Annoying creationists

[kjkent1]

The only problem you have is that you don’t have any selection pressures. For example, I just won my malpractice case that you so much enjoyed trying to discredit me with. It only took 6 years to win the case. Now you watch the web carefully because the Superior Court only posts their decisions for a week or so.

I do enjoy making evolutionists twitter.

Good for you on your malpractice defense. There will be no living with you now.

Anyway, kindly get back to the thread topic. We're all waiting for you to prove evolution impossible.

 

[kleinman]

Annoying Creationists

The only problem you have is that you don’t have any selection pressures. For example, I just won my malpractice case that you so much enjoyed trying to discredit me with. It only took 6 years to win the case. Now you watch the web carefully because the Superior Court only posts their decisions for a week or so.

Good for you on your malpractice defense. There will be no living with you now.

Thank you and what makes you think that you could live with me before. Don’t you find it amazing that in all the years I have practiced medicine and the tens of thousands of people I have cared for, I have had only two malpractice cases, both which I have won with motions for summary judgment. I think it’s because when people tell me they smoke cigarettes, I tell them it’s good that they are doing this, they pay lots of taxes and keep us doctors busy.

Anyway, kindly get back to the thread topic. We're all waiting for you to prove evolution impossible.

Ok, ok, be patient, the wheels of science grind slowly but they do grind fine. However, I am not proving evolution impossible; I am proving the theory of evolution (common descent by mutation and selection) mathematically impossible. So where were we? Oh yes, we were mapping out the function:

F(n,G,g,mr,nsp) = gfc

Where,

n = population size
G = genome length
g = number of sites
mr = mutation rate
nsp = number of selection pressures
gfc = generations for convergence.

And we were starting the mapping process by making a series of cuts in the 6-dimensional space in the mr, gfc planes and we have obtained the data from these planes for the following cases:

G=256, population=64, g=16, nsp=3:
mr|gfc to perfect creature
10^-7|95975841
10^-6|4278078
10^-5|1105535
10^-4|47279
10^-3|5029
1|662
2|572
3|12890
4|>1,000,000 did not converge

Note that I have included in this series a few more data points with more realistic mutation rates (1 mutation per 100,000 bases per generation down to 1 mutation per 10,000,000 bases per generation). Even this very simple series, the basis of which is the baseline case which Dr Schneider included in his publication shows that with a realistic mutation rate, it can take almost one hundred million generations to accomplish the sorting process for his three selection conditions. This series show that the maximum rate of sorting is around 2 mutations per genome per generation (2 mutations per 256 bases per generation) or about 1 mutation per 128 bases per generation.

G=512, population=64, g=16, nsp=3:
mr|gfc to perfect creature
1|2412
2|1251
3|973
4|13251
5|22790
6|>1,000,000 did not converge

This series show that the maximum rate of sorting is around 3 mutations per genome per generation (3 mutations per 512 bases per generation) or about 1 mutation per 170 bases per generation.

G=1024, population=64, g=16, nsp=3:
mr|gfc to perfect creature
1|18030
2|9701
3|4679
4|2991
5|1299
6|1979
7|3782
8|7254
9|16243
10|>1,000,000 did not converge

This series show that the maximum rate of sorting is around 5 mutations per genome per generation (5 mutations per 1024 bases per generation) or about 1 mutation per 205 bases per generation.

Here are a couple more series which show the effects of mutation rate on the generations for convergence to perfect creature.

G=2048, population=64, g=16, nsp=3:
mr|gfc to perfect creature
1|42641
2|36817
3|11480
4|6049
5|10364
6|9708
7|3634
8|5382
9|5089
10|6170
11|9851
12|3952
13|10880
14|38738
15|6145
16|105799
17|259469
18|>1,000,000 did not converge

This series show that the maximum rate of sorting is around 7-12 mutations per genome per generation (9 mutations per 2048 bases per generation) or about 1 mutation per 228 bases per generation.

G=4096, population=64, g=16, nsp=3:
mr|gfc to perfect creature
1|163722
2|48755
3|57212
4|30316
5|19675
6|23306
7|17689
8|21947
9|13328
10|8935
11|16043
12|10225
13|17029
14|16052
15|17186
16|7715
17|14761
18|8772
19|12152
20|8806
21|7647
22|6724
23|4959
24|4096
25|13463
26|8325
27|8067
28|17877
29|46876
30|1000000 did not converge

This series show that the maximum rate of sorting is around 24 mutations per genome per generation (24 mutations per 4096 bases per generation) or about 1 mutation per 170 bases per generation.

So, what can you say about the F(n,G,g,mr,nsp) = gfc 6-dimensional surface when you make a series of cuts through this surface with planes that are parallel to the mr, gfc coordinates. You can say that the surface has a paraboloid shape with the with the number of generations for convergence becoming extremely large to accomplish the sorting process when the mutation rate used is a realistic value or if the mutation rate used becomes unrealistically large.

The next coordinate we will consider will be g (the number of binding sites) and how it affects the gfc (generations for convergence).

I hope you all are enjoying the analysis of the mathematics of the mutation and selection sorting/optimization process.

 

[Mister Earl]

What weights are you using for the selection pressures? I note you left out that part. Also, when will you address Rocketdodger's results?

 

[kleinman]

Annoying Creationists

What weights are you using for the selection pressures? I note you left out that part. Also, when will you address Rocketdodger's results?

All parameters not listed in the post were left at Dr Schneider’s baseline values. If I don’t mention a value for a particular parameter in the model, it is because they were not changed from the default values.

