Relationship between "entry frequency" and "reproducibility" in EA
Hope you are well,
This is Dr. Neko.
This time,
focusing on Power_Scal,
I would like to describe the relationship between “entry frequency” and “reproducibility.”
＜Power_Scal product page＞
https://www.gogojungle.co.jp/systemtrade/fx/23789
Even now, Power_Scal above is continuing to perform well, and
as one of the features of this EA,
there is a point that it has a high “entry frequency.”
First, please take a look at the backtest results below.
As shown in the test results above,
Power_Scal
has
a backtest from 2005 onward,
with a trading frequency of about 17,000 trades.
For an EA with a single position (maximum number of positions = 1),
this is a very high trading frequency, and
it is rare to see an EA with nearly six months of revenue curves showing a consistently rising trend.
Now, for the EA,
what are the benefits of having a high “entry frequency,”
and
I would like to explain this time.
～Conclusion～
【For the EA, higher reproducibility (reliability) of trades】
I already stated the conclusion, but
this is a consequence of
having more trials
and thus more opportunities for reliability.
For example,
consider a die that tends to roll a 1.
Here,
【if you roll 10 times and get “1” five times】
and
【if you roll 100 times and get “1” fifty times】
which would be more reliable?
This is,
【getting five “1”s out of 10 rolls】
is far less reliable than
The 10-roll scenario can occur by chance,
so it is not reliable due to insufficient trials,
whereas the 100-roll scenario is not likely to be due to chance.
(Conversely, getting 50 “1”s out of 100 is practically impossible by coincidence.)
The same applies to EAs,
where an EA with 2,000 trades showing a rising trend and
an EA with 20,000 trades showing a rising trend
differ in reliability by a factor of about 10.
This completes the explanation of
【high reproducibility (reliability) due to a large number of trials】.
In the case of Power_Scal below,
which shows a rising trend over 17,000 trades,
it has exceptionally high reliability compared with typical EAs,
and
this is reflected in its current forward performance.
＜Power_Scal product page＞
https://www.gogojungle.co.jp/systemtrade/fx/23789
※The above explanation assumes
that the price movement of the currency pair in that period does not deviate significantly from normal movement.
In extremely abnormal market conditions,
performance may deteriorate to some extent.
(Even so, compared with ordinary EAs, reliability remains higher,
and this does not change significantly.)
I hope the above will be helpful as a reference for future methods of evaluating EAs.
Dr. Neko
