Recently, Than Christopoulos produced a silly and absurdly wordy insistence that Bayes’ Theorem proves Jesus rose from the dead in Paulogia…It’s Time To Stop. I was recently interviewed about this on Modern Atheism (you can watch that here) and as promised I am here writing up the main issues. Christopoulos frames this as a response to Paulogia who has repeatedly roasted him and Michael Jones over their attempts to argue this, thus explaining this desperate wordwall as their end-state failure-mode. It’s 12,000 words of almost pointless rambling, almost all of which is trivially true, while the rest of it is badly worded or face-palmingly wrong. Here, by contrast, I will explain why all apologetics for the resurrection fail regardless of whether it’s translated into math—but especially when it is.

It would be tedious and pointless to fisk every point they muck up or get wrong. What I will do here is more productive: I will zero in on the most crucial mistakes they are making, which establish their entire position as another example of Formalized Gullibility as a Modern Christian Methodology, vectored through yet another instance of Crank Bayesianism.

Outline of the Dispute

Gary Habermas’s “minimal facts” apologetic has catastrophically failed—from practically lying about the state of the field, until reality caught up with him and Habermas had to repeatedly trim his “minimal facts” from what was once around sixteen down to now only three “facts” that most mainstream scholars agree with (Jesus was crucified, some disciples had or claimed experiences convincing them he had risen from the dead, and then at least one skeptic, Paul, did too).

That approach having been chewed into the grave, gullible and desperate Christians flipped the other way and started insisting on a “maximal facts” apologetic, whereby you just “believe” everything the Gospels say (because “how could they lie?”), and from that you can prove Jesus rose from the dead. Which ironically is illogical even granting all its premises. But its premises are also universally rejected by mainstream scholars, being accepted only by Christian apologists. So it’s basically returning to a Scopes Monkey Trial scale of apologetic credulity on line with Young Earth Creationism and Flat Earth Theory. As such this turn is an example of what I have been documenting lately as The New Christian Apologetics, which is just a garbage recycling of gullible approaches we refuted and made fun of twenty years ago. But lo. Here they are. Forgetting all that happened. And instead of acknowledging that, they resort to just getting angry and flipping tables when we remind them of it (“It’s Time To Stop, Paulogia!!! Stop!!! Stoooooopppppppp!!!!!!!!”).

The “new” approach retranslates the old dumbery into fancy mathematics, but that’s just lipstick on a pig, to try and steal authority from the appearance of doing math to sell what has always been a failed argument. But rewording it in a mathematical language does not fix any of it. The overall approach is first to insist on a Historical Reportage Model (HRM), whereby some (notably rarely Bayesian or mathematical or even logical) arguments are chucked up to try and game you into assuming the Gospels are just faithful historical reports and not the propagandistic mythologies the bulk of mainstream scholars have proved they are. I have an updated bibliography on that being a fact in chapter two of Obsolete Paradigm—the field has moved even more in this direction than its already-majority consensus ten years ago, leaving Christian apologists behind to diverge from mainstream thought and continue insisting on the ridiculous.

With that premise “established” (through hysterical handwaving rather than even a pretense of careful Bayesian analysis), they then use it to build what they call the Maximal Data Case (MDC), translating their gullible HRM position into numerical premises in Bayes’ Theorem to get it to be “probable” that Jesus rose from the dead. But since the premises are bogus, so is the conclusion. Their MDC is just GIGO. Paulogia has pointed this out repeatedly. And all they do in response is misstate his every point so as to go on making the same refuted arguments over and over like a broken record. That this is the new normal for Christian apologetics I discovered with Andrew Loke, Cameron Bertuzzi, and, yet again, Michael Jones. Christian apologetics, like the flat Earth movement, is a zombie profession now, having no point or value anymore other than serving their own desperate egos and hope of stopping the bleed.

Bogus Premises, Bogus Math

I am only going to cover the key errors here, but they are sufficient to demonstrate how incompetent Christians are at evaluating evidence. Which adds to my accumulated evidence that this is why anyone is still a Christian: if they evaluated evidence rationally (without delusion or fallacy, and without the suicidal gullibility that would literally trap them in any other cult if chance would have had it), they’d realize their religion is bunk and leave. This creates a filter, whereby only fools remain in the fold. Which is why they keep saying foolish things with such alarming overconfidence. This article by Than Christopoulos, defending ideas endorsed by Michael Jones, exemplifies the point.

In its simplest format, Bayes’ Theorem entails that for any claim about history or the world, the product of the prior probability ratio and the likelihood ratio equals the posterior probability ratio. Or put more simply: the prior odds given all past evidence, times the consequent odds given all current evidence, equals the final odds on anything being true. Any claim is a causal theory: you are proposing an explanation of all the observed evidence, and determining how likely your proposed explanation is compared to others (see What Is Bayes’ Theorem & How Do You Use It? and for the deep logic behind Bayesian reasoning as causal reasoning, the real solution to Hume, see Judea Pearl’s Causality: Models, Reasoning and Inference). Bayes’ Theorem does this by accounting correctly for the effect of all pertinent evidence. Any final Bayesian conclusion on anything is bogus (indeed, crank) that leaves out pertinent information (take note of that; it’s going to come up). As is any Bayesian premise that does not derive a probability from that evidence logically (take note of that; it’s going to come up).