What results has rocketdodger presented? Rocketdodger spent an afternoon throwing together some code and then claims that n+1 selection pressures evolve more rapidly than n selection pressures. I’ll stick with Dr Schneider’s peer reviewed and published computer simulation that has been scrutinized for more than 20 years and for which Dr Schneider has published and posted extensive documentation, especially since the mathematical results from Dr Schneider’s model is reflected in real empirical examples of mutation and selection.

The weights for all the selection pressures were left at the default values of 1. Why don’t you study the effects of varying weights on the rate of convergence?

 

[kjkent1]

The weights for all the selection pressures were left at the default values of 1. Why don’t you study the effects of varying weights on the rate of convergence?

I've already done that. Varying the weights creates a normal distribution centered around all three mistake weights set to the same number. The effect of the mistake weights are relative, not absolute, i.e., setting the weights all to 1 produces the same result as setting all the weights to 100.

The above is true unless you set a mistake weight to zero (0). When you do this, the "perfect creature" reported by the program is no longer actually perfect, because it contains missing and/or spurious bindings at the time perfection is reported. This is due to the fact that when the mistake weight is zero, the algorithm no longer detects the presence of that mistake.

Also, whenever a mistake weight is set to zero, Rseq never converges on Rfreq, clearly demonstrating that the genome being created does not model any known independently living life form, because all such life forms can be shown to have genomes where Rseq ~ Rfreq. I.e., ev without convergence, proves nothing about evolution.

Now, here's some questions that you can't answer without modifying the program:

1. The mistake weights can only be set from 1 to 100. There is no way to model what would occur were a mistake weight several magnitudes greater than some other.

2. How do we know that a creature with a small genome that is already generally perfect (Rseq ~ Rfreq) does not mutate by replication, for example, and then develop a new function without having to repeat a long iteration of fitness trials so as to cause the genome to converge again on perfection?

How do we know that the same cannot be said for a frame shift, a fusion, a deletion, a transposition, etc?

The early Earth had plenty of time to produce a small perfect creature, which could then act as a springboard for future and much faster evolutionary change.

Bottom line, kleinman, old boy, we only really need one self-replicator with a genome of any size wherein Rseq ~ Rfreq. As soon as this occurs, evolution is off and running.

And, while you can contend that the odds of this occurring are too remote to be possible, the fact is that we are here typing about it, and the evidence of common descent is overwhelming.

Therefore, that's what happened, unless you want to claim that God or some extraterrestrial set the ball rolling. Which you certainly can claim -- but the odds in favor of that are infinitely worse than random chance.

 

[kleinman]

Annoying Creationists

The weights for all the selection pressures were left at the default values of 1. Why don’t you study the effects of varying weights on the rate of convergence?

I've already done that. Varying the weights creates a normal distribution centered around all three mistake weights set to the same number. The effect of the mistake weights are relative, not absolute, i.e., setting the weights all to 1 produces the same result as setting all the weights to 100.

It should be obvious to you why setting all the mistake weights to the same number does not alter the sorting process of the algorithm.

The above is true unless you set a mistake weight to zero (0). When you do this, the "perfect creature" reported by the program is no longer actually perfect, because it contains missing and/or spurious bindings at the time perfection is reported. This is due to the fact that when the mistake weight is zero, the algorithm no longer detects the presence of that mistake.

Setting a mistake weight to zero reduces the number of selection conditions. A “perfect creature” is simply a creature that has satisfied the sorting conditions applied to the population.

Also, whenever a mistake weight is set to zero, Rseq never converges on Rfreq, clearly demonstrating that the genome being created does not model any known independently living life form, because all such life forms can be shown to have genomes where Rseq ~ Rfreq. I.e., ev without convergence, proves nothing about evolution.

Rseq ~ Rfreq is not a selection condition. Selection is based on mistakes identified by the selection conditions. I have many examples where Rseq ~ Rfreq yet you don’t have a “perfect creature” and visa versa. Kjkent1 is trying to make ev something which it isn’t. Ev is simply a sorting algorithm based on selection conditions and what ev shows is that it is far easier for the algorithm to sort based on a single selection condition than sorting on all three conditions simultaneously. This in fact is how mutation and selection works in reality.

Now, here's some questions that you can't answer without modifying the program:

1. The mistake weights can only be set from 1 to 100. There is no way to model what would occur were a mistake weight several magnitudes greater than some other.

Invite Paul to make such a change. It shouldn’t take much programming, perhaps changing a byte integer to a long integer and then increasing the limits on his entry field.

2. How do we know that a creature with a small genome that is already generally perfect (Rseq ~ Rfreq) does not mutate by replication, for example, and then develop a new function without having to repeat a long iteration of fitness trials so as to cause the genome to converge again on perfection?

Kjkent1, you are arguing that evolution does not take a series of small steps but somehow makes large step changes. That is you don’t need a series of microevolutionary steps to achieve a macroevolutionary change, you just get a macroevolutionary change in one fell swoop. There is no mathematical or empirical basis for such a concept. This is raw speculation.

How do we know that the same cannot be said for a frame shift, a fusion, a deletion, a transposition, etc?

We know that frame shifts occur with HIV and HBV yet combination selection pressures still profoundly slow the evolution of these viruses. It is not the type of mutations which dominates the mutation and selection sorting/optimization process; it is the number of selection conditions which dominates the process. The mutation and selection sorting/optimization process is mathematically predictable and empirically measurable and it simply does not work the way evolutionists allege.

The early Earth had plenty of time to produce a small perfect creature, which could then act as a springboard for future and much faster evolutionary change.

