Home » Biology, Design inference, Extraterrestrial life, Intelligent Design, Science fiction » Another Mars Mystery – Design, Natural or Hoax?

Another Mars Mystery – Design, Natural or Hoax?

Fox news reports that an armchair astronomer, David Martine, claims that he’s discovered evidence of intelligent life on Mars. In this YouTube video Martine speculates that it could be a bio lab, or a dwelling or garage (he hope’s its not a weapon. NASA is investigating.

So, is this evidence of intelligent design? Is it a natural phenomenon of some sort? Or is it a hoax (albeit an intelligently designed one)? And how might one go about making the determination?

Thoughts anyone?

  • Delicious
  • Facebook
  • Reddit
  • StumbleUpon
  • Twitter
  • RSS Feed

28 Responses to Another Mars Mystery – Design, Natural or Hoax?

  1. 1

    This looks like an ideal opportunity for design detection science. Can someone calculate the CSI present in that biostation?

  2. It’s a hoax.

    He’s superimposed an image atop some sort of other image.

    Given that the ‘cylinder’ is fairly sharp in its image, we should be seeing shadows from the surrounding mountains and outcrops. There not there. He’s just pulling our chain.

    Dear Prof. Gumby: I suppose you’re made out of candy, huh?

  3. Is there anything even like “Google Mars”? I don’t really think so.

  4. Can someone calculate the CSI present in that biostation?

    I just love how people fail to grasp ID at even the most fundamental level.

    ID is not intended to be a methodology that suggests you stop using common sense.

  5. Stop looking at my property!

  6. 7

    ID is not intended to be a methodology that suggests you stop using common sense.

    I thought ID was intended to be methodology to replace overreliance on common sense. Common sense can let us down in all sorts of ways. And clearly “oooh, it looks designed” is not a very convincing or rigorous argument. A mathematically rigorous definition of a CSI metric would go a long ways towards combatting the perception of ID as too much “uncommon sense” and not enough science.

  7. Can you tell us when your book on ID is coming out?

    Let me put it another way. I don’t have to calculate the CSI in your post in order to infer it had an intelligent cause.

  8. Prof Gumby:

    The “mathematically rigorous” talking point is a dead giveaway, given what has happened at UD for three months. Cf the answers to it, e.g. here.

    There is indeed a Google Mars [I was too busy elsewhere y/day to comment on that], and it is clear that the maps produced thereby are intelligently produced — just like the drawings of 100 years ago that showed poorly resolved structures as a pattern of canals. In the former case the design inference would indicate that if the maps were accurate they would indicate an artificial entity on Mrs, and as we know on having better imagery, the maps indicated the intent of the astronomers to represent what they thought they saw on Mars. On either view they were designed, the issue was whether the design was on the ground on Mars or in the minds of the astronomers of those days.

    Now, the issue now — as it was then, is whether a small and poorly resolved feature (we are looking at something that seems to be one to a few pixels wide by several pixels long with a zigzag effect [notice how it seems to align with the pixel pattern]) is or is not a glitch of the map generating process — there are many obvious digital artifacts in the maps and videos; or, whether it was planted there separately as a prank or whether it possibly is a feature on the ground.

    If it is a feature on the ground, it could be a complex of buildings, which would be functional, specific and information-rich, warranting an inference to design even if we do not know how they would have been put there. Buildings are not credible results of chance and necessity; being functional, specific, complex and organised. In short, we see how the explanatory filter would be applicable.

    If the result is a glitch, then it is an artifact of the software and the photos, a matter of chance and necessity.

    If it is a prank put in, it is obviously a design.

    But, right now the design inference filter is telling us that the level of information we can currently see would be well within the Chi metric’s zone where chance or mechanical necessity of the processing system could throw up sch a feature by noise.

    As happened earlier with the Mars Face feature as well. (That one went away on getting a higher resolution view, as happened with the canals.)

    This is not an issue for the Chi metric or the like to resolve under current circumstances.

    It is — as was the case 100 years ago and again a few decades ago — an issue for getting a higher resolution image of the specific zone. (A Mars Orbiter with a good telescope would be nice to have . . . )

    That there is insufficient data at present to clearly resolve a matter in a specific case does not lead properly to the conclusion that the Chi metric is ill defined and meaningless.

    Save, to those who brought the talking point to the table up their sleeves to begin with and are looking for a hook to hang it on.

