Signal and Noise


Well my point was straightforward. I made the observation that the CI values often are actually closer to the random model than to 1.0 (the ideal common descent model), and that this better fits the design model than the common descent model. As I explained, design trends emerge simply due to the constraints of natural laws, the same planet, biosphere, etc. Or as Fernald put it:

Although the variety of eyes in the animal kingdom seems astonishing, physical laws have constrained solutions for collecting and focusing light to just eight types of eye optics.

You said that the consistency index data “on the face of it, it offers strong evidence for common descent.” Actually, if anything, on the face of it the data are evidence for design.

Of course, we all understand common descent can explain the evidence. But that is because common descent can explain practically any evidence. If CI ~= 1.0, common descent could explain that. Common descent can explain a wide range of outcomes, but that is not a virtue when it comes to theory evaluation.

How could you see this as strong evidence for common descent? What this points out is how far data can be from the model and (i) not only do no harm, but (ii) actually be greeted as “strong” evidence. This is not a virtue.

For instance, let’s consider this in Bayesian terms, in order to see what evolutionists are claiming here. In order for these observed CI values to be “strong” evidence, the probability of observation (low CI values) given the hypothesis of common descent, P(O|H), would need to be very high. In other words, on common descent, the CI value is very strongly expected to be down there closer to random than to 1.0. This means that for larger CI values, as it gets closer to 1.0, P(O|H) necessarily decreases significantly, and becomes pretty small.

But this clearly is not the case. On common descent, this simply does not happen. High CI values would do no harm, and in fact would be closer to the ideal common descent model. Evolutionists would of course have no problem with such findings, and certainly would interpret it more positively. P(O|H) would increase, not decrease.

The fact is, this is not good evidence for common descent, and is a better fit for design, as I have explained. So here is a notional illustration of this. For a given #taxa, you have a “random” CI band, and the observed CI values above that, and then way up higher the ideal common descent model. Evolutionists have for years complained “why would God make it appear to have evolved?!” Well here is the answer: it doesn’t. Clearly the CI values, by the evolutionist’s own reckoning, are a better fit to design than to common descent.

Clearly the CI values are not “strong” evidence for common descent. You can always claim that common descent is capable of explaining the evidence, but that is a very different claim. Chris lectures me on evidence interpretation, but perhaps you guys need to think a bit harder about this. Agreement is probably too much to expect, but at least to acknowledge, “OK, this is their position—this is why they say what they say, and this is the strength of their argument”, rather than straw-manning it.

When I criticize evolution, I first want to understand it in all its strength. I want to understand why evolutionists say what they say. And who knows, perhaps that exercise will change my mind. But it will do no good if I’m merely knocking down strawmen. I’m gratified when evolutionists say “I disagree with your conclusion, but I appreciate that you have represented my position in its strength.”

Instead Chris has taken the low road. He has gone the extra mile, putting words in my mouth to make it appear I am making basic, high school math mistakes and rendering my argument as absurd. I shudder to think what is said about me when I am not here to read the posts.


I like this post, especially your graph. It is very elucidating.

Can you explain how you derive it? I have several key questions about it:

  1. How did you arrive at the conclusion that Design should have less consistency (at about 0.6) than common descent (at about 0.8)? Walter ReMine (a YEC that rejects evolution) argues exactly the opposite, that design should produces more consistency than common descent (he argues it should be close to 1). Which of you is right? And how do we know?

  2. If possible (though I understand this is not likely), can you please show your math. How did you arrive at such precise numbers, with non-overlapping uncertainty?

  3. You also point out that you are using an “idealized” common descent model (which presumably does not model any noise). So why should we trust it? Don’t we want a realistic common descent model that includes noise, and would therefore reduce the expected CI?


Why are you blaming me for interpreting your words in the way that mathematically literate English speakers would normally understand them? You seem to be implying that I am remiss for not being able to read your mind and understand that the words you wrote were not to be interpreted according to standard English usage.

Wouldn’t it be better to say, “I realize now that what I wrote was unclear. Let me speak more clearly now”? That is almost verbatim what I said in a different thread last week.