There are effectively three premises in Bayesian reasoning: the prior distribution (the relative frequency of your proposed cause having been the cause of previous similar outcomes), and two likelihoods: the likelihood of the evidence on your explanation, and the likelihood of all that same evidence on any other explanation. And usually it’s fine to just consider the next most likely explanation (the “best” explanation other than yours). But when all explanations are unlikely, you might need to take seriously conjunctions of explanations. Christopoulos does not do this. He never sums-in the frequencies and likelihoods of every explanation alternative to his. This creates the only explicit mathematical mistake in his analysis: His “HAlt” must be inclusive of all logically possible alternative explanations of the evidence. All his other mistakes are of deriving the premises (the probabilities he puts in—hence, his inputs, the Garbage In), not the ensuing math, which all looks fine so far as I could tell (I did not check closely because any further errors there wouldn’t matter).

But this one mathematical mistake matters in a crucial way. When he says his premise HAlt is “any competing hypothesis (e.g., hallucination theory, theft, legend),” that’s incorrect. It’s every competing hypothesis: P(HAlt) = P(hallucination) + P(theft) + P(legend) + P(dreamquesting) +P(shrooming) + P(schizotypal cult formation) + P(heist) + P(fabrication) + P(mythologization) + P(wizards) + P(demons) + P(aliens)” + every other logically possible theory, and every other admixture of theories—because several things could have converged to cause the present state of evidence as well, and we have to include them all. If the array is (P(hallucination) = 0.1) + (P(theft) = 0.2) + (P(legend) = 0.2) + (P(dreamquesting) = 0.1) + (P(shrooming) = 0.05) + (P(schizotypal cult formation) = 0.15) + (P(heist) = 0.15), and everything else sums to less than 0.01, and we are rounding to the nearest whole percentile so they collectively vanish from the math, then P(HAlt) = 0.1 + 0.2 + 0.2 + 0.1 + 0.05 + 0.15 + 0.15 = 0.95 or 95% which leaves only a 5% chance of “resurrection.” I just made up all those numbers. But the point is that the sum of all “not resurrection” theories can exceed “resurrection” even if “resurrection” exceeds each and every one by itself. Christopoulos fails to account for this, and that alone is fatal as-is. I’ve caught Christian apologists making this mistake before. It’s a variant of the conjunction fallacy (or the Linda the Banker Problem).

I think Christopoulos could “fix” this, though it would make his already far too wordy wordwall even wordier and the math harder (both for him to do and his audience to follow), and wouldn’t likely save his argument—because of the other errors he makes, which automatically exclude key hypotheses and conjunctions of hypotheses at the very point of his forming his premises. For example, the most central mistake that destroys his entire case is conflating the Gospels with the testimony of Paul, whereby he “invents” eyewitness accounts (that don’t exist) of a walking, breathing, talking, eating, cuddling, sticking-around Jesus (what he calls a “multimodal group experience”). But this is a factual error. The facts are of an evolution from a much more typical religious experience to that myth. So when Christopoulos tries to compare his explanation with “hallucination” he credibly derives a likelihood for that—but as that only applies to evidence that doesn’t exist, it is the incorrect likelihood, and thus a bogus premise. No amount of accurate math after that point can save him. The correct (most probable) HAlt is isolated dreams/hallucinations evolving into legends of a multiomodal group experience, which is far, far, far more likely than either “celestial spirit magic” or “multimodal group hallucinations.”

So those are the kinds of mistakes (bogus premising) that I will spend the rest of the time here discussing. But I did need to pause on this one mathematical mistake because any attempt he makes to fix the one requires him to fix the other. And this matters to coming points I will be making, because I shall illustrate his mistakes with an a fortiori thesis of a heist. I will propose that as a covering hypothesis inclusive of all heist scenarios, not exclusive to only one—like the Gospel of Afranius or anything else, not even limited to the entirely plausible hypotheses of Robert Price in “Explaining the Resurrection without Recourse to Miracle,” in The End of Christianity. So P(heist) = P(Afranius) + P(heist[1]) + P(heist[2]) + P(heist[3]) + … P(heist[n]). I think this is actually the least likely of the plausible explanations of the data, so I am not even defending it. What I am saying is that even the worst hypothesis on the table is better than theirs, so a fortiori theirs is necessarily the least likely hypothesis on the table. I will also mention that even hypotheses so improbable they aren’t on the table (as theirs shouldn’t be either) are more likely than theirs (like an alien heist; or sorcery; or demonic trickery). But you will grasp the point more firmly if you see the same unfold for an entirely plausible scenario. I’ll get to why their hypothesis isn’t even plausible in the end—everything they say about its prior probability is laughably silly and illustrative of how deluded and incompetent Christians are (which is comparable to Young Earth Creationists and Flat Earthers now), but that will be my closing point.

In between we’ll run the principal problems with the other two premises: the likelihoods.

