Thoughts on the "Lifespan of EA" ~ The relationship between the number of EA users and the lifespan of EA ~

I hope you are doing well. This is Dr. Neko.

Now, today,
about the concept of “EA lifespan,”
I would like to briefly share my thoughts.

Regarding EAs,
there are various patterns such as “long-lived” and “short-lived,”
and I think there are many forms of EAs.

And,

“If an EA sells too well,

its edge may disappear, leading to a shorter lifespan,”

is a sentiment I sometimes see, I think.

As for this [EA lifespan],

different people may have different views,
but in my case, I have produced and sold many EAs,
and I have checked the results of numerous EAs by other authors for nearly 10 years,
and my own impression is that
the number of units an EA sells does not actually have much influence on its lifespan.

Why the number of units sold doesn’t affect lifespan,
is because if you think about the market mechanism of FX,
it is quite straightforward.

What it means is that
・[Foreign exchange rates (FX) move toward the side with more sell or buy orders]
as a market principle.

In other words,

“As the number of EA users increases,”
“the amount of one-way orders at nearby price points increases,”
“and, by market forces, moves more easily in that direction.”

This is the reason.

In other words,
“The more users an EA has, the easier it becomes to win,”

is the actual phenomenon, considering market principles, I think.

In other words,
it is not that the increased number of users makes the EA stop functioning, but
that the market state shifts away from the EA’s logic (entry conditions),
causing its logic to stop functioning, which is
the correct way to think, I feel.

And, in reality,
if the impact were only from the EA’s users’ orders,
the impact on the currency market (the direction of price) would be not so large.

(Considering the size of the FX market,

the order volume for one Japanese EA is a negligible amount, so it can be ignored)

However, if the EA’s logic is a
logic that captures [the fundamental pattern of market movement],
then things change substantially.

Not only that EA, but
there will be many orders from other EAs, discretionary large traders, major investment institutions, automated trading using AI, and market open/close effects in the stock market, all moving in the same direction as that EA,
so

・If that EA’s logic captures [the fundamental pattern of market movement],

the lifespan of that EA could become dramatically longer,
I think this is a view you could take.

And when that EA’s logic stops working,
it would be when, as above, many entries including large investors disappear.
I think that would be the case.

This, in fact, is
“If there is an opportunity to win, even large investors will continue to enter at that timing, so long as that timing remains favorable,”
so deliberately stopping is only possible when
the timing shifts to a point where you can no longer win.

In other words,

【Only when there is a major market shift (significant deviation in market state) that makes winning with that logic impossible】

will that logic stop working.
That is what I feel.

In short,

・Whether the underlying market movement patterns that support the EA’s logic
shift or not.

I think this is the pattern of an EA that has been winning for a long time but then stops.
If the market movement patterns do not change, the logic should continue to work.
That would be the case.

Lastly, for reference,
I would like to note how to think about EAs that quickly stop winning in a short period of time.

This is not simply a logic that captures the pattern of market movement,
but rather a pattern of simply applying past market conditions to the entry condition expressions,
in my view.

If the logic had properly captured [the pattern of market movement],
it should not decline so quickly,
so a quick loss of winning would be a sign of a merely curve-fitted EA,
and that would be a realistic assessment at present, in my view.

(1) The EA’s lifespan depends not on the number of users, but on whether the fundamental pattern of market movement shifts,
and thus becomes a lifespan issue.

(2) “An increase in EA users” would, considering market principles, likely make winning easier.

(3) An EA that captures the fundamental market movement pattern is more likely to remain functional longer (has a longer lifespan).

This is, of course, only my personal viewpoint, but
as one perspective on the EA lifespan,
I hope you may find it参考 as a reference.

Dr. Neko