    GEM of TKI

  9. UPDATE: The video has been removed by the Youtube user. There being no explanation, we are left to infer that most likely he has been satisfied the video was a mistake. Glitch or ice patch or the like, and as noted, well within the reach of the chance and/or necessity [here, aka "it ain't a bug itz a FEETURE!"].

  10. F/N: A suggestion, found at the Daily Galaxy (on how a glitch could produce the item):

    >> “It looks like a linear streak artifact produced by a cosmic ray,” said Alfred McEwen, a planetary geologist at the Lunar and Planetary Lab at the University of Arizona and the director of the Planetary Imaging Research Laboratory. McEwen is the principal investigator of the High Resolution Imaging Science Experiment (HiRISE), a powerful telescope currently orbiting Mars. [Okay, there is an eye in the Mars sky]

    Cosmic rays are extremely energetic particles emitted by the sun and other stars. For the most part, the Earth’s protective magnetosphere blocks them from hitting the planet’s surface, McEwen explained. “But with space images that are taken outside our magnetosphere, such as those taken by orbiting telescopes, it’s very common to see these cosmic ray hits. You see them on optical images and a lot of the infrared images too,” he told Life’s Little Mysteries. [Read: Why Do Photos from Deep Space Take So Long to Get to Earth?]

    As a cosmic ray passes through a camera’s image sensor, it deposits a large amount of its electric charge in the pixels that it penetrates. If the particle passes through at a shallow angle to the plane of the camera, it affects several pixels along its path. The result is a bright streak on the image.

    The digital compression software that converts the image into a JPEG file then “sort of smears out the image, giving it that pixelated look,” McEwen said. What started as a clear streak in high-resolution turns into a streak that, in the armchair astronaut’s words, looks like it is “made up of cylinders.”

    McEwen said that the cosmic ray streak would be much easier to recognize in the raw, pre-compressed image, but many orbiters and telescopes have contributed imagery to create the Mars map, and Google doesn’t identify the source image.

    “I can’t tell whether this image was taken by Viking or what,” McEwen said. “The people at Google need to document what the heck they’re doing. They should be able to identify what the source of their information is, and let people know so they can go back and look at the raw data.” >>

  11. 12

    KF:

    The “mathematically rigorous” talking point is a dead giveaway, given what has happened at UD for three months. Cf the answers to it, e.g. here.

    It would be a “giveaway” were I trying to hide anything. I only thought that given the inability of ID supporters to calculate the CSI of MathGrrl’s example cases, that here was another opportunity to develop the metric.

    I would agree with you that a better telescope is desirable. More data are always better.

  12. I don’t see why we need a calculation of CSI in order to determine design. If we encountered something like, say, a bulldozer on mars, and we knew beyond any shadow of doubt that no human from earth brought it there, and then with no calculation at all we would know a)it was designed by some intelligence and b)that intelligence is extraterrestrial. That would be true if we encountered some object we didn’t recognize, and didn’t know what it was for, but had metal parts of some unknown metal, flashing lights of some sort, and the like. In such a case, why would we need a calculation of CSI to infer design? Same for the image on this video. Suppose it turned out it really was some sort of building structure…what would we need to calculate to correctly infer design?

  13. Prof Gumby:

    You are spouting more counter-factual talking points.

    And the tangential red herring to turn this into yet another round on the long since adequately answered demands, is telling. FYI, the first case MG gave was immediately answered right there in her guest post, by several people, myself included — copying existing information is not origin of new info but may point to a process that is additional FSCI. The rest of cases were also answered, over and over, with data where that was accessible, She showed that NO answer would ever satisfy her and so that matter is closed, especially after she made snide allusions to the inquisition etc.

    And the whole mathematical rigour game was misdirected, complex specified info is an empirically observed fact, and it is convertible into reasonable metrics and that can be applied to cases.

    I suggest you work your way through the already linked response on “rigour” and the round up on the issues she raised — dozens of them, never engaging in responsive serious dialogue — here.

    But after three months of reiteration of answered talking points as though there had been no reasonable response, it became very plain that there was a willful intent to suppress the truth that reasonable answers exist and were given, again and again. If you want to parrot such tactics, we will know what to make of your unresponsiveness to the evident truth.

    Good day

    GEM of TKI

  14. DM:

    I actually gave a rough approximation to a CSI calculation above: the number of pixels is too small to be even worth more than that.