To that end they use only two items of so-called evidence: the appearances and the conversion of doubters. So I’ll cover those and then conclude with their silliness about priors. Because they don’t quantify the usual apologetic nonsense about people being willing to die for “multimodal sensory reports,” which is good because there is no evidence any witnesses even died, much less for that (see Did the Apostles Die for a Lie? and Interpreting 1 Clement’s Supposed Descriptions of Fabulous Murders)—and people die for lies all the time, from righteous fanatics defending a cause to amoral death row serial killers (like Ted Bundy).

And I won’t belabor the facts about the empty tomb claim, for example, as they also don’t use that as a premise in their math. The evidence indicates that was made up anyway, and that even if not, it has too many commonplace natural explanations, so it cannot get any usable result for them. That’s why Habermas dropped it from his minimal facts, and why most scholars deem the empty tomb false or unproven and thus unusable as a premise. We have no eyewitness nor even a named (much less verified) source for it. Instead, they use only the crucifixion (which they admit is as expected on all natural cause theories) and the appearances and doubters as their evidence. Which are, curiously, the only components of their “maximal case” they actually try to quantify and defend, yet happen to be exactly the same as the remaining three of Habermas’s minimal facts. Things that make you go hmm. But whatever. Let’s proceed.

Conflating the Gospels and Paul

As apologists always do, they distort the evidence (“E” in their equations) by conflating Paul’s account and the Gospels’ accounts, rather than including the chronological development, which extends not just from Paul to the Gospels, but across the Gospels, as the legends grow and become more mythical and ridiculous over time (something real stories don’t do—unless you admit the mythical and ridiculous versions no longer preserve the real story). Which requires accounting for evidence missing in Paul (the silences, which have to be causally accounted for without gerrymandering made-up accounts of it not in evidence), and hence the disconnect between what he says (he only ever knows about internal visions and ancient scripture as sources) with what the Gospels say (which explode Paul’s far more natural and credible accounts into fantastical stories, naming no sources they spoke to, not even themselves).

That Mark is a mythological fictionalization is clear from All the Fantastical Things in the Gospel according to Mark and Mark’s Use of Paul’s Epistles, with notes in Why Did Mark Invent an Empty Tomb? And that these tales didn’t exist when Paul wrote is detailed in Then He Appeared to Over Five Hundred Brethren at Once! Paul very clearly and repeatedly says the risen Jesus was never met in the flesh but as an internal mental experience (Romans 16, Galatians 1, 2 Corinthians 12, 1 Corinthians 9 and 15, etc.). Paul describes Jesus “appearing” most directly as a voice; but psychedelic visual phenomena would not be atypical in the anthropology and psychology of religion (including inner “ascents to heaven” like Paul also describes, most likely in a dream; or amorphous lights when in a waking ecstasy, as Acts describes; I provide a bibliography to the history and science in Element 15 of chapter 4 of On the Historicity of Jesus, and add more since in chapter 2 of Obsolete Paradigm).

The idea of multimodal sensory encounters not only never comes up in Paul, our only eyewitness, but are conspicuously absent precisely where he’d have cited them as evidence of the nature of the resurrection in his extended articulation of that in 1 Corinthians 15. And everywhere else. When he says his claims were a “stumbling block” and “foolishness,” the most directly obvious reason is that their only evidence had been mystical visions (theirs or prior prophets’), not having met an actual man. They therefore could not present forensic evidence of someone surviving their execution, and so Paul had none to present, or to explain as what people were not believing. Whereas revelatory claims were typically regarded as foolish by educated Greeks (that was a stock attack on religious claimants in extant philosophical critiques, reflected in the historical fiction of Acts 17 and Paul’s frequent allusions to the idea), and as stumbling blocks to skeptical sectarians (claiming to have had revelations contrary to opposing Jewish sectarian beliefs would cause stumbling like what Paul alludes to in Galatians 3:1).

It is not rational to circularly “presume” these all refer to something else (some “other” thing being a stumbling block or called foolish). Because that violates the very Bayesian methodology they want to employ here. The normal, natural, commonplace history and science of religious experience perfectly explains everything Paul says and doesn’t say. Therefore, one cannot just “dismiss” that to get this evidence to be “more likely” on the fantastical hypothesis. It’s not. It’s less likely. A point I’ll get to next. But for now the principal point is that the actual evidence goes from a perfectly normal, natural, nonsupernatural account of what happened from an eyewitness, to a fantastically mythological account attributed to no witness at all, and composed a lifetime later in a foreign land and language.

And note I do not mean simply that the names of the Gospels came later (see Three Things to Know about New Testament Manuscripts), but that their authors do not identify themselves in the text as the witnesses of anything, nor describe ever consulting any named witnesses. You cannot “insist” they are doing that when they didn’t. That is simply not in evidence. It does not matter that “maybe coulda” they did mean that somehow; no evidence is no evidence. Claiming they meant what you want is to invent evidence, not account for it. The Bayesian method is simple: the perfectly normal, natural, nonsupernatural explanation of these texts matches the evidence we have perfectly. And that’s that.