Once upon a time…

Bottom line, kleinman, old boy, we only really need one self-replicator with a genome of any size wherein Rseq ~ Rfreq. As soon as this occurs, evolution is off and running.

That explains why the world is filled with tribbles.

And, while you can contend that the odds of this occurring are too remote to be possible, the fact is that we are here typing about it, and the evidence of common descent is overwhelming.

A child of five would understand this -- send someone to fetch a child of five!

Therefore, that's what happened, unless you want to claim that God or some extraterrestrial set the ball rolling. Which you certainly can claim -- but the odds in favor of that are infinitely worse than random chance.

A mentsh tracht und Gott lacht.

While kjkent1 is dreaming of tribbles, let’s see what else we can learn from the mathematics of ev.

What happens if we make a cut in our 6-dimensional surface in the g, gfc plane?

G=4096, population=64, mr=1 mutation per genome per generation, nsp=3:
g|gfc to perfect creature
2|140441
3|55585
4|52475
5|65441
6|26551
7|56496
8|79879
9|32391
10|59240
11|10684
12|53572
13|30162
14|35564
15|54498
16|42641
17|61064
18|54802
19|31806
20|45631
40|24544
80|28916
160|34389
200|28826

It appears from this series that g has little effect on gfc, the generations for convergence. Any of you want to hazard a guess why? Tomorrow we will look at the effects of population on the rate of convergence of ev.

 

[kjkent1]

It should be obvious to you why setting all the mistake weights to the same number does not alter the sorting process of the algorithm.

It's obvious to me, but it may not be obvious to other readers.

Setting a mistake weight to zero reduces the number of selection conditions. A “perfect creature” is simply a creature that has satisfied the sorting conditions applied to the population.

Rseq ~ Rfreq is not a selection condition. Selection is based on mistakes identified by the selection conditions. I have many examples where Rseq ~ Rfreq yet you don’t have a “perfect creature” and visa versa. Kjkent1 is trying to make ev something which it isn’t. Ev is simply a sorting algorithm based on selection conditions and what ev shows is that it is far easier for the algorithm to sort based on a single selection condition than sorting on all three conditions simultaneously. This in fact is how mutation and selection works in reality.

kleinman doesn't know what he's talking about. The relationship between/convergence of Rseq ~ Rfreq is one of the most important components of "ev" and Schneider's paper. Kleinman doesn't like this particular fact because it wreaks his entire computational theory. But, I'll repeat it so it doesn't get lost: a genome where Rseq is not ~ Rfreq is likely not a living organism; therefore any result from ev that does not show this relationship has nothing to do with reality.

Kjkent1, you are arguing that evolution does not take a series of small steps but somehow makes large step changes. That is you don’t need a series of microevolutionary steps to achieve a macroevolutionary change, you just get a macroevolutionary change in one fell swoop. There is no mathematical or empirical basis for such a concept. This is raw speculation.

It is? You need to get out more. Mutations with new functionality occur all the time. Most of them don't add any fitness benefit to the environment. But some do. Ask Daniel Tammet how he manages to memorize Pi to 24,000 places. Surely you wouldn't suggest that he's just got an extra good memory? And, what about the pesky 2p and 2q chromosomes found in non-human primates -- but fused into one chromosome in humans? I'd say that was a pretty gross macroevolutionary change, wouldn't you?

No, of course not. You would ignore it and say it's not even empirical data, because it messes with your theory. Well, you're theory is crap, Alan. It doesn't comport with reality.

We know that frame shifts occur with HIV and HBV yet combination selection pressures still profoundly slow the evolution of these viruses. It is not the type of mutations which dominates the mutation and selection sorting/optimization process; it is the number of selection conditions which dominates the process. The mutation and selection sorting/optimization process is mathematically predictable and empirically measurable and it simply does not work the way evolutionists allege.

Now THAT was total speculation. I've previously posted a peer-reviewed paper that showed a frame shift which provided immunoprophylaxis escape for an HBV variant while under combination therapeutic pressure -- so you're statement above is absolutely falsified.

Once upon a time…

...there was a physician in Clovis, Ca whose obsessive-compulsive behavior makes it impossible for him to maintain a relationship with anyone, because his low self-esteem prevents him from admitting that he could ever be in error about anything...the end.

 

[Mister Earl]

I'll repeat it so it doesn't get lost: a genome where Rseq is not ~ Rfreq is likely not a living organism; therefore any result from ev that does not show this relationship has nothing to do with reality.

Good luck. I've been barking up that tree for days. I pointed out that the paper he clings to also clearly states that it would have to factor in a great number of additional variables to accurately model realistic evolution. He ignores this part, of course. And every time I mention it, I get an immediate response from everyone's favorite nutbag, saying "But but but look at the MATH!", followed by a good number of posts containing no math whatsoever.

Bottom line, I don't think Kleinman is in any kind of mental state that would facilitate an opinion change. This latest thread direction is entirely for his own benefit. It is his intention that he sway our opinions. I think he's going to have to do a better job... so far I haven't seen him post anything that even gives me the shadow of a doubt that evolution doesn't happen they way present theory states. The arrogant tone doesn't exactly help, either...

 

[Mister Earl]

So just to summarize, Kleinman, I will no longer try to change your opinion, because I no longer believe that is possible. I will, however, patiently wait as you continue to present data and see if you can prove anything that'll sway me.

 

[rocketdodger]

What results has rocketdodger presented? Rocketdodger spent an afternoon throwing together some code and then claims that n+1 selection pressures evolve more rapidly than n selection pressures. I’ll stick with Dr Schneider’s peer reviewed and published computer simulation that has been scrutinized for more than 20 years and for which Dr Schneider has published and posted extensive documentation, especially since the mathematical results from Dr Schneider’s model is reflected in real empirical examples of mutation and selection.

