A working version of FMM’s design detection game is available.
Download and install the applicable version of “Processing”.
https://processing.org/download/?processing
Get the fifthmonarchyman progam code from here, and paste it into the Processing script area.
Sample data here
http://pastebin.com/raw/MjV8RmvW
You need two files in the same folder as the Processing executable.
real.txt and fake.txt
The testing and such starts here
test strings
The place to clarify the requirements for submitted strings was some time ago.
Operational definitions. Look it up.
It means defined by what you do to create them. Not what can’t be done.
If you want your claims of being able to recognize some property of data, you need to submit to adversarial testing. It’s the nature of the beast. You have made an extraordinary claim.
Patrick,
You were drunk when you posted this? Is that why your post is so stupid?
Maybe something I expect to have happened has not been made clear.
In order for your game to be interesting, you need to have some independent way of confirming the properties of your strings. Independent of the game.
The property can’t be defined as a negative. One can’t say that a snippet is part of a non-computable number.
I thought we had reached some meeting of the minds regarding this. I thought that “real” strings could be strings defining a real complex object. Say a biological object. That would rule out strings that are the result of an algorithm.
Then I thought you were claiming to be able to distinguish such strings from randomly generated strings. We even discussed the operational definition of random, and you agreed that random number generators would be adequate.
Then you discussed strings in which the order of the data is scrambled, or in which the string is cut and reassembled.
I thought these were the terms, and taking these terms in good faith, I spent several days searching for suitable data. Strings not created by algorithms, real published data having an open source, so as the game progressed, anyone could check the data submitted to the game against the original published data.
I spent time scrambling the data and obtaining equivalent randomized data.
I assumed that when you claimed you could distinguish patterns in non random data, you meant you could tell “real” from “fake”. Those are your terms. The game as programmed appears to support my interpretation. One “guesses” at which is real, and the game confirms the guess.
The game program itself is not yet functional as a guessing game. the “real” data is always on top. So I compensated by submitting unlabeled data. I expected you to play the game and report back which set of data was which.
It is possible that it is interesting, just not in the way you expected
Was either of them non random non computable?
While I was trying to produce my submissions, I ran across some attempts to distinguish the output of random number generators.
Some common rnd() functions produce patterns that can be observed when the numbers are converted to graphics.
So I do not dismiss Fifth’s claims out of hand. His approach is not entirely outside the realm of mathematical research.
What I find to be outside the realm of reasonable is his approach to testing and validation. I will not repeat my concerns.
I’m really sorry fifth, I felt the game/project was doomed from the start because I think one shouldn’t be looking at strings for design unless the strings represent some underlying physical phenomenon where we have some idea what uncertain (random) processes will generate.
I felt your critics were humoring you so that they could score points against an IDist.
Petruska, Patrick and others have criticisms I sympathize with. Sorry, old friend.
I think everyone thought the game had a point beyond different strings have different patterns.
stcordova,
I for one would be delighted to find that an intelligent design creationist had actually come up with an objective metric for identifying design. fifthmonarchyman’s failure to clearly define his terms and support his claims is his fault, not that of his critics.
IDCism has never been more than a religiously-based political movement. While it’s important to “score points” in the political realm to prevent that sectarian dogma from gaining the support of government power, I’d be very interested in seeing something more than the usual arguments from incredulity. Unfortunately, it doesn’t appear that we’re going to get that in this thread.
Patrick,
What you call incredulity, I call proof by contradiction — meaning certain features of the universe are not consistent with natural expectation. It’s not, “I can’t see how that can happen”, but rather, “the features are not consistent with natural expectation.”
What ever one makes of the violation of expectation, is up to them. There may not even be a formal right or wrong answer short of God showing up and giving a visual demonstration.
But it is science to say a physical phenomenon is not the expected outcome of typical (dare I say natural) processes. It is wrong to keep pretending the features of life are consistent with chemical expectation from supposed pre-biotic conditions (non-life). IDists can argue that angle (violation of expectation) quite well.
The issue I have with FMMs game is we are dealing with strings rather than physical systems where we can even state what expectation is and deviation from expectation.
I don’t think it’s right to insinuate the IDists don’t have a legitimate objection to OOL or the evolution of taxonomically restricted features (like say eukaryotic genomes). ID lives because OOL and evolution of taxonomically restricted features (TRFs) are not demonstrated to happen by processes repeatable in the lab.
For one to insist TRF evolution and OOL are true in the absence of repeatable experiments to that effect I find problematic for those claiming such ideas are actually empirical and scientific. The claims may be naturalistic claims, but naturalism is not automatically scientific nor empirical since the heart of science is repeatability, neither of which is on the side of OOL or evolution of TRFs.
I don’t think that is a fair assessment. Speaking for myself, I have some interest in pattern recognition dating back to my working life and that is why fifth’s post caught my eye. More generally, what kept people involved I think is that finally an ID proponent comes up with a proposal for a design detection method that we can actually try out in practice. Something that potentially goes beyond ‘it looks designed therefore I think it is’ – which is what in my view all other proposed ‘methods’ boil down to. We gave it a fair hearing and offered constructive criticism.