You cannot then claim the evidence is “improbable” on that hypothesis. To the contrary, it’s improbable on the other hypothesis. Someone recording history faithfully with the intent to confirm it to others typically does what that entails: name sources, explain how their accounts were confirmed, and why we should believe any of it. Instead, the last redaction of the last Gospel has its (plural) authors tell us their source was something they insist was written by an unnamed guy conspicuously never in any prior account, whom they (plural and still unnamed) claim was a witness—but they never say how they know that. Was that previous written Gospel actually authored by any such person? Did that person even exist? Or was it actually just another bogus Gospel like the forty or more we have or know about? Since the “author” they mean here was almost certainly in earlier redactions intended to be a fictionally reified Lazarus, a “made up guy” bears a stronger Bayesian probability here (I document all the evidence in chapter 10.7 of On the Historicity of Jesus, to which we can add Herman Waetjen, The Gospel of the Beloved Disciple: A Work in Two Editions, and Ben Witherington, “What’s in a Name? Rethinking the Historical Figure of the Beloved Disciple in the Fourth Gospel,” in John, Jesus, and History: Aspects of Historicity in the Fourth Gospel). And so on. But either way we’re in the same place: they can’t establish the things they want to be evidence are even true to begin with.

I show this conflation of Paul and the Gospels to be the most common and fundamental mistake Christian apologists make when trying to convince rational people the resurrection of Jesus is plausible in Resurrection: Faith or Fact? Though to study all the mistakes they make, Pearce’s New Take-Down of Resurrection Apologetics Is a Must-Have. When these are mistakes. Often I suspect these are deliberate tricks and distortions; as apologists tend too often to be liars, and continuing to defend things that are false after being shown they are false is the surest way to convince us not to trust you or your judgment anymore. But either way, it has the same effect on the math: by creating false inputs as E, their conclusions are bogus no matter how precise their math. Real (correct) inputs would have the opposite effect: the actual evidence and its actual chronology and what it actually contains and what it actually doesn’t are exactly what is expected on ordinary and common sequences of events throughout history and not at all what we expect on the supernatural hypothesis.

To be fair, that it is indeed “not at all” like what we expect on their hypothesis I have only touched on so far. But we’ll get the big reveal on that soon.

Ignoring (or Lying About) the Actual State of Scholarship

The second factual mistake to note is one that is a kind of new regressive slide into backwards apologetics: the claim that the Gospels are faithful historical records. That has been so thoroughly and empirically refuted in the field that the mainstream consensus is now unanimous against it. The only “scholars” who deny this are Christian apologists. Who have been completely sidelined and abandoned on this by their more honest and less delusional peers. The uniform opinion of mainstream scholars is that the Gospels do not fit the genre of historical biography but of mythical biography, otherwise known as hagiography.

The most typical manifestation of which are the fantastical biographies of Catholic Saints that would get farcically spewed out across the Middle Ages. But the most pertinent manifestation of which are the precedents the Gospel authors were following (see Robyn Faith Walsh and the Gospels as Literature): the Jewish novels (like Tobit, Joseph and Aseneth, even Daniel) and the Greek mythographies (like the Lives of Aesop, or the narratives of Osiris, Dionysus, Romulus, Hercules, etc.). This was so decisively confirmed three decades ago by Richard Burridge that it has become the mainstream consensus of experts since. I attended an anniversary panel with him and others where he and they all confirmed this point in no uncertain terms, against growing apologetical disinformation about what he actually argued and proved (see my account in The Burridge Panel).

This is why their entire premise of HRM is bollocks and really just represents historical denialism—outright tinfoil hat—and not a credible basis for formulating evidence in a Bayesian equation. Their entire analysis falls on this alone. The rest is just “other standard shit they pull” that is only worth pointing out so we can make them talk to the hand from now on and shut that shit down. They are the ones who need to stop. Because it’s embarrassing. To outsiders it looks just like the “Earth is flat” and “Adam rode dinosaurs” shit. And they don’t even see it. Yet everyone is laughing at them for a reason.

Their position is even more ridiculous than that, since HRM should entail commonality with ancient historical biography—and yet even historical biography was routinely then fabricated and thus not a reliable comparison point for any of their Bayesian premises. Contrary to their gullible browbeating, there were no constraints on making things up, especially for propaganda, to defend any claims and ideas—and especially for speeches, and the paranormal. Ancient observers pointed this out repeatedly. But modern experts have done so in even more detail. I just mentioned the work of Robyn Walsh above, for example. But in OHJ I already cited the world’s leading scholars on this point for ancient biography: Mary Lefkowitz, Janet Fairweather, and Barbara Graziosi. This is the mainstream consensus in ancient biography studies.

Hence even treated under the HRM (a model there is no evidence the Gospel authors followed but abundant evidence they didn’t, as already noted), we know the Gospels are unreliable. The entire trial sequence is anachronistic (essentially impossible), the depiction of the Pharisees is anachronistic (Jesus is often made to take the position against them that they actually held); they are full of improbable contradictions, and outright falsehoods, including a completely refuted three-hour darkening of the sun, and Luke’s fabrication of the wrong chronology of events for Jewish rebels, and for Paul’s debates and travels (see How We Know Acts Is a Fake History). The list is vast. So even if they were following HRM, they are so full of lies and fabrications, or at least an overwhelming cloud of suspicion of it, that their Gospels would never be admitted as evidence in a court of law. They should not be admitted into ours.