He is talking about the post I made that shows how stupid you are when it comes to interpreting results from ev.

Tell us how these facts fit into your theory, Kleinman:

When the mistake weights are [1,0,0], [0,1,0], or [0,0,1] the simulation never stops on Rseq >= Rfreq (which, by definition, means these weights are invalid).

When the mistake weights are [0,1,0] or [0,0,1] the simulation evolves a perfect creature in one generation (which, by definition, is impossible -- thus these weights are invalid).

When the mistake weights are [100,100,100], convergence happens faster than [1,100,1] or [1,1,100] and convergence has a higher rate than [100,1,1].

 

[Mister Earl]

I wonder if Kleinman moved to a new thread, or if he's running additional simulations. It's out of character for him to not comment in this amount of time.

 

[sol invictus]

Some people in another thread requested to see my estimate of the scaling of the generations required for convergence to a perfect creature with the number of selection pressures.

Suppose we have a population of creatures, and we'll suppose for simplicity that they all start off with the same genes. They reproduce asexually, and each gives birth to one child per generation. So the population can be divided into a bunch of independent lines of descent, and we want to know how long we have to wait, on average, for a given line to mutate in N specific beneficial ways. To model selection pressure we can simply assume that once any of those beneficial mutations occurs in a particular line it is fixed from then on (so it never un-mutates). This reduces the problem to a rather trivial combinatorics exercise. (It's the same question as: if you buy tickets for N lotteries, all drawn simultaneously at regular intervals, how many rounds will you have to play on average before you've won all the lotteries at least once?)

Suppose for a moment we only cared about one gene, and the probability per generation that gene will mutate in the way we're interested in is p. Then the number of generations it takes for the mutation to occur is on average simply t_avg = 1/p.

Now suppose there are N mutations we want to keep track of, and for simplicity assume that each has that same probability p to occur in any given generation. It's true that the chance for them all to occur in a single generation is very small: p^N. But what's the average time it takes for all N mutations to occur?

That's slightly more complicated, but the answer is the following: the chance that after t generations a given mutation has not occurred is (1-p)^t. So that chance it has occurred is 1-(1-p)^t. Then the chance that all N have occurred is (1-(1-p)^t)^N. We can use that to compute the average time exactly, or make an estimate.

For N large, the answer is approximately t_avg ~ (1/p)(d + log[N]), where d is a number of order 1. So the time grows very slowly with N at large N, and the average time per mutation decreases at large N as log[N]/(pN). If p = 1/1,000,000 and N = 100, the average number of generations for all 100 mutations to occur is only 5.2 10^6, or 5.2/p.

So in this model, the rate of evolution per pressure INCREASES with the number of pressures as N/log(N), contrary to kleinman's repeated assertions that "evolution is profoundly slowed by multiple pressures". Furthermore this simple model seems to coincide with the results of at least two somewhat more sophisticated computer models, one being ev and the other being something rocketdodger coded a while back.

This type of scaling is the basic reason why so many mutations have occurred over the course of evolution, even though the chance of them all occurring in any one generation is vanishingly small.

 

[kjkent1]

When the mistake weights are [0,1,0] or [0,0,1] the simulation evolves a perfect creature in one generation (which, by definition, is impossible -- thus these weights are invalid).

The funniest thing about this particular outcome is that if we were to accept it at face value, then ev proves that ANY arrangement of amino acids that accidentally occurs in nature will spontaneously bind together and start the cycle of organic life!

Thus, kleinman destroys his entire theory by conceding that ev proves abiogenesis.

I doubt that Schneider would agree with this ridiculous outcome from his algorithm, but hey, it's a "valid" ev result!

 

[Paul C. Anagnostopoulos]

I doubt that Schneider would agree with this ridiculous outcome from his algorithm, but hey, it's a "valid" ev result!

I guess this means that a person has to think about what is actually going on in Ev, in order to assign a valid interpretation to these fringe results. He can't just say "Oh look, viable creatures from only one selection pressure!"

~~ Paul

 

[m_huber]

I guess this means that a person has to think about what is actually going on in Ev, in order to assign a valid interpretation to these fringe results. He can't just say "Oh look, viable creatures from only one selection pressure!"

~~ Paul

Isn't that the point of this whole discussion?

 

[kleinman]

I apologize for not responding in a more timely manner but I'll back with you on Monday. You all enjoy your weekend.

 

[Mister Earl]

I hope you enjoyed yours. I spent mine getting bitten by a ravenous cockatiel in my sleep. Ended up cooking some spaghetti, just to placate the bird. Spaghetti > Birdseed apparently.

 

[kleinman]

Annoying Creationists

Have any of you mathematically incompetent evolutionists learned anything about the mathematics of the mutation and selection sorting/optimization process yet? Let’s find out.

It should be obvious to you why setting all the mistake weights to the same number does not alter the sorting process of the algorithm.

It's obvious to me, but it may not be obvious to other readers.

Well then, why don’t you explain it to the readers?

Setting a mistake weight to zero reduces the number of selection conditions. A “perfect creature” is simply a creature that has satisfied the sorting conditions applied to the population.