    There is a plausible explanation on a Cosmic ray strike that makes sense, and the uploader of the video has withdrawn it. Chance happening triggering the usual ionisation, triggering an artifact on the image.

    Case closed.

    GEM of TKI

  15. kairosfocus,

    the first case MG gave was immediately answered right there in her guest post, by several people, myself included — copying existing information is not origin of new info but may point to a process that is additional FSCI.

    That was certainly asserted, but no explanation of why you came to that conclusion was provided. Please remember that my first scenaro dealt with a case where the gene duplication resulted in increased production of a protein. Neither you nor anyone else explained why “Produces at least X amount of protein Y.” is an invalid specification, according to Dembski’s explanation of CSI.

    The rest of cases were also answered, over and over, with data where that was accessible,

    No, they were not. If they had been, you would have simply provided links to the answers. You cannot do so.

    And the whole mathematical rigour game was misdirected, complex specified info is an empirically observed fact,

    There is no way of supporting that statement since you have never provided a rigorous mathematical definition of CSI. Without such a definition, your claim is literally nonsense. You could replace “complex specified info” with gobbledegook and it would convey the same amount of meaning.

    We’ve been over this too many times for me to expect you to actually answer the questions I posed. I’m not going to continue the debate in this thread, although I will be happy to discuss the topic further if you actually do come up with a rigorous definition and some detailed example calculations. Until then, I will simply point out when you make claims not in accordance with the facts.

  16. PS: Prof Gumby may be interested to see the metrics he thinks do not exist in action here on, in reply to Dr Liddle.

  17. MG:

    Please, stop willfully supressing the truth. We can all follow the links, and no I will not engage in a furtehtr round of fruitless attemtps to again and again answer your demands that you will never acknowledge have been reasonably answered to.

    Onlookers:

    Observe my standard reply to MG;s willful distortions, from the last thread here (I got sick and tired of going over material again and again, so I just point to it now):

    ______________

    >> On CSI and its “rigour,” that has been addressed over and over again, in most specificity to the issue of rigour, at 34 – 5 in the previous. Similarly, the talking points MG tends to use over and over as thought hey have not been cogently answered, were last dissected in 23 – 24 similarly. And, the overall summing up of the issues MG has needed to explain herself on has been kept up in the editorial response to Graham at no 1 in the CSI newsflash thread; which MG has persistently ignored.

    When it comes to ev, 137 above shows my links to the places in the CSI Newsflash thread where it is dissected by Mung. (One of MG’s tactics seems to be to wait until something is buried under enough posts in a thread, or has been continued in a successor thread, before repeating the assertion that was rebutted.)

    She knows or should know better than she has acted. >>
    _______________

    When MG can show that she is reading seriously and responding ont eh merits, we can have a fruitful discussion. But so long as what we are seeing is willful repetition of false talking points in the teeth of three months now of repeated cogent response on the merits, that is more than enough.

    I think the saying is: don’t feed the trolls.

    Good day

    GEM of TKI

  18. kairosfocus,

    Anyone can read the thread to which you linked and see that the comments you specify do not, in fact, address the issue. I pointed that out in detail in several comments of my own following yours.

    If you had a rigorous mathematical definition of CSI, as described by Dembski, you would simply be copying it in response to my challenge. You haven’t done so because you don’t have such a definition. The same holds for any detailed calculations based on that definition.

    This is already more than I planned to write before seeing some real answers from you. As always, if you choose to actually define CSI and show how to calculate it, I’ll be delighted to continue the conversation.

  19. Onlookers:

    Plainly, the point needs to be repeated:

    __________

    On CSI and its “rigour,” that has been addressed over and over again, in most specificity to the issue of rigour, at 34 – 5 in the CSI footnote thread. Similarly, the talking points MG tends to use over and over as thought hey have not been cogently answered, were last dissected in 23 – 24 in the same thread. And, the overall summing up of the issues MG has needed to explain herself on has been kept up in the editorial response to Graham at no 1 in the CSI newsflash thread; which MG has persistently ignored.

    When it comes to ev, 137 in the same shows my links to the places in the CSI Newsflash thread where it is dissected by Mung.

    (NB: One of MG’s tactics seems to be to wait until something is buried under enough posts in a thread, or has been continued in a successor thread, before repeating the assertion that was rebutted.)