The trouble now is that fmm seems to have watered down his claims from ‘a reliable way to detect design in a string’ to ‘identifying differences between strings’. By now we have hundreds of his posts that often consist of only a few lines, and it is hard to precisely understand his claims. What he should do is sit down and spend some time to carefully compose an OP that clearly, logically and unambiguously lays out his hypothesis and how his ‘game’ addresses it. That way we can minimise the misunderstandings and repetitions that plague these threads.
Fmm claims he loves science. That is great, but he should put his money where his mouth is and write down his ideas according to the conventions of science. These conventions are there for a reason. Internet chatting so often creates confusion rather than clarity of understanding.
fG
All,
As you know I have been in a stupor for the better part of a week. The fog is just now beginning to clear.
I think the recent misunderstanding with petrushka’s multiple strings might be the partially the result of that fog.
I think that I misunderstood what he was trying to do with his multiple string dump. If all he want’s is me to give him my results for each string individually I can do that one by one. And despite what I might have said I think I can say whether an individual string is nonrandom and noncomputable
I have some stuff to take care of this morning and then I will hit the reset button and attempt to explain all this. I have a feeling that we are not as far apart as it appears right now.
I only want to remind every one that we are still waiting on OMagain to make the game shareable, So that testing and results will be transparent and public
What we are dealing with now is my original crude programming attempt from several months ago.
peace
my version will be nothing more then what you already have, just on a website. So, what it does not it will do then, what it does not do now, it will not do then.
If you really give a shit, start an OP just for testing and results and use what you have now.
Mine will be exactly the same as yours, can I make that very clear? If yours is crude, so will mine.
All yours does is plot two graphs from two data sets. that is all mine will do. I may make the attempts shareable, but that is nothing you can’t do right now with a new OP.
stcordova,
No, you’re saying “I don’t see how a natural process could result in this.” Neither you nor any other IDCist have demonstrated any contradiction. It’s a classic argument from incredulity.
What’s worse is that you use bogus examples like a bunch of coins coming up all heads or dominos standing on their edges without tying those to modern evolutionary theory or even discussing cumulative selection. You use strawman arguments about modern dogs evolving into modern cats. You mock trained, working scientists rather than encouraging the wonder of discovery. You do all this to children who have yet to learn that some authority figures should not be trusted.
What you are doing isn’t teaching, it’s child abuse. You are not behaving like a good person.
What we need is the ability to compile and view results from the individual tests and the ability to have double blind tests in which you don’t know what string you are testing or which assay you are performing.
In a perfect world a person would need to do one test every time they enter the website and at that point they could submit strings or view results that others have gotten.
If you are unable to move us toward that
I will continue to work on it as the spirit moves me.
peace
That should be doable. You should carry on regardless however.
The misunderstanding I suppose is in the concept of a “submitted string”,
It never occurred to me that you would not be interested in actually playing the game but would instead simply want to submit strings for others to evaluate.
I not used to that kind of lack of curiosity,
It’s really simple if you have a string that you want some one to look at let me know. When I get time I’ll take a gander.
There are no “requirements”
Some strings work better than others but any numerical sequence will load into the game.
If we can distinguish the submitted string from random ones and strings from close models then we can say it is nonrandom and noncomputable. That is pretty much it.
I would only once again emphasize that the intent of the game was to work on this stuff together and try to move the ID conversation forward a little bit.
I suppose there are those who are happy with things just like they are.
To each his own
peace
Is there a common characteristic to the nonrandom and non computable strings or do they differ like the random computable strings?
like I said I think this misunderstanding was the result of Petruska’s string dump coupled with my flu.
I thought he was looking to play the game when in reality he just wanted others to look at strings he was submitting and give our appraisal.
I still find this to be an odd approach.
If I had to guess I’d say he wants to show that the game causes false positives and negatives and is generally worthless.
I would think a better way to demonstrate this is to play the game and attempt to show that the results are not repeatable and are subject to personal bias.
that is my intention as soon as we have a working shareable game to discuss
peace
As far as I know there is no common characteristic to point to.
Every string is different.
All we can say is that with nonrandom noncomputable strings some pattern will distinguish them from randomized strings and strings from models that are close.
peace
This definition is of no use unless there is an independent method for determining non-computability, or an agreed upon method for producing non-computable strings.
fifthmonarchyman,
I’m glad to hear you’re on the mend.
The best way for you to demonstrate your claims is to provide examples of how you use your game to determine whether or not a string is designed. These are your claims, after all. You are the one with the burden of proof to support them.
Several people have posted reasoned comments explaining this to you. If you are interested in demonstrating the value of your game, you could do worse than to review petrushka’s comment on his expectations and faded_Glory’s comment on scientific conventions. Address petrushka’s expectations, resetting them if necessary, and follow faded_Glory’s recommended format for laying out your definitions, rules of your game, and the claims you are making.