Rseq ~ Rfreq is not a selection condition. Selection is based on mistakes identified by the selection conditions. I have many examples where Rseq ~ Rfreq yet you don’t have a “perfect creature” and visa versa. Kjkent1 is trying to make ev something which it isn’t. Ev is simply a sorting algorithm based on selection conditions and what ev shows is that it is far easier for the algorithm to sort based on a single selection condition than sorting on all three conditions simultaneously. This in fact is how mutation and selection works in reality.

kleinman doesn't know what he's talking about. The relationship between/convergence of Rseq ~ Rfreq is one of the most important components of "ev" and Schneider's paper. Kleinman doesn't like this particular fact because it wreaks his entire computational theory. But, I'll repeat it so it doesn't get lost: a genome where Rseq is not ~ Rfreq is likely not a living organism; therefore any result from ev that does not show this relationship has nothing to do with reality.

Really, legal beagle? Do you want to explain how to analyze and solve an implicit functional equation? Dr Schneider was surprised when his model showed the peculiar result that Rseq ~ Rfreq. Dr Schneider still hasn’t given a good explanation how this relates to the mutation and selection sorting/optimization process. Because Dr Schneider chose to focus on this peculiarity, he missed the obvious result that his model shows, that is that combination selection conditions profoundly slow evolution by the mutation and selection sorting/optimization process. If Dr Schneider had done the parametric study he called for in his publication, he would have discovered this result himself.

Kjkent1, you are arguing that evolution does not take a series of small steps but somehow makes large step changes. That is you don’t need a series of microevolutionary steps to achieve a macroevolutionary change, you just get a macroevolutionary change in one fell swoop. There is no mathematical or empirical basis for such a concept. This is raw speculation.

It is? You need to get out more. Mutations with new functionality occur all the time. Most of them don't add any fitness benefit to the environment. But some do. Ask Daniel Tammet how he manages to memorize Pi to 24,000 places. Surely you wouldn't suggest that he's just got an extra good memory? And, what about the pesky 2p and 2q chromosomes found in non-human primates -- but fused into one chromosome in humans? I'd say that was a pretty gross macroevolutionary change, wouldn't you?

I don’t argue that polymerases can’t mutate so that they can cleave xeno-molecules like nylon, what I argue is that the evolutionists’ gross over-extrapolation of this microevolutionary process to the evolution of lizards to birds is irrational, illogical and mathematically impossible. The reason I make this argument is based on the mathematics of Dr Schneider’s ev simulation of random point mutations and natural selection which show that combination selection pressures profoundly slow the sorting/optimization process and hundreds of real examples of mutation and selection which demonstrates this same finding. You evolutionists have completely bungled the understanding of the mutation and selection sorting/optimization process.

No, of course not. You would ignore it and say it's not even empirical data, because it messes with your theory. Well, you're theory is crap, Alan. It doesn't comport with reality.

Really? You think the greater the number of selection conditions the faster the evolutionary process proceeds? Where are all you citations which show your contention?

We know that frame shifts occur with HIV and HBV yet combination selection pressures still profoundly slow the evolution of these viruses. It is not the type of mutations which dominates the mutation and selection sorting/optimization process; it is the number of selection conditions which dominates the process. The mutation and selection sorting/optimization process is mathematically predictable and empirically measurable and it simply does not work the way evolutionists allege.

Now THAT was total speculation. I've previously posted a peer-reviewed paper that showed a frame shift which provided immunoprophylaxis escape for an HBV variant while under combination therapeutic pressure -- so you're statement above is absolutely falsified.

Are you contending that combination therapy should not be used on HBV? Are you contending that frame shifts overcomes the mathematical fact the Dr Schneider’s ev program demonstrates that combination selection pressures profoundly slow the mutation and selection sorting/optimization process?

Once upon a time…

...there was a physician in Clovis, Ca whose obsessive-compulsive behavior makes it impossible for him to maintain a relationship with anyone, because his low self-esteem prevents him from admitting that he could ever be in error about anything...the end.

Why legal beagle, how many times have you said good bye on this thread and you keep coming back to this discussion? I think you are the one obsessed with this discussion. You can’t produce the mathematical or empirical data to support your irrational and illogical theory of evolution so you look for any way you can to discredit me. Do you want to talk more about my malpractice case which I won yet you tried to discredit me with that? It so typical of you evolutionists to jump to conclusions when you don’t have the facts to support your irrational and illogical thinking, lizards must evolve into birds because a polymerase can mutate so that it can cleave nylon, perhaps in one of your 10^500 alternative universes but not in this one.

Hey legal beagle, since you are such an expert on Dr Schneider’s ev model, why don’t you tell us why variations in g have such little effect on the generations for convergence?

I'll repeat it so it doesn't get lost: a genome where Rseq is not ~ Rfreq is likely not a living organism; therefore any result from ev that does not show this relationship has nothing to do with reality.

Good luck. I've been barking up that tree for days. I pointed out that the paper he clings to also clearly states that it would have to factor in a great number of additional variables to accurately model realistic evolution. He ignores this part, of course. And every time I mention it, I get an immediate response from everyone's favorite nutbag, saying "But but but look at the MATH!", followed by a good number of posts containing no math whatsoever.

Mister Earl, what mathematical model of mutation and selection are you clinging to? Oh wait, you evolutionists don’t need any mathematical model to prove your irrational and illogical theory. You have your speculations and extrapolations to prove your point. Mister Earl, do you think that increasing the number of selection pressures will accelerate the mutation and selection sorting/optimization process?

Bottom line, I don't think Kleinman is in any kind of mental state that would facilitate an opinion change. This latest thread direction is entirely for his own benefit. It is his intention that he sway our opinions. I think he's going to have to do a better job... so far I haven't seen him post anything that even gives me the shadow of a doubt that evolution doesn't happen they way present theory states. The arrogant tone doesn't exactly help, either...