Fabricating 21 Eyewitness Affidavits

But the most spectacularly ridiculous claim they pull out of their universally rejected premise of HRM is that we have “at least 21 distinct reports” of “multimodal sensory experiences” of a risen Jesus (or what I affectionately call “cuddle dinner party Jesus,” based on Acts 1, John 21, and Luke 24). Their argument requires that this mean 21 sworn affidavits written or dictated by eyewitnesses. Because otherwise their probabilities don’t follow. There is nothing improbable about some random dude claiming on the street that a million eyewitnesses saw Godzilla rise out of the sea. That’s not “one million eyewitness accounts to a multimodal sensory experience of Godzilla.” That’s not even one such account. It’s an unsigned, unsourced claim patently absurd on its face, contradicting all prior evidence, and typical of a commonplace urban legend genre.

We don’t even have one eyewitness account (Paul never gives us “an account”). We have only one eyewitness report, but as I just explained, it uniformly conforms to standard religious experience psychology. It contains no reference to (and rather the conspicuous absence of) any “multimodal sensory experiences” of a risen Jesus. So their evidence at this point is simply fabricated. No honest Bayesian result can follow from that. What they have are late fantastical legends in a foreign language for foreign audiences, and no comment on them at all from anyone who was actually there—not even to endorse any of it. And even those all embellish just one legend: the so-called Gospel According to Mark. We have no independent sources at all. Just Mark, and redactors of Mark, and eventually a wildly implausible rewrite of Mark. And that’s that.

That is all 100% expected on any natural cause theory (after all shared contingencies cancel out: see Proving History, index, “coefficient of contingency”), much less all of them. And you can’t get any probability higher than 100%. So there is no way to get a Bayesian likelihood ratio out of this that favors a supernatural resurrection. That result is simply impossible. Not because a supernatural resurrection is impossible; but because there is no evidence favoring it over more common and typical causes of similar evidence throughout world history. They need evidence (not a claim; evidence) that Mark didn’t make it all up, or that anyone after him had any source independent of him. And there isn’t any. To have that you’d need those “21 sworn eyewitness affidavits,” vouchsafed in some chain of custody we can be confident was free of alteration (thereby meeting the “ancient documents” rule of evidence in any modern court), and sufficient evidence to be confident those weren’t faked to begin with, or weren’t lies to begin with. That’s just not the kind of evidence that exists here. And that’s that.

Ignoring the Expected Rate of Defections

They then claim the “skeptics” Paul and James (the so-called Brother of the Lord) were convinced by appearances of Jesus, but that’s fudging. There is no evidence anywhere in the New Testament that any James was an unbeliever when Jesus died. And we have no accounts of him being converted from a skeptic by the risen Jesus—at all, much less by a “multimodal sensory experience” rather than just an ordinary dream, ecstasy, inspiration, or hallucination. Whereas a brother of a movement-building charismatic would be very inclined to inherent that mantle when opportunity arose, so even on their theory there is no unlikelihood of his claiming whatever he needed to on any theory contrary to theirs.

But even 1 Corinthians 15:6 never says the James there named was this James, or was an unbeliever at the time. So this is a fabricated line of evidence. A “maybe” illicitly converted into an “actually,” a very common Christian apologetic trick (the fallacy of possibiliter ergo probabiliter). And you can’t input fabricated evidence. No theory has to explain evidence that doesn’t exist. Which leaves Paul. Just one guy. Who tells us his experience was mystical and internal, not a “multimodal sensory experience” (much less experienced at the same time as anyone else). But let’s pretend we had two guys anyway.

They argue that “this is what we would expect to see given the resurrection” and not otherwise, and “therefore” it is (by their arbitrary calculation) 800 times more likely that way, and so that’s what they plug into the equation. But this is erroneous. That only two skeptics would defect (having or claiming inner experiences convincing them to) is actually not what we expect on a real resurrection but exactly what we expect on it not having been real (see Dating the Corinthian Creed and Daniel Bonevac’s Bayesian Argument for Miracles). Defectors, indeed defectors claiming to have had marvelous experiences convincing them to defect—indeed defectors actually having such experiences (dreams, inner voices, ecstasies, hallucinated auras, or whatever)—are not unusual in the history of religion. You can find many cases, of people being convinced by religious experiences to defect to practically every mystical religion on the planet (I had one as a teen convincing me of the supernatural truth of Taoism, as I’ve discussed many times before), including from hostile positions: Muslims having a vision of Christ becoming Christians; Christians having powerful dreams or inspiring moments convincing them to commit to Islam; even people abandoning atheism or Christianity for Hinduism after visions of Hindu gods and goddesses.

In fact, every religion has examples of defectors claiming religious experiences that moved the needle for them. They are simply not common. And here is where Christopoulos makes a mathematical mistake prior to the equation: the likelihood of an observation is not the frequency of an occurrence but the frequency of the observation. The likelihood of “two defectors” claiming visions is not the frequency of a defector claiming visions—which could be, say, 1 in 1,000, and thus 1 in 1,000,000 to get two of them, or whatever the ratio is of skeptics converted by a purported mystical experience to skeptics not converted. Rather, it is the frequency of observing at least two defectors in any given religious movement—which is effectively 100% for all religious movements (rounding to the nearest whole percentile). In other words, that a tiny rare few and no more get so converted is exactly what we expect from a false religion. A true religion should produce a far higher rate than that—a miraculous rate of mystically converted defectors; not an ordinary rate.

