What is an ID proponent's basis of comparison? (edited)

No that's not correct. In fact, in some robot control system evolutionary algorithms, I know for sure that the initial movements of the robot are generated purely randomly. There are no initial parameters! I'm not sure about some of the other examples I quoted, but I don't believe the initial parameters are critical to the end result. Maybe others have knowledge here....

And that does not relate to its probability of occurring. Some very complex things are virtually certain to occur; decomposition for instance. Other, simpler things are much less likely to occur; for instance, the components of an object being arranged completely uniformly.

quote:

---

Sorry Smooth Operator, one of the authors of the original no free lunch paper you quote has written an explicit refutation of the use made of NFL theorems by Dembski.

---

No, he made a rant. What he wrote is not a refutation but a rant.

quote:

---

Note that he is also quite critical of some evolutionary biologists so he is clearly not biased in their favour.

---

Actually he is. He can be critical of anyone he want's to but that doesn't mean he doesn't support them.

quote:

---

He says in essence that Dembski's use of the NFL theorems is so vague and imprecise that he actually demonstrates nothing. Here's a link.

---

I know, i read that. Saying that it's vague and imprecise is not a refutation but a rant.The only thing he actually said, that is wroth commenting on is this:

quote:

---

Rather it is a co-evolutionary process. Roughly speaking, as each genome changes from one generation to the next, it modifies the surfaces that the other genomes are searching. And recent results indicate that NFL results do not hold in co-evolution.

---

Basicly he did a new experiment which shows that NFL does not hold for co-evolution. And now if you actually want to see a real refutation. Look at what Dembski wrote about his new work.http://www.designinference.com/...ong_competitive_agents.pdfDembski shows that the NFL theorems still hold even in co evolution. But basicly he didn't have to since Wolpert's real article is supportive of NFL in biological evolution even in co-evolutionary cases.

quote:

---

However, in the typical coevolutionary scenarios encountered in biology, where there is no champion, the NFL theorems still hold.

---

Wolpert's own words.Coevolutionary free lunches | IEEE Journals & Magazine | IEEE Xplore

quote:

---

You are also misunderstanding what the NFLs claim. They say that if you want results when searching a specific space you need a tailored search algorithm.

---

Yeah, obviously.

quote:

---

This is not the same as your claim that you need intelligent input for the algorithm to work on. You can see this by reading your own quote above.

---

That's the same thing as you said. The specific algorithms that are "tailored" are being inputed with information by an intelligence.You first have to create the algorithm to work with it, obviously!

quote:

---

No that's not correct. In fact, in some robot control system evolutionary algorithms, I know for sure that the initial movements of the robot are generated purely randomly.

---

No they are not. They are programmed to act like that.

quote:

---

There are no initial parameters!

---

Yes there are. The program itself is the initial parameter. The rundom number generator itself is based on an algorith that produces pseudo-random numbers.

quote:

---

I'm not sure about some of the other examples I quoted, but I don't believe the initial parameters are critical to the end result. Maybe others have knowledge here....

---

Well you obviously don't have enough knowledge to be speaking about it now do you? If the initital parameters were not critical than every algorith would produce the same results on all search spaces. Which clearly doesn't happen.

quote:

---

And that does not relate to its probability of occurring.

---

Well you obviously never saw a lock than. Are you telling me that it is easier to open a lock by pure chance that has 100 combinations and a one that has 100.000.000.000 combinations?

quote:

---

Some very complex things are virtually certain to occur; decomposition for instance.

---

Yes becasue they are predispositioned by natural laws. There is no natural law to open a lock or get a DNA sequence into a specific, biologicaly functional sequence.

quote:

---

Other, simpler things are much less likely to occur; for instance, the components of an object being arranged completely uniformly.

---

That's becasue there is no natural law for it.