All you need to do Mister Earl is produce some mathematical or empirical evidence which show the combination selection pressures accelerates the mutation and selection sorting/optimization process. Sadly, you like other evolutionists are mathematically incompetent.

So just to summarize, Kleinman, I will no longer try to change your opinion, because I no longer believe that is possible. I will, however, patiently wait as you continue to present data and see if you can prove anything that'll sway me.

Mister Earl, you never had any mathematical or empirical evidence to change my opinion. You think irrational and illogical speculations and extrapolations constitute a scientific argument. I’ll continue to show the mathematical data from Dr Schneider’s peer reviewed and published mathematical model of mutation and selection which Dr Schneider believes represents the essential variables in this process (and so do I) and the empirical evidence which substantiates the results from Dr Schneider’s model.

What results has rocketdodger presented? Rocketdodger spent an afternoon throwing together some code and then claims that n+1 selection pressures evolve more rapidly than n selection pressures. I’ll stick with Dr Schneider’s peer reviewed and published computer simulation that has been scrutinized for more than 20 years and for which Dr Schneider has published and posted extensive documentation, especially since the mathematical results from Dr Schneider’s model is reflected in real empirical examples of mutation and selection.

He is talking about the post I made that shows how stupid you are when it comes to interpreting results from ev.

Which post was that? Was it this one?

I found a combination of parameters last night that led to over 100 pressures being faster than a single one, but I forgot what it was

Or was it this one where you claim that all fitness landscapes are convex?
http://www.internationalskeptics.com/forums/showpost.php?p=3170363&postcount=6644

Do you even know what a fitness landscape looks like, Kleinman? Here is a hint for you -- regardless of the number of selective pressures, the landscape is always a convex shape. There is no such thing as local optima that the population can "get stuck on" or whatever nonsense it is that you assert.

Or perhaps it was this one where you claim that if you get enough selection conditions that the mutation and selection sorting/optimization process speeds up?
http://www.internationalskeptics.com/forums/showpost.php?p=3126277&postcount=6348

At first, it appeared as though Kleinman might be right -- adding selective pressures in low numbers really does slow the average rate of fixation way down. However, once a "critical mass" has been reached, adding pressures actually drives the rate back up. Eventually, with enough pressures, you get to a point where the average rate of fixation is equal to or faster than it is under only a single pressure.

Maybe it was this quote you made?
http://www.internationalskeptics.com/forums/showpost.php?p=3127246&postcount=6359

Actually, Adequate and I sort of are. My simulation shows that if you introduce enough strong pressures, and if you measure the rate of fixation on average, then fixation is just as fast (or even faster).

Rocketdodger, you claim that n+1 selection pressures evolve more rapidly than n selection pressures but you have no real examples of your ridiculous claims.

Tell us how these facts fit into your theory, Kleinman:

When the mistake weights are [1,0,0], [0,1,0], or [0,0,1] the simulation never stops on Rseq >= Rfreq (which, by definition, means these weights are invalid).

There is no selection pressure for Rseq >=Rfreq. This demonstrates a fundamental misunderstanding you have of Dr Schneider’s ev model and how his selection process works. The best documentation for how Dr Schneider’s selection process works was not written by Dr Schneider but by I.G.D. Strachan and is available at: http://www.iscid.org/papers/Strachan_EvEvaluation_062803.pdf

Here is a pertinent quote from the paper:

For the recognizer gene, Schneider uses a simple form of neural network, called a Perceptron, consisting of a numeric weight vector w, and a threshold value Ɵ. The recognition action is performed by computing the dot product of an input vector x with the weight vector, and subtracting the threshold from it. If the resultant value is greater than zero, then the system is considered to have \recognized" the input, and if it is below zero, then recognition is not deemed to have occurred. Hence the recognition criterion is:

w · x - Ɵ > 0 (2)

The numerical values of the weight vector and the threshold are encoded onto the recognizer gene, by using sequences of 5 bases for each number (as reported in the paper, but the number of bases can be specified in the input file to the program). By encoding A = 00; C = 01; G = 10 and T = 11 in binary, the five bases yield a 10-bit number that represents integers in the range -512 : : : +511.

The inputs to the perceptron are computed by assigning a 1-of-4 code to each of the six bases in the scanned potential site: A -> (1; 0; 0; 0); C -> (0; 1; 0; 0); G -> (0; 0; 1; 0) and T -> (0; 0; 0; 1). For a sequence of six bases, there are therefore 24 inputs to the perceptron, in six groups of 4. These are described in the paper as a “weight matrix”, reflecting the 6 x 4 grouping. Each of the six rows in the matrix can only have a single 1, and the rest are zeros. However, the linear algebra in equation (2) implies no matrix operation, only the vector operation of the dot product (taking the pairwise products of the two vectors and summing the result). Because the inputs to the perceptron in this simulation are always only 0 or 1, the equation reduces to summing the weight matrix row entries corresponding to the base found at each position. The operation is illustrated schematically in Figure 1.

The figure can be viewed by clicking the above link. This paper written by I.G.D. Strachan gives and excellent review of the mathematics which Dr Schneider employed in order to develop his selection scheme and mathematical model.

When the mistake weights are [0,1,0] or [0,0,1] the simulation evolves a perfect creature in one generation (which, by definition, is impossible -- thus these weights are invalid).

No it doesn’t rocketdodger, what this demonstrates is how easy it is for the weight matrix to find matches on the genome and how difficult it is for selection to remove improper matches to the weight matrix in one portion of the genome while maintaining matches in other portions of the genome. This demonstrates how difficult it is to sort multiple conditions simultaneously.