Until you increase the clarity by doing that, this discussion is going to continue to bog down.
My intention is to see if fifth can demonstrate his claimed ability.
But before that, I would like to know what the claimed ability is.
That requires an operational definition of the terms real string or noncomputable string.
Operational does not mean defining what the string is. It means describing an objective procedure for producing such string. Or, alternatively, an independent and agreed upon way of verifying the “realness” of strings.
Playing the game makes no sense until the definitions are agreed upon.
Actually in the game that I originally gave to OMagain the strings are randomly shuffled at each time you guess. I’m not sure about the game you have
This is correct.
In the first two strings you gave me I assumed we were evaluating the first string and that the second string was a “fake”. I loaded it this way and as you know immediately was able to distinguish a pattern and a difference between the strings I reported this to you.
At that point you asked me which sting was real and which was fake.
That is where the confusion began, If I thought we were evaluating both strings I would have randomized them both myself and also come up with models to compare with both on them.
Instead of one comparison we would need at least 4 and probably more.
It is a much more time consuming process to do all that. I thought we were just playing the game.
Does that clear it up for you
There is only one pastebin of the original code, it is in the OP.
We have been using Maguire’s definition of an integrating function
ie the knowledge of m(z)does not help to describe m(z′),when z and z′are close
so we determine non-computability by seeing if a close model helps to describe the string we are evaluating
does that make sense to you?
peace
It doesn’t make any difference whether the game is clear.
I am not now and never have been interested in this or any other game. I don’t do games.
I am interested in two and only two aspects of your program
One is your operational definition of “real” or non-computable. You need either to provide a procedure for producing such strings, or an independent (independent of the game) procedure for validating strings.
A second interest is whether, in fact, you can demonstrate your claimed ability to put real and fake strings into buckets.
And the procedure for that is…
we see if we can distinguish between the real string and a close model
peace
And the procedure for that is…
short answer is “the game”
long answer is there is no single procedure because if there was a procedure then it would be computable.
peace
This iss absolutely essential for validating your game.
Give us a procedure for generating or acquiring “real” strings, and a method independent of the game for validating them as “real”.
Without those items, the game is worthless.
So you have this claim that cannot be verified.
fifthmonarchyman,
Let’s aim for more precision here, please.
If I understand your responses to petrushka correctly, you’re claiming that you can tell if a particular string is designed or not by making noisy copies of it and running it through your game. If that’s the case, I suggest the following:
1) Take two pairs of the strings petrushka provided.
2) For each of the four strings:
2a) make your noisy copies
2b) play your game
2c) determine if the string is designed or not
3) Report your findings, including explicit details of each of the steps.
This will demonstrate exactly what you mean and will provide a test of your claim.
If non-computability cannot be verified, you could at least provide a objective procedure for acquiring real strings.
You seem to have gotten into this using financial data. Is that acceptable? Temperature or barometer readings? Genomes?
Anything at all the is accessible and verifiable?
what claim exactly?
peace
Do you understand that any phyisical object or process can be represented as a numerical string?
How it is represented depends on what you want to represent,
A string is “real” if it comes from a real object or process
peace
Your claim that you can segregate strings into integrated or real or non-computable (or whatever) and those that are not.
I see two problems, the first of which is the most serious.
Before you claim you can identify a class of string, you must have an independent way of verifying that your identification is correct.
You could have an independent procedure, but you have suggested this is not possible.
But to play the game, you must have a supply of real strings.
The alternate way of supplying them is to define how they are acquired. Experimental data? Temperature histories?
What?
no offense Patrick but I have zero confidence that your aim is more precision.
peace
Stop being a shithead.
I have suggested a number of possible sources.
Look at my list and say yes or no.
If my list is unsatisfactory, provide an acceptable list.
Off to guano, please.
They can be acquired any way you choose,
It’s up to you to decide what you want to evaluate
The way to do this is to see if different observers have the same results
you know when they actually play the game .
peace
fifthmonarchyman,
Assume I’m posting in good faith. In fact, I am. I don’t think you can actually do what you claim, but I’d like to see your claim fairly evaluated.
What, if anything, do you object to about the process I suggested?
I have provided six strings acquired from real objects. Two are the original data, and the rest are either random strings or scrambled versions of the original.
Is this acceptable? I have asked a bunch of times without getting a yes or no answer.
why not try it for yourself and see what you see?
I have never in my life seen such a lack of curiosity
nothing much except it is something you could do if you were actually interested
peace
like I said any string at all will load.
That is something you would already know if you just gave it a try,
peace
fifthmonarchyman,
Given the vagueness of your descriptions, I could not.
Since the process is correct, use it to clearly describe your game and support your claim.
That isn’t the question I asked. The question is are the strings I supply as real acceptable to you as real?
Yes or no?
Amazing how difficult it is to get a straight answer from a creationist or theist of any stripe.