Which means their evidence of “converted skeptics” is actually evidence against resurrection, not for it. Their likelihood should flip the other way around (becoming 800 to 1 against resurrection) or worse. And here is the point again that I’ve made already: this should also happen for all the other appearances. Because a real resurrection should have produced waaaaayyyy better appearance evidence than we have, while what we actually do have is exactly what we expect on mere ordinary natural phenomena (religious psychology and mythmaking). So its likelihood ratio should also flip the other way around or worse. Because the actual evidence is of estatic experiences under altered states of consciousness to a primed and desirous group inspired by a charismatic leader—and not “multimodal group hallucinations.” Mass ecstasies are normal. Mass multimodal group encounters are fiction. And that’s exactly the order in which we find those claims in the Bible: our one eyewitness attests the former; only much later foreign legends attest the latter. That simply isn’t the order of evidence a god would produce—if producing evidence were what he wanted to do.

As I explained over a decade ago, in my own definitive Bayesian refutation of the resurrection in The Christian Delusion (and see also my Bayesian extension of that argument to what happened after the religion started, which was also unextraordinary, in The End of Christianity):

Only an ordinary explanation can easily explain why Jesus only appeared to die-hard believers, and then, much later, to only one of millions of outsiders across the entire planet. If God himself were really appearing to people, and really was on a compassionate mission to reform and save the world, there is hardly any credible reason he would appear to only one persecutor rather than to all of them. But if Paul’s experience was entirely natural and not at all divine, then we should expect such an event to be rare, possibly even unique—and, lo and behold, that appears to be the case.

Paul’s conversion thus supports the conclusion that Christianity originated from natural phenomena, and not from any encounter with a walking corpse. A walking corpse—indeed a flying corpse (Luke 24:51 and Acts 1:9–11) or a teleporting corpse (Luke 24:31–37 and John 20:19–26)—could have visited Pilate, Herod, the Sanhedrin, the masses of Jerusalem, the Roman legions, even the emperor and senate of Rome. He could even have flown to America (as the Mormons actually believe he did), and even China, preaching in all the temples and courts of Asia. In fact, being God, he could have appeared to everyone on earth. He could visit me right now. Or you! And yet, instead, besides his already fanatical followers, just one odd fellow ever [then] saw him.

If Jesus was a god and really wanted to save the world, he would have appeared and delivered his Gospel personally to the whole world. He would not appear only to one small group of believers and one lone outsider, in one tiny place, just one time, two thousand years ago, and then give up.

So their evidence E2 (appearances) is evidence against, not for, the resurrection, and so its likelihood should be flipped or worse, and thus their 10,000 to 1 in favor should be 10,000 to 1 against; and their evidence E3 (only one or two defectors) is also evidence against, not for the resurrection, and so its likelihood should also be flipped or worse, and thus their 800 to 1 in favor should be 800 to 1 against, for a combined 8,000,000 to 1 against. Their own likelihoods should signal that the evidence they have is eight million times more likely if the resurrection didn’t happen than if it did.

So even their absurd priors can’t save them here. But because that’s also a common and outrageous game typical of crank Bayesianism, we’ll close with that anyway, just to beat their dead horse into lunch for my aged tabby cat Ophelia.

Crank Bayesianism 101: Fake Your Priors

I won’t waste any time on their distracting nonsense about Hume. Arguing over what some three hundred year old dead guy said or meant is a waste of everyone’s time. The only things that matter are what is true now. Not what some long dead guy said. They don’t correctly describe what Hume said or meant anyway; and what they do argue is mostly the very thing their own citation of John Earman meant by “vapid” (no one ever actually reads John Earman’s book). What matters though is not whether that’s true—everything they said about that could be 100% true and wonderful, and it makes no difference. Because what they end up doing to determine the prior probability of a supernatural cause is gullibly absurd.

The prior odds are the odds given only “k = relevant background knowledge” as they put it, and thus not yet considering any evidence particular to this one supernatural claim. Background knowledge means everything we know—you can’t leave any human knowledge out. So it includes biology and physics (and hence the observed implausibility of supernatural beings and powers, and the utter lack of evidence they exist or operate), and ancient comparanda in religious conversion, experience, and mythography (indeed even the forty or so other Gospels and thirty or so Acts that are by all agreed to be fake—faking things was the overwhelming norm for Christian authors then), and the relative frequency of lying, exaggerating, mythmaking, gullibility, error, and the like, across human time.

The prior odds here are functionally the base rate of supernatural causes of outcomes in the world. Which basically means magic. Whether a god’s magical powers or a demon’s or a wizard’s or a psychic’s, doesn’t matter here. Even if you pick the reference class of people who wake up after being pronounced dead, the base rate of any kind of magic causing that is vanishingly small. If you pick the class of people who say a god came to them in a dream, waking or not, or talks to them from time to time, or shows them dazzling lights or weird phenomena, the base rate of an actual god causing it is vanishingly small. We know full well such things have far more common causes in nature. When we can check—actually run the evidence to ground and confirm a cause—magical tales are almost always fabricated. And the ones that aren’t are always the product of error, someone mistaking a natural phenomenon for supernatural.