Because the data is highly inconsistent with the null hypothesis, and the null hypothesis seems to be the only available model with which we can compare the actual distribution.[quote=“Cornelius_Hunter, post:14, topic:35048”]
What this points out is how far data can be from the model and (i) not only do no harm, but (ii) actually be greeted as “strong” evidence. This is not a virtue.
It’s neither a virtue nor a vice. It’s just a fact that noisy data can provide strong support for a hypothesis.[quote=“Cornelius_Hunter, post:14, topic:35048”]
For instance, let’s consider this in Bayesian terms, in order to see what evolutionists are claiming here. In order for these observed CI values to be “strong” evidence, the probability of observation (low CI values) given the hypothesis of common descent, P(O|H), would need to be very high.
Not at all. The Bayes factor is the ratio of the likelihoods(*) of the data for two models. The likelihood of the data given common descent is small, but much larger than the likelihood of the data given the null hypothesis. (Note that in Bayesian or maximum likelihood model testing, the absolute likelihoods can be very small indeed, a fact that is irrelevant to choosing the better model.) That’s why I told you before that you had to offer an alternative model that did a better job of explaining the data.

(*) More properly, the likelihood of the data integrated over model parameters.

As @Swamidass pointed out, you haven’t justified your distribution under the design model at all.


First rule of responding to paper reviewers: if the reviewer completely misread your crystal clear prose in the most idiotic way possible, it’s still your fault. If you’re unable to communicate successfully with a reviewer, then you’re failing in your job as paper-writer.


Very inspired by your diagram @Cornelius_Hunter! Here is my version…

A key variable you are missing is the “speed” at which a features evolves (which is often directly measurable). Depending on this parameter, common descent will make different predictions. This isn’t cheating, because we can directly measure the speed in many cases (especially with DNA), and this gives us a way of analytically computing the expectation.

Any how, I think (in my opinion), ReMine’s model corresponds to the blue line, the features used in this analysis correspond to a mixture of slow and fast evolving features, so in this case CD corresponds to the green line. And your model corresponds to the orange line, because it is not specified at all.

So the data ends up fitting CD the best. I should point out that I believe in design too; design by common descent. So in no way does this rule out that God created us. Rather, it tells us that common descent is the best design principle to understand biology.



Well you are going from bad to worse here Steve. The null hypothesis is generated using a random number generator to scramble the data randomly. In other words, the null hypothesis is physically meaningless. It completely disregards the science. There is zero science accounted for. It tosses out the biology, physics, ecology, etc. What you are claiming here is that either the world is utterly chaotic and random, or else common descent is true. Any species can take on any character trait, at anytime, anywhere. There is no context. This is fact-free science. And this is how you judge these data to be “strong evidence for common descent.” Either a physically meaningless reality obtains, or else common descent it true. This is utterly pathetic and ridiculous. And you guys lecture me on philosophy?

Your reasoning is circular. You characterize the data as “noisy” because it does not fit your model. Your interpretation is “theory-laden.”


This is truly astonishing. Not only did I justify it, but that justification, as I indicated, runs all through the evolution literature! To say that I haven’t justified my notional design distribution model at all is not only false, but an incredible example of having it both ways. This is what evolutionists have been insisting all along!! It goes all the way back to Darwin. I mean, you gotta love evolution. Just so readers can fully appreciate the humor here, I’ll paste in what I said:

So I haven’t justified the design model at all?

What is so ironic here is that evolutionists have been pounding this message home all the way back to Darwin: “Design and creationism aren’t constrained to the common descent pattern, but we are!!” they have repeatedly proclaimed. And now, suddenly it’s “what ever could you be talking about? What a strange notion this is.”

Sigh. This is why these discussions become so tedious. So, here we go:

And Chris’s incredible strawman:

As is obvious, that is a misrepresentation of my point. And here is Steve’s response:


As for Remine, his model is not mainstream or consensus. That is not to dismiss it, it is just a statement of fact. One can find all kind of outlier ideas in evolutionary thought as well, but I would never use them in a rebuttal or criticism of evolution. There’s no question, for objective observers, that the consensus views are that the common descent model calls for a higher CI than the design model.

1 Like

It is correct that ReMines’ model is not mainstream. But to be clear, I am not saying that “because he is wrong, so are you.” I bring him up to point out that “design” is a broad concept that includes several models. Some of these models make entirely different predictions than you are making. I even was sure to point out that your model produces different predictions than his (blue vs orange curve).

And, I have already sketched out the principles behind CD. It depends on another variable: the rate of change. It is fairly easy to see how this enables us to compute the expected range of CI we can expect for a given feature.

