Most profitable strategies are not necessarily curve fitted.

if you want to split hairs that is true

but historical data is the only thing we have to develop the strategy.

it is easy to say markets repeat:but because it is emotions that drives markets,not rationality, it is not predictable.a person,blinded by fear or greed,will do anything.so markets will do anything; this 'anything' is not easily described by rules,so to capture it with an ea is not easy.

That does not stop me from trying.

Hehe

In limited number of indicators given enough time it will at some point figure out the most profitable strategies but you can't really tell when. This however does NOT really matter because the most profitable strategy on historical data will have high degree of curve fitting (won't it?)

In my humble opinion the Generator should be used more as an source of good ideas and not as the one-and-only tool. Discussing it in such philosophical detail suggests that it's the one-and-only-tool and it is not

since all technical analysis uses historical data, it follows that technical analysis does curve fit.

since human emotions repeat, price patterns repeat.

All  strategies are a result of curve fitting.

the advantage with the generator is that it curve fits at a speed which we can only dream of doing manually.so we can analyse a huge number of strategies.

I sometimes let it run for days to see what combinations it might produce.

Sometimes there is an idea that I have never thought about and often there is nothing to look at.

I have not seen it decide to use more than a few indicator slots unless I set up something in logical operators.

In limited number of indicators given enough time it will at some point figure out the most profitable strategies but you can't really tell when. This however does NOT really matter because the most profitable strategy on historical data will have high degree of curve fitting (won't it?)

In my humble opinion the Generator should be used more as an source of good ideas and not as the one-and-only tool. Discussing it in such philosophical detail suggests that it's the one-and-only-tool and it is not

Hmmm

Miroslav

1> Is it possible to add a PAUSE in Generator?

2> Could you give some more information about "how Genεrator works" in more detail?

Τhanks in Advance

There have been numerous questions about the pause feature in generator.
1) Since you can imagine that such a huge number of strategies cannot be stored or tracked when generating ... the generator creates random strategies and tests them against the best strategy so far created (comparison is based on Net Profit or whatever you have chosen in the settings). If the strategy is better - it gets the new first place. If it's not better the generator just makes a new random one. The "time limit" is just to make sure you don't wait a zillion-million-billion years to generate the super-uber strategy
I mean if you are picking random sand grains on the beach it does not make sense to "pause" does it? Starting and stopping would mean the exact same thing.

2) I think I also explained that in 1 without noticing

Have a great day

1> Is it possible to add a PAUSE in Generator?

2> Could you give some more information about "how Genεrator works" in more detail?

Τhanks in Advance

I need that tou calculate hours of working of Generator.

The generator must work 293640906760361053031104848.00274^100 years. I'm afraid that the Time will end earlier.

The Universe is 14000000000 years old.

If you have one MA with Period, Shift, Smoothing method and base price, the combinations are: 200*200*4*7 =  1120000 * 8 variants of logical conditions = 8960000

For one MACD indicator you have 10752000000 combinations.

If you have MA and a MACD, the combinations are: 8960000 * 10752000000 = 96337920000000000

The program comes with 100 indicators and 18 slots.

The rough number of the available strategies is 200000000^(100*18)

No criteria.

How big is this number of combinations?

I need a rough estimation.

If is very big, we need to use not everything in the generator but selection of indicastors.

I need that tou calculate hours of working of Generator.

How many spectra is the maximum number which can be calculated by generator?

The number of calculations is not limited. I'm using an integer number for the counter, so it will overfloat at 2,147,483,647 calculations. It may start showing negative numbers after that.

You see the counter. The Generator doesn't keep strategies except the "best one". All other are ether deleted or send to the collections repo. The collection also filters out the worst strategies keeping only the top 100. You can generate all day long without problems.

During the use of generator, I suppose a large number of spectra is calculated.

Suppose we have in our bank to use and select all available in repository.

How many spectra is the maximum number which can be calculated by generator?

I Measure thousands.

Do you know roughly how many are in total in this moment?

The answer looks trivial but is useful to estimate how much of the space available each time Generator searches with parameters selected by a user.

Thanks in Advance.