    She knows or should know better than she has consistently acted.
    __________

    Until MG shows serious responsiveness and contrition over things like the snide invidious association with the inquisition she has made, we need not take her further rhetoric seriously.

    I say that with regret that it has come to this, but she and those she associates with have now poisoned the matter to that extent.

    GEM of TKI

  20. 21

    Uhm….by the way.

    I stopped reading the Mathgrrl threads when they became the opportunistic and repetitive talking points she intended them to be, (and of course, my own exchanges with Mathgrrl ended in a whirlwind of logic whereby she needed something from me in order to know what she herself was talking about).

    And I have no desire to go back and read them now…..so I have a question.

    Did Mathgrrl overturn countless obsevations and thereby demonstrate that CSI doesn’t exist?

  21. 22
    CannuckianYankee

    Too bad. The video has been taken down by the user (Martines?). Which upsets me to the max, ’cause I was hoping to get a closer look at “Ski Station Beta.” Not that I could go there anytime soon, but I can still dream. :)

    You people keep moving this discussion to different threads and it’s hard to keep track. Could the moderators please make a separate thread and leave it open for the ongoing discussions with MathGrrl? I keep missing a whole lot of the discussion.

    Thanks.

    KF:

    “As happened earlier with the Mars Face feature as well. (That one went away on getting a higher resolution view, as happened with the canals.)”

    The last I heard about the Mars Face was in a 90s sci-fi movie called The Red Planet, in which the face actually is a large building holding the original seeds of earthly life. I guess that’s why the movie was such a dud – panspermia and all. But it was a design explanation for how evolution got started on Earth.

  22. 23
    CannuckianYankee

    Upright,

    I guess you’re right. Moderators, please ignore what I said about starting a separate thread.

  23. MathGrrl @16:

    Neither you nor anyone else explained why “Produces at least X amount of protein Y.” is an invalid specification, according to Dembski’s explanation of CSI.

    And you never explained why you thought it was a valid specification. Yet it was your guest post and your challenge and therefore it was your responsibility to to do so.

    Here, let me help. Here’s a starting point for you:

    I, MathGrrl, believe that “produces at least X amount of protein Y” meets the criteria of a specification because Dembski defines a specification as [fill in your answer] and “produces at least X amount of protein Y” shares the following features in common with that definition.

    It’s not up to us to make the case that your specifications don’t qualify as specifications. It’s up to you to argue that case that they do qualify and that they are in fact specifications.

  24. Did Mathgrrl overturn countless obsevations and thereby demonstrate that CSI doesn’t exist?

    No, in fact MathGrrl believes that Tom Schneider’s ev program can generate CSI, but she just can’t seem to bring herself to explain why she believes it.

    I’ve repeatedly asked her to discuss that claim of hers with me with no success.

  25. Funny all this is coming up now in a thread about hoaxes.

  26. Onlookers:

    Just for the record, on assumption you can follow the links to see context tracing to info theory and related areas, here is the first part of my response to Dr Liddle, 147 the Obvious thread:

    ______________

    >> I: Preliminaries:

    . . . . here is the basic definition of information in the UD glossary for several years now, that is extracted by way of testimony against interest from Wiki:

    “ . . that which would be communicated by a message if it were sent from a sender to a receiver capable of understanding the message . . . . In terms of data, it can be defined as a collection of facts [i.e. as represented or sensed in some format] from which conclusions may be drawn [and on which decisions and actions may be taken].”

    Perhaps, you are concerned about how it is quantified, well at first level there is a Hartley suggested metric picked up by Shannon and others that uses symbol frequencies viewed as frequentist probabilities, and then deduces for the kth symbol:

    Ik = log(1/pk) = – log pk

    This is then extended to the average quantum of info per symbol by a weighted sum:

    H = – [SUM on k] pk log pk

    So far we are basically dealing with measures of symbol frequencies on the idea that the rarer symbols are more surprising, more unexpected, and thus more informative, and using a log metric to allow additivity. It turns out — by the peculiarities of probabilities — that a string of [flat] random characters will under this metric, give the peak value of information per symbol [by contrast in English about 1/8 of characters in a typical message will be an E].