When the mistake weights are [100,100,100], convergence happens faster than [1,100,1] or [1,1,100] and convergence has a higher rate than [100,1,1].

Mistake weights of [100,100,100] give the same results as mistake weights of [1,1,1]. Changing the relative weights changes the shape of the fitness landscape. You remember what the fitness landscape is? That’s the surface you claim is always convex. Nothing affects the fitness landscape like setting two of the three selection conditions to zero. It becomes trivially easily to satisfy the remaining selection condition. Reduce the number of sorting conditions to one and that condition is extremely easy to sort for and satisfy.

Suppose…

Suppose…

So in this model, the rate of evolution per pressure INCREASES with the number of pressures as N/log(N), contrary to kleinman's repeated assertions that "evolution is profoundly slowed by multiple pressures". Furthermore this simple model seems to coincide with the results of at least two somewhat more sophisticated computer models, one being ev and the other being something rocketdodger coded a while back.

Sol, do you still want to lose $10,000? Ev shows nothing of the kind and rocketdodger shows nothing.

When the mistake weights are [0,1,0] or [0,0,1] the simulation evolves a perfect creature in one generation (which, by definition, is impossible -- thus these weights are invalid).

The funniest thing about this particular outcome is that if we were to accept it at face value, then ev proves that ANY arrangement of amino acids that accidentally occurs in nature will spontaneously bind together and start the cycle of organic life!

What it proves by definition is that neither of you have any understanding of the mutation and selection sorting/optimization process which ev is modeling.

Thus, kleinman destroys his entire theory by conceding that ev proves abiogenesis.

Legal beagle, are you looking for another ambulance to chase?

I doubt that Schneider would agree with this ridiculous outcome from his algorithm, but hey, it's a "valid" ev result!

Who knows what Dr Schneider thinks of his algorithm anymore? Perhaps he has changed fields and now is a noodleologist?

I doubt that Schneider would agree with this ridiculous outcome from his algorithm, but hey, it's a "valid" ev result!

I guess this means that a person has to think about what is actually going on in Ev, in order to assign a valid interpretation to these fringe results. He can't just say "Oh look, viable creatures from only one selection pressure!"

Well, well Mr RcaPaulcity, are you going to claim that evolution of resistance from single drug therapy does not give viable creatures. I have posted hundreds of real examples of mutation and selection which shows that combination selection pressures profoundly slow the mutation and selection sorting/optimization process and your own algorithm shows the same thing. What real example of Rcapcity have you shown?

Are you now going to claim that ev is evolving real viable creatures? Paul, you’ve earned this:

I guess this means that a person has to think about what is actually going on in Ev, in order to assign a valid interpretation to these fringe results. He can't just say "Oh look, viable creatures from only one selection pressure!"

Isn't that the point of this whole discussion?

M_huber, this is the point of this discussion. Ev does demonstrate something about the mathematics of the mutation and selection sorting/optimization process. Dr Schneider and other evolutionists missed the fundamental principles of the mathematical behavior on this phenomenon. Dr Schneider took a peculiar finding from his particular model and chose to focus on this and missed the fundamental nature of this sorting/optimization process. This is an example how the evolutionist bias has interfered and continues to interfere with the proper mathematical and empirical interpretation of the mutation and selection sorting/optimization process.

I hope you enjoyed yours. I spent mine getting bitten by a ravenous cockatiel in my sleep. Ended up cooking some spaghetti, just to placate the bird. Spaghetti > Birdseed apparently.

My weekend was better than Friday; I mixed up my hair spray and underarm deodorant spray. My hair smelled ok but I couldn’t lift my arms for a day.

Hey Mister Earl, isn’t it dangerous keeping a carnivorous cockatiel around the house?

Before I go on and give the data for varying population and it’s affect on the generations for convergence in ev, do any of you experts on ev want to explain why varying g has little affect on the generations for convergence in Dr Schneider’s model?

Do not use personal attacks to argue your point. This is your last warning before suspension.

Replying to this modbox in thread will be off topic  Posted By: Lisa Simpson

 

[Mister Earl]

[Have any of you mathematically incompetent evolutionists learned anything about the mathematics of the mutation and selection sorting/optimization process yet? ... Really, legal beagle?

I was hoping your brief vacation would see you returning slightly more civil. I was wrong, I guess.

I don’t argue that polymerases can’t mutate so that they can cleave xeno-molecules like nylon, what I argue is that the evolutionists’ gross over-extrapolation of this microevolutionary process to the evolution of lizards to birds is irrational, illogical and mathematically impossible.

I suppose I could ask you to prove that it is mathematically impossible, but I'm pretty certain I'd just get the same rehashed statements from before.

The reason I make this argument is based on the mathematics of Dr Schneider’s ev simulation of random point mutations and natural selection which show that combination selection pressures profoundly slow the sorting/optimization process and hundreds of real examples of mutation and selection which demonstrates this same finding. You evolutionists have completely bungled the understanding of the mutation and selection sorting/optimization process.

I wonder what Dr. Schneider himself would think. I'm going to try to contact him via email, and bring him in on this discussion.

Really? You think the greater the number of selection conditions the faster the evolutionary process proceeds? Where are all you citations which show your contention?

Rocketdodger posted this point in a much more fluent way than I could. You may want to review his "Kleinman FAQ".

Are you contending that frame shifts overcomes the mathematical fact the Dr Schneider’s ev program demonstrates that combination selection pressures profoundly slow the mutation and selection sorting/optimization process?

Why is it your entire argument resides on "This is a mathematical fact" when you've never proven this? We've shown you why EV isn't a perfect simulator for biological evolution. I see no need to rehash this.