In fact we have no evidence that even the mechanism exists—or even could exist, in the way we can prove Dysonian alien visitors can (but just probably don’t). The number of supernatural causes ever discovered for anything is zero. After centuries of extensive scientific inquiry over the entire observable universe, Naturalism Is Not an Axiom of the Sciences but a Conclusion of Them. Worse, supernatural causal theories require massively convoluted and wildly implausible suppositions far more ad hoc and needless than even the rarest and thus most improbable natural causes attested (like heists or cons), indeed than even unattested natural causes (like psionics or aliens), and thus suffer from The Argument from Specified Complexity against Supernaturalism and the fact that Christianity Is a Conspiracy Theory. It’s like Adam riding dinosaurs or the Earth being flat. It’s simply a ridiculous theory. And that is not an a priori supposition. It’s the universal conclusion of all background knowledge to date.

Nevertheless, they are correct to note that the prior probability of the supernatural cannot be zero. Paulogia never said it was, so they were distracting their readers by arguing something hardly anyone ever claims. But some few do get this wrong. They are just rare (or amateurs) and thus the straw men, not the steel men of what people like Paulogia and I argue (see How Not to Be a Doofus about Bayes’ Theorem and Correcting 5 Mistakes Atheists Make About Epistemology). But Christopoulos commits to the non sequitur of “it’s not zero, therefore it’s absurdly high.” Which is even dumber than thinking it’s zero. So he kind of foots his own mouth here.

This is where they make farcical claims like “we granted the lowest prior,” in fact “even with the lowest possible prior,” the “the lowest plausible prior imaginable,” of “1 in a million.” Wait. What? An event happening to someone that is as rare as 1 in a million per day happens 8,000 times a day on Earth right now. That’s not even rare. If miracles were that common, they’d be routinely proven facts of science. We’d have miracle wings in hospitals, actual magic-working wizard guilds, and the like. The supernatural is provably nowhere near that common. More competent Bayesians don’t make this mistake. Daniel Bonevac (as cited above) set this prior at “1 in 10 billion,” which is something that happens to about one person every day, and thus still absurdly high for a thing never ever confirmed to even be plausible much less happen. You would not say the prior probability that gremlins downed Flight 370 is anywhere near as high as “1 in 10 billion.” Or accept that as the prior probability that someone on trial for murder cast a spell to magically alter the memories of witnesses to report them as somewhere else at the time. Or as the prior probability that the Angel of Death committed that murder at God’s behest.

Hence even the gullaballoo Christian apologist Lydia McGrew set this prior far more credibly at 1 in 10⁻⁴³ in her own Bayesian case for the resurrection. That’s a ten followed by forty three zeros. An event that rare happens only once per entire galactic lifetime. That’s a more realistic estimate. In my forthcoming book From Ontology to Morality I settle on less than 1 in 10⁻²³ which is more generous but still more realistic. To say it’s 1 in 1,000,000 is simply ridiculous. And based in no evidence or logic. For comparison, about one art heist valued over a million dollars is successfully completed every three years globally—and those are easily millions to one against (counting all the million-plus dollar art that is not stolen compared to what is). And “million to one” is a routine event (once every three years for expensive art). Indeed, lotteries are routinely won that have lower frequencies of success than that. And “an angel ate my homework” is in no way as common as people weekly winning lotteries. Or art heists.

If we imagine the resurrection as a comparably difficult heist, and you say “but it’s so unlikely that would succeed,” you’re right. Successful million-dollar art heists are also “so unlikely to succeed.” But succeed they nevertheless sometimes do. Given enough time, eventually one succeeds. Hence the probability that all the evidence of the resurrection is actually the product of a perfectly natural heist, one that hit all the holes in the swiss cheese and thus was brilliantly pulled off, is vastly greater than “magic.” So whatever low prior odds you’d put on “it was a heist,” the odds on “it was magic” must necessarily be lower. Because amazing heists are documented known facts. Magic is not. A magical resurrection is millions if not billions of times less common than even a clever million-to-one heist being pulled off—much less far more common and thus more probable causes like “ordinary ecstatic religious psychology followed by fantastical mythmaking.” The James Randi Prize ran for fifty years with no takers. No one in half a century could prove a supernatural power existed at all, even for a million dollar prize. And all other efforts to find even one provable instance have always failed. Mysterious events, when a cause is found, always end up having natural causes. And that includes some pretty incredible events (see Everything You Need to Know about Coincidences). And that’s way more often than a million times to one. For there to be no documented supernatural causes or powers across all the history of modern science means they have to be galactically rare. Not “amazing art heist” rare.