    How do we get to the more conventional sense, where meaning is important, and information is often expressed in coded clusters of symbols with defined vocabularies and rules for meaning? (In cases where information is implicit in a structure or an organised cluster of components to do a function, the structured set of yes/no questions that specifies the outcome is going to be similarly in symbols and will have rules of meaning and a vocabulary.)

    Sets of symbols imply possible configurations, only some of which are meaningful or functional, i.e. we are looking at islands of function [or more broadly zones of interest] in the space of possible configurations. That means the observed event E comes form a zone of interest T, that can be separately specified or observed.

    This is of course the root of the idea that we have deeply isolated zones of interest or islands of function in vast seas of possible configs.

    As you know, if we are dealing with a binary string, each additional bit doubles the config space of possibilities. Beyond 500 bits, the resources of the solar system of 10^57 or so atoms will not be enough to scan more than 1 in 10^48 of the possibilities so a zone of interest that is sufficiently narrowly defined will be so isolated that it is not accessible to random walk driven trial and error search. And, unless a search is matched based on knowledge, the typical search strategy will on average do about as well as that. It is intelligently injected active information that makes a difference and makes reaching such zones within resources a routine observation, e.g. posts in this thread.

    This is of course the heart of the concept complex specified information [CSI], and functionally specific complex information [FSCI] is the case where the zone of interest T is specified on a function, e.g. an AA chain must fold properly and work as a Type-X enzyme. (Cf Durston et al’s estimates for 35 protein families for cases in point.)

    A simple way to model and measure such FSCI, towards use in the decision-making explanatory filter to decide whether on best explanation something is intelligently caused is:

    1: Define specificity, S such that once we observe a specification S = 1, and if not S = 0. (Typically this will be on observed function and on observing or inferring the likely effect of significant random perturbation that moves us a reasonable Hamming distance away from an original observed E within T.)

    2: Identify a measure of information, such that I is calculated per symbol frequencies or if a storage medium is used, it can be estimated from the storage used at least to an order of magnitude. [E.g. File-X is 127 kbits]

    3: Using the log reduction of Dembski’s Chi:

    Chi_500 = I*S – 500, bits beyond the solar system threshold

    Chi_1000 = I*S – 1,000, bits beyond the observed cosmos threshold.

    4: On the case of a random string, S = 0.

    5: On the case of a string that is set to a fixed repetitive pattern [say Thaxton et al's THE END, repeat], I = 0 or a value near to 0.

    6: Taking as a text, [Dr Liddle's] 127 up to the smiley, you have 3282 128-state ASCII characters. Or,3282 * 7 = 22,974 bits, an I value; and as text in English S = 1. Chi_500 = 22,474 functionally specific bits beyond the threshold.

    [Note, in my first response to MG's guest post, I did a similar calculation, on her post, using the X-metric, which does essentially the same job, but does not do it the way Dembski did. In short, right form the beginning MG was presented with a quantification of the FSCI -- the functionally specific form of CSI -- in her own post, using as a base the very same unit of information commonly used to measure computer file sizes. Then, dummy variables were applied to identify that it was beyond a 1,000 bit threshold, and that it was functionally specific. So, FROM THE VERY BEGINNING MG WAS PRESENTED WITH A CONCRETE APPLIED CASE OF CSI IN ACTION, WITH A VALID METRIC BASED ON MEASURES COMMONLY USED, AND APPLYING THE SAME ISSUE OF ISLANDS OF FUNCTION THAT SHE TRIED TO DISMISS AS IRRELEVANT TO THE MEANING AND VALIDITY OF THE DEMBSKI METRIC. This alone suffices to show that her performance these past three months does not pass the basic smell test. We have been dealing with an ideologicical rhetorical agenda, not open-minded, give and take dialogue. That is also why she has refused to acknowledge the facts Mung has unearthed on the nature of ev and other GA's, or my point that such GA's all work within already set up islands of function, i.e. you are already on the target, just you are trying to hit the gold ring. So, ev etc are irrelevant to the real issue of macroevo, getting so viable body plans requiring 10 - 100 Mbits or more of new bio info. And, as other dismissal of the cases where I tracked down actual cited claimed additions of information by such algorithms, her refusal to acknowledge that the numbers for info added came from sources including herself, speaks volumes. None of it good. And so on and so on for three months. I only write this to help those who may be unacquainted with the things that have happened and why it is that my response to her and those of like ilk at this point is iron-toned. I am now convinced I am dealing with ideological zealots and those who have no compunction about saying over and over again what is not true, and what hey know or should know is not true, for rhetorical advantage, and not with sincere dialogue. I know, I know, that sounds strong, and maybe we can all get along nicely.But, for there to be dialogue, there has to be a two sided responsiveness; for three months, there has been NONE from MG and ilk. If you doubt me on this, contrast the exchanges I have had with Dr Liddle. (And, Dr Liddle, I understand the potential career cost of going beyond a certain point with us here. Sadly, too many are in denial over the sort of ruthless -- and in some cases outright sicko vicious -- ideologues we have been dealing with. But exposing that is part of what now has to be done. And yes, I am very aware of the turnabout tactic that will be tried to twist this around to make me and others who point out what is going on seem to be the troublemakers. That is one of the most vicious of all tactics: blame the victim, or the one who stands up to stop the bully-boy thought police game.) 'Nuff said. ]