Why legal beagle, how many times have you said good bye on this thread and you keep coming back to this discussion? I think you are the one obsessed with this discussion. You can’t produce the mathematical or empirical data to support your irrational and illogical theory of evolution so you look for any way you can to discredit me. Do you want to talk more about my malpractice case which I won yet you tried to discredit me with that? It so typical of you evolutionists to jump to conclusions when you don’t have the facts to support your irrational and illogical thinking, lizards must evolve into birds because a polymerase can mutate so that it can cleave nylon, perhaps in one of your 10^500 alternative universes but not in this one.

Personal attacks prove nothing. Neither against you, Kleinman, or by you.

Hey legal beagle, since you are such an expert on Dr Schneider’s ev model, why don’t you tell us why variations in have such little effect on the generations for convergence?

Rocketdodgers model showed different results than yours.

Mister Earl, what mathematical model of mutation and selection are you clinging to? Oh wait, you evolutionists don’t need any mathematical model to prove your irrational and illogical theory. You have your speculations and extrapolations to prove your point. Mister Earl, do you think that increasing the number of selection pressures will accelerate the mutation and selection sorting/optimization process?

I've got a simulator at home called "Framsticks". It's pretty nice. I've run a multitude of simulations on it, where I've cranked up the selection pressures to obscene levels. One of two things happen: First, complete extinction. Second, a multitude of genomes with superior fitness levels. Unlike EV, you can configure framsticks to utilize population levels and multiple, varying, selection pressures. I've run the same simulation with only one selection pressure, and at a low level, and those genomes only reach about 15% of the fitness level of the previous test in the same number of generations. If you'd like, you can download Framsticks, and I can send you the files to replicate my tests. Would you like to do this? If not, why not?

All you need to do Mister Earl is produce some mathematical or empirical evidence which show the combination selection pressures accelerates the mutation and selection sorting/optimization process. Sadly, you like other evolutionists are mathematically incompetent.

More ad hominem attacks. Spare me your egotistical nonsense. Either give us definitive proof, or go cry somewhere else. Your pointless bluster doesn't cover up the fact that your argument has very little weight.

Mister Earl, you never had any mathematical or empirical evidence to change my opinion. You think irrational and illogical speculations and extrapolations constitute a scientific argument. I’ll continue to show the mathematical data from Dr Schneider’s peer reviewed and published mathematical model of mutation and selection which Dr Schneider believes represents the essential variables in this process (and so do I) and the empirical evidence which substantiates the results from Dr Schneider’s model.

Grab framsticks, and when I get home tonight I'll email you the two trial files. Then you can come back tomorrow and offer your point of view how my simulations aren't valid when you don't get the results you want.

Who knows what Dr Schneider thinks of his algorithm anymore? Perhaps he has changed fields and now is a noodleologist?

I'll email him and try to get him to join us.

My weekend was better than Friday; I mixed up my hair spray and underarm deodorant spray. My hair smelled ok but I couldn’t lift my arms for a day. Hey Mister Earl, isn’t it dangerous keeping a carnivorous cockatiel around the house?

Very much so. We've had him for fifteen years, so I think he's just getting old and ornery. When he's in a good mood, it offsets a great deal of the biting. I just take his actions into context. (Old, ornery bird).

 

[kjkent1]

Have any of you mathematically incompetent evolutionists learned anything about the mathematics of the mutation and selection sorting/optimization process yet? Let’s find out...(clip)

kleinman, your reply to my posts and to others clearly demonstrates why you cannot successfully discuss your theory with anyone who doesn't already agree with you.

Your comments are entirely non-responsive in every instance. You don't address your opponents' arguments -- you ignore them and restate you own position -- or, you request that your opponent prove something which is completely irrelevant to the argument at the time that you raise the issue.

Thus, there's little point in my responding to each of your individual points.

I will say this: you concede that polymerases can "mutate so that they can cleave xeno-molecules like nylon," yet you dismiss as irrational that those same mutations cannot aggregate to create a morphologically distinct organism over time.

Well, there is a really long fossil record showing that what you say is irrational, actually occurred. Now you can moan all day long that it's mathematically impossible. But, if, as above, you concede that a gross mutational change can occur in the K172 bacteria, then you cannot avoid the evidence in the fossil record that such changes have actually occurred over the ages.

You can, of course, cling to the position that Schneider's math shows that creating a perfect creature is impossibly slow, but that's irrelevant, because you have long conceded that (1) Rseq ~ Rfreq is meaningless, and (2) that ev doesn't model any gross mutational behaviors. In the first case, without convergence, there is no means of measuring whether or not a creature has evolved, and no means of identifying a potential life form from a dead string of amino acids. I.e., how do you know that a creature filled with missing and/or spurious bindings cannot reproduce? Apparently, since you dismiss the relationship of Rseq and Rfreq, you cannot know anything about the target organisms, other than their slowness in removing mistakes.

In the second case, the target organism could spend a very long time without much substantive change, and then suddenly change quite rapidly, due to an unexpected and beneficial mutation. Now, you can certainly argue that this is contrary to some "gradualist" (point mutation) theory of evolution -- but, it's not contrary to the general theory of evolution by mutation and selection -- it fits just fine.

What this leaves you with is one argument: Schneider's algorithm doesn't demonstrate the entire process of evolution. Well, not even Schneider himself contends that to be true.

So, what ARE you arguing, kleinman? Of course, I already know the answer: Genesis 1:1.

But, the relevant issue is that you have yet to prove Genesis 1:1, and you will not do so by showing that evolution is not purely the product of point mutation.