in the October 1985 TASAC but that's unavailable
now unless you buy the CD-ROM for all the back issues (about $400 !!). Or
you can buy the book (yep! I wrote it.).
I'm developing a web site (http://seagym.com/mae) to explain all that and
I'm enthusiastic about answering questions so feel free to ask.
John Sweeney
>Hi Walter,
>
>Gee you really force me into studying this MAE-thing <G>
>
>Do you or anybody else know if there is anything on the web by John Sweeny
>regarding this subject ? Or did he write about it in TASAC ?
>
>I checked out Technical Analysis fm A to Z by Steven Achelis. Nothing there.
>
>The MS6.5 handbook pg. 322 says it is the worst intraday price move against
>a position, measured from the entry price.
>
>I interpret this as the maximum open drawdown of the specific trade your
>looking at.
>
>And than the system MAE would be the max. open drawdown over all the trades
>done in that systemtest.(?)
>
>Perhaps I simply don't get the feel for what he means.
>
>Regards,
>
>Frans
>
>
>
>
>At 15:58 7-09-98 -0400, you wrote:
>>Hi Frans
>>
>>I put this on the bulletin board because of possible confusion others may
>>have about MAE (Maximum Adverse Excursion).
>>
>>If you do a system test of 2 MA crossovers on a tradeable, you may have 120
>>test reports. Each test report, i.e., 5 and 20 SMA cross will produce a
>>system MAE, because the MA's will cross and get you out at that MAE level.
>>
>>If you pick one system report and go to trades you may see 10 to 100 trades.
>>Each trade will have a MAE, none will exceed the system MAE.
>>
>>I do not use the system MAE to calculate "proven correct" for the
>>tradeable. Instead I look at the MAE for each trade of the system (i.e., 5 &
>>20 SMA cross). These are arranged like a distribution under a bell curve.
>>What you are looking for is a "fat" tail. The distribution is put into
>>"groups" or "bins". See Sweeney for a complete explanation.
>>
>>The same thing is done with the MFE (Maximum Favourable Excursion). This is
>>superinposed over the MAE. You are looking for the most number of MFE in the
>>smallest size of MAE.
>>
>>This is the basis I use for the distance part of the "proven correct"
>>calculation.
>>
>>When I used the expression "that what goes, goes", what I meant was that
>>some trades have a MAE of "0". They just take off and never look back.
>>
>>... Will write later.
>>
>>Walter
>>
>>
>>
>>
>>
John Sweeney, Tech. Editor Technical Analysis of Stocks & Commodities
Technical Analysis, Inc. The Traders' Magazine
4757 California Ave. S.W. Phone: 206 938-0570 Fax: 206 938-1307
Seattle, WA 98116-4499 USA Web: http://www.traders.com/
____________________________________________________________________
Contents may not reflect official opinion of Technical Analysis, Inc.
|