    7: On seeing such a positive value of FSCI beyond the threshold, I comfortably infer that the cause is not mechanical necessity without intelligent direction, nor is it chance, but intelligence.

    8: Similarly, I have several times cited cases from Durston’s 2007 Table I, and have set up the inference to design for these protein families thusly:

    RecA: 242 AA, 832 fits, Chi: 332 bits beyond

    SecY: 342 AA, 688 fits, Chi: 188 bits beyond

    Corona S2: 445 AA, 1285 fits, Chi: 785 bits beyond . . . results n7

    8: In both cases, the matter is simple and in the case where we directly know the cause, the inference from FSCI — as usual — is accurate. As was shown, a random text and a repeated block would not pass the threshold, for opposite reasons.

    9: A simpler, brute force X-metric does much the same, based on storage use.

    That should be enough for basic backdrop. >>
    _______________

    Let the record reflect that the roots, context and utility of the CSI concept and one metric model have been shown in outline, in a context that leads back tot he fuller discussion inclduing addressing any number of hyperskeptical talking points as seen these past three months.

    And it the above sounds like I typed it through gritted teeth, it is because tha tis precisely correct.

    Anytime you see someone blandly declaring that the CSI concept has no reasonable warrant and is not a meaningful or measurable entitty, that person is either parroting a talking point or is saying what she or he s=knows or should know is false and misleading.

    Period.

    And in the case of MG, she owes a major apology for the snide and utterly out of order insinuation that the situaiton is in any material respect parallel tothe case of the Inquisition’s threatening Galileo with torture intruments to get him to publicly back down.

    As for those who have gone on to play at outing tactics, THAT is intended to intimidate people who would think to say anything supportive of ID. Those who have gone on to make false accusations of UD being a nest of perversion do so knowing that where I come from, that is a mortal insult worth someone’s life.

    That is the level of poisonous, vicious mind we are dealing with, and we had better stop thinking that if we are nicey-nicey maybe they will love us and be nice to us. No, at this end of the spectrum, we are dealing with the same sort of ideological, indoctrinated fanatic that I cut my intellectual eye teeth on many years ago now: the Communists.

    That is what Dawkins et al REALLY mean when they suggest that those who dare differ with them are ignorant, stupid, insane or wicked.

    Such can only be exposed and broken when their follies and agendas are manifest to all. And, fence sitting is enabling behaviour.

    (That’s why I wrote this post, as a warning. Bone up on Alinsky, folks, that is what we are dealing with.)

    Good day,

    GEM of TKI

  27. PS: Onlookers, observe how MG is — as usual — dismissive without being substantial; cf with the just above which clips a summary response on the merits from an exchange with Dr Liddle.

    We can safely predict that no answer on the merits, no worked examples [never mind that some of these show how the Dembski and Durston metrics fit together, never mind her assertion/suggestion that such was not possible, etc etc], no analysis will ever receive a more responsive answer from her.

    Conclusion: she is playing ideological, rhetorical games in the teeth of what she knows or should know, she is not responding on the merits — where, if I was wrong, she and her ilk would love to pounce on the errors I have made, in painstaking details, i.e. it it that very dismissiveness that is the key clue that this response is based on a bluff intended to drag us into cycle after cycle of trying to prove something to the utterly unreasonable.

    And, it is further to be noted that she has been especially unresponsive on mathematics, save to make errors like dismissing a log reduction calculation as a probability calculation. For MONTHS, she has failed to explain this blunder.

Leave a Reply