That is not what Kolmogorov complexity describes. For example:

quote:

---

abababababababababababababababababababababababababababababababab
4c1j5b2p0cv4w1x8rx2y39umgw5q85s7uraqbjfdppa0q7nieieqe9noc4cvafzf

---

The first string is very easy to define; it is "ab" 32 times. The second string is much harder to describe short of just writing the string down. Both are equally likely to occur given random chance, as they both have 64 characters.So, as you should see now, they are not related. You didn't understand what you were talking about, and neither did your source.

quote:

---

The first string is very easy to define; it is "ab" 32 times. The second string is much harder to describe short of just writing the string down. Both are equally likely to occur given random chance, as they both have 64 characters.
So, as you should see now, they are not related. You didn't understand what you were talking about, and neither did your source.

---

No, you are the one that forgot about Shannon's entropy. It and Kolmogorov complexity together, both describe an event that's size is inversly proportional to the chance that it occures.So how are they not related? Both have the same chance of occuring since both have 64 characters. If one had 32, it would be easier to happen by chance.Edited by Smooth Operator, : No reason given.

You are incorrect.*sigh* There's no possibility of your mind being open to evidence, based on what I've seen of your behaviour so far.

So the original author of the paper is not in a position to judge when it has been misused? You're on very weak ground here. You are showing strong confirmation bias. Again, not true. How can every algorithm apply to all search spaces?Edited by Richard Townsend, : No reason given.

quote:

---

You are incorrect.

---

Why? Are you telling me that robots actually think for themselves?

quote:

---

*sigh* There's no possibility of your mind being open to evidence, based on what I've seen of your behaviour so far.

---

Evidence? What evidence? What actual evidence did you show me? Where?

quote:

---

So the original author of the paper is not in a position to judge when it has been misused?

---

Yes he is, and he can be righ or wrong. This is science, nobody is right 100% of the time.

quote:

---

You're on very weak ground here. You are showing strong confirmation bias.

---

No, I actually showed you evidence. Unlike you, who showed me rants. Cite me one good argument that Wolpert made against Dembski's use of the NFL.His rant was nothing like Dembski did, when he responded to his co-evolutionary tests. But, you didn't even bother to read what I showed you. Admit it!

quote:

---

Again, not true. How can every algorithm apply to all search spaces?

---

It can't that's trhe point! LOL, it's like talking to a kid! It can't apply the same way because of the INITIAL PROGRAMMING that has been done to the algorith itslef. If those initial parameters were not important AS YOU CLAIM than they would all be the same.Edited by Smooth Operator, : No reason given.Edited by Smooth Operator, : No reason given.

No, I'm telling you that the initial movements of the robots are generated purely randomly. You seem not to believe this. I'm saying your behaviour shows you are not interested in evidence that contradicts your current beliefs. Edited by Richard Townsend, : No reason given.

Are you just mentioning random vaguely-related concepts hoping one of them supports you? You were quoted three articles that disproved your theory, you provided feeble and easily refuted complaints for two, and the third you didn't even understand. You have been proven wrong on this criticism and your attempts to justify yourself are just digging a bigger hole.

quote:

---

No, I'm telling you that the initial movements of the robots are generated purely randomly. You seem not to believe this.

---

Because it is impossible. Everything a machine does is programed in advance.

On this point you are quite right. I do admit it. It was not a rant - that's why I accuse you of confirmation bias. You're not willing to consider his arguments. You misunderstand the the concepts of algorithm and parameters. The initial parameters are the INITIAL INPUTS TO to the algorithm. You're right that the algorithm is pre-defined before the evolutionary process starts. But this is irrelevant to my original point, which is that 'specified complexity' can be generated (and has been generated) by the purely automatic execution of this kind of evolutionary algorithm, with no or very basic specified input parameters.This is an indisputable fact.That's why I think the more interesting question is - can such an evolutionary system arise without human intervention. Do you see the difference? Edited by Richard Townsend, : No reason given.Edited by Richard Townsend, : No reason given.