At this point, however, you have not specified your model. Please explain to us: what is your model of design? Clearly, it is not ReMine’s model, but what is it exactly? And how does it give rise to your precise CI predictions with narrow standard deviation? What is the math/principles/calculations you use to arrive at the expected consistency? Just saying “design” is very poorly specified, which is why you ended up on the orange curve.


EDIT: To the best of my ability, removed any discussion of disputed terminology, etc.

You allude to new evidence, the testimony of “objective observers.” In my web searches, I was unable to identify any such folk who held the opinion you ascribe to them. It is quite possible I overlooked them. Who are these people? What are the analytic methods they used to assign quantitative CI predictions to design models vs. common descent models? Where have they published their analysis?

I would like to understand the issue better, so I would appreciate any details you could offer. Thanks.


Well how gracious of you. Am I supposed to be thankful for this undue generosity. :slight_smile:

So let’s look at your strawman. You said:

But of course, I never said that. I didn’t imply it, I didn’t suggest it, I didn’t say it at all.

And I never said that either. “Randomness hypothesis is victorious” ?? You are contriving things whole cloth.

And, again, I never said that. “the null hypothesis is assumed” ??

That is why I pasted in what I, in fact, really did say. There’s no mystery here. So at the risk of extreme tedium, here it is again:

As any objective observer can see, I never said the bizarre version that Chris came up with, complete with detailed charts and graphs!

Yet when I point out this objectively undeniable fact out, I am told that it is I who am doing the misrepresenting:

As you can see, being an evolutionist mean never having to say your sorry. You gotta love it. Perhaps I should feel honored to be given so much attention, in spite of the fact it is all false. And of course, I was offered “a graceful way out.”

Where would we be without some humor. But kidding aside, this is unfortunately how these discussions often (de) evolve.

Now that the air is cleared, here’s a subject that I think many readers in this thread, not just myself, would find interesting. If you have the time to provide this requested information, I would love to hear what you have to say:


I don’t see a point to the generalities being exchanged here. Let me be concrete. If we compare genomes of two great apes, based on common descent, I can predict the following with considerable precision: the transition/transversion ratio of the single-base differences between the genomes; the relative rate of transitions at CpG sites, the relative rate of transversions at CpG sites. What predictions does your model make for these quantities, to say, +/- 20%?


Forget justification, you haven’t even described it verbally or mathematically.


Hi Cornelius,

In order to represent your discussion as fairly as I possibly can, I have edited a couple of previous posts in order to eliminate any ascription to you of null hypotheses or randomness–other than the reference to randomness that you yourself made. I also adjusted my description of what CI you think evolution would predict from 95% to 80% (not that I could have known your specific number at the time of my original post). Finally, in one post I clipped a large block of text just to reduce the noise and friction.

“[A]s far as it depends on you, live at peace with everyone.” - Romans 12:18

Regarding how to measure the validity of an evolutionary explanation for consistency index: you publicly and quite clearly proposed an arithmetic mean. Thus I feel that my analysis on that issue treats what you wrote with 100% honesty and fairness, and needs to be kept in place so that readers can be aware of the mathematical thinking that points to the best inference from data in my opinion. If you feel that my mathematical formulations are incorrect, I invite you to provide an alternative explanation with a clearly defined, superior mathematical basis.

Best regards,

Chris Falter


you really should stop digging

A “scientific” theory is a well-supported hypothesis, not a “Law of God.” Scientists earn Nobel Prizes for falsifying scientific theories. In other words, a scientific theory is the latest best guess.


I am not surprised that @Cornelius_Hunter never responded to your question…

I wonder what @Cornelius_Hunter thinks about the hyper-speciation that has to occur when the animals are released from the ark … in order to have a an Earth filled with all the different kinds it is filled with …

George, more than once it has been mentioned that Cornelius Hunter believes in an old earth. He also believes in the general chronology presented by paleontology. Thus the ark scenario that you present does not apply to him at all.

Sooner or later, George, you are going to accept that ID and YEC are not synonymous. Probably the forum editors have a lunch tab on the line regarding when that will happen. Before September 2017, @jpm pays for the monthly run to the Chinese buffet. After that, @BradKramer pays. So far, things are looking good for @jpm, but you might surprise us. :smile:


Sure. My stance would be that the “laws” of science are approximations of the laws of God with respect to matter and energy.