I suppose by this point, we’re not getting any math. 
1 Like
Hi E
The math is not hard to generate once the model parameters are agreed upon.
Notice that Bill’s supposed “null model” isn’t a model at all. It can’t be rejected, since it has no predicted parameter values. The way he’s proposed setting up the “test”, evolution is the actual null model which, if you can reject it, supposedly implies separate creation. Obviously, he doesn’t know that, but that’s what it is. Still, even this approach is flawed, because rejection of the null model shows only that the model of evolution is wrong in some way. Conceivably, it’s wrong because there is no evolution, but that seems the least reasonable possibility. More reasonably, one or more of the model parameters is incorrect, or would be if he actually had any parameters or any model or any test.
2 Likes
Hi John
It can be rejected by showing a high probability that gene duplication and associated variation can account for the variation we are observing between the species we are testing.
It also maybe wrong because evolutions contribution to observed genetic variation is limited.
This is what you previously stated:
What you stated implied that math has already been attempted/done.
If you’re saying the math hasn’t been done, then your previous claim has no basis. You don’t draw a conclusion on the outcome of testing a model before building the model and doing the tests.
2 Likes
Then pick some model parameters and generate the maths.
(This is a rhetorical request. I know you have no idea how to generate the maths and wouldn’t be capable of it even if you did.)
2 Likes
Nope. That’s not how you reject a null model. That would be failing to reject the null model of evolution. You have it exactly backwards. A statistical test operates by testing the data against a null model, which is either rejected or fails to be rejected (at some particular confidence level). Your supposed null model isn’t even a model, since it has no expected parameter values. It doesn’t even have any parameters. The only thing you have proposed that in any way slightly resembles a model is the evolutionary one.
Possible, but nothing you have proposed would test that.
1 Like
The math has been done for bacteria and primates. These are specific tests with specific parameters. We now have AI that can generate models rapidly. What needs to be discussed is the test and test parameters.
Here is a sample of a simulation.
https://grok.com/share/bGVnYWN5_0c8640ac-4f24-4295-99c4-6c8b3d739bce
Posting AI-generated text that you don’t understand and can’t confirm to be correct is not generating the maths.
4 Likes
Would you like me to etch every equation in stone or is this too much technology for you 
I’d like you to admit that you either haven’t read it or don’t understand it.
1 Like
You clearly failed to read the post you were quoting:
No Bill.
A null hypothesis is an assertion that some statistical parameter has a specific value.
“Design or special creation”/“the claim that most the genetic variation we are observing is based on the original design of the species” is not such an assertion therefore
“Design or special creation”/“the claim that most the genetic variation we are observing is based on the original design of the species” CANNOT BE A NULL HYPOTHESIS.
Repeated attempts to claim it as a null hypothesis will simply be taken as further evidence that (i) you still know nothing about Statistics; and (ii) that you are still unable to contribute anything substantive to conversations.
If you could find a relevant parameter and value, then those could be a null hypothesis – but until then you have no null hypothesis.
I believe Behe himself has admitted that the assumptions in that model were unrealistic. That means that this model provides no basis for anything.
1 Like
The null model assumes no relationship between reproductive variation and the gene pattern observed.
How is this meaningfully different than a typical null model that assumes no relationship between the parameters being tested?
“reproductive variation” and “gene pattern” are vague concepts, not rigorously-defined, and thus measurable, parameters. Therefore they cannot be statistically tested.
1 Like
I agree they should be defined.
This is a bald assertion in bold. Why don’t you think things through before you commit to such silliness?
No Bill. 
It is a conclusion not an assertion. This is why I prefaced it with “therefore”:
Projection much?
1 Like
Lack of confirmation (or even claim) that you have read and understood the link you posted noted.
1 Like
Likewise neither “climate change” or “no climate change” can be a statistical null hypothesis, because they do not, in and of themselves, assert a value for a specific statistical parameter.
They can however lead to a prediction about a specific statistical parameter.
E.g. "“no climate change” would predict that there is no increase in ocean temperature over time at some specific measurement point (or points).
This could then be tested with a statistical hypothesis test of whether the series of those measurements is correlated to time, which would have, as its null hypothesis that the two are uncorrelated..
What you get is:
-
Scientific hypothesis.
-
Prediction of scientific hypothesis.
-
Statistical hypothesis testing of this prediction.
It is only in step 3 that you get a null hypothesis.
1 Like