But notice that even if we granted all their bogus inputs into the likelihoods, for a combined Bayes factor of 8,000,000:1 for the resurrection, this would fail to overcome even Bonevac’s already absurd estimate of a prior odds of 10,000,000:1 against (that would get only a 20% chance the resurrection happened), and would come nowhere near overcoming my or McGrew’s priors (those would get them a probability so far below 1% as to make no relevant odds). This is why it is necessary for these guys to make up a ton of falsehoods about the empirically confirmed frequency of miracles being “better” than events that happen 8,000 times a day, which is the most absurd claim I have ever heard.

And That’s Not Even the Whole Problem

Still, if we had confirmed 8,000 miracles a day by now, they’d be in a better place, but they’d still have some hurdles. Ancient sorcerers were reputed to be able to raise the dead. It was one of the standard list of things expected of a real mage. So if today sorcery were normal, if we had sorcery wings in hospitals, and sorcerers routinely produced evidence accepted as reliable in courts of law, then we could get a decent prior for a resurrection. Except then, even if we proved all the rest, we’d still have to rule out mere sorcery to rule in a god stepping in. So if a god wanted to prove it was them and not just someone else’s demonic or natural magic, he’d have to do something they could not do, something that shows godlike power unachievable by any mere demon or wizard. That entails something far more spectacular than what the Gospels report. If you even want to prove it’s God and not Ardra using completely natural physics from a hidden spaceship in orbit, you need something Ardra cannot do. And Ardra can do even far more marvelous things than the Gospels imagine for Jesus. What could God do better? Like rearrange the stars to spell out the Christian creed when viewed from Earth, in a language everyone can miraculously read. Or have the risen Jesus fly down and make his sales pitch to every adult ever born, throughout all recorded history. Or, hey, resurrecting not just one guy under dubious circumstances but every recently deceased person on the planet, who then report their brief experience in paradise above or purgatory below, every one’s account corroborating the other. You know, actual godlike stuff.

But we can’t even prove any magic exists, much less godlike magic. We can’t even prove magic is plausible—like we can Ardra’s miracle-faking spaceship. God arranged the world so that is plausible but miracles are not. On any assumption of theism he chose to do that, take note. And either that or the nonexistence of God is why all opportunities to find any at all have failed. All we get are myths, legends, frauds, and unverifiable claims in distant jungles. All we get are “the cataracts on Sam’s mom’s eyes.” Which is extremely unlikely unless that’s all there is.

To hold out hope for something still is way worse than holding out hope of winning the Mega Millions lottery, which actually happens routinely, and thus far more frequently than miracle stories turning out to be real, or even alien saviors, which must then have odds billions of times worse. And yet winning lotteries has odds against worse than 1 in 290 million (at which as a prior even their faked evidence would get a probability for the resurrection of not even 3%). So real miracles must be less common explanations of miracle tales than Mega Millions lotteries or alien rescues. McGrew and I conclude it must be a lot less common. Bonevac was thus being ridiculous. And yet these guys couldn’t beat even his absurdly high prior for miracles. They have no chance of beating ours. And that’s even with the evidence they faked up. When we fix all that faked evidence back to what really actually exists (and what really actually would have), they can’t even beat a modest prior, much less an extraordinary one.

Conclusion

So not only are their priors baloney (and baloney in, baloney out), but their likelihoods are baloney. A 1 in 1,000,000 prior odds of the supernatural existing in the world is downright comical.

Far more plausible priors for supernatural causes existing at all range between

• 1 in 100,000,000,000,000,000,000,000

and

• 1 in 10,000,000,000,000,000,000,000,000,000,000,000,000,000,000.

That’s the expected frequency of supernatural causes of observed effects in the world, given the entire history of all observed effects, and all observed causal mechanisms, down to the quantum level. Supernatural causes just aren’t credible. They are always the least likely explanation of anything. And your priors must reflect that—and you can’t do that honestly with laughably absurdist attempts to claim half of all miracle claims are supernaturally caused, much less even as many as one in one million of them, a frequency so high that supernatural events should be a routine part of our experience—in which case that prior would be credible. It’s just, alas, we don’t observe ourselves to be in that world. Which incidentally would have to be by God’s choice. So even if God exists, he has telegraphed to us that he does not want us to use miracle claims as evidence of it, or of anything. If God wanted that, he’d have given us the rationally appropriate evidence. He didn’t. So Christian apologists are defying God’s will by falsely pretending he did.

It’s just more likely, of course, that there is no God. It’s even more likely wizards or demons or psi-mutants rose Jesus from the dead, since as a supernatural effect, it is far more congruent with those far weaker causes than a cosmic god. And here “far more congruent” means, literally, “having a far higher likelihood.” But natural causes of all observed effects are far more likely than demons or psychics or wizards—or even aliens, who are far more likely than demons or psychics or wizards. Inventing elaborate just-so stories to explain why God behaves just as feebly and ineffectively as mere aliens or wizards is why Christianity Is a Conspiracy Theory.

But natural causes have far higher combined likelihoods than even aliens or wizards, much less gods. These apologists have instead invented evidence that doesn’t exist (like 21 eyewitnesses to multimodal sensory experiences of a risen Jesus), they left out evidence that does exist (like the actual chronology of the evidence’s exaggeration over time), and they illogically derived their likehoods even from the evidence they did present (like failing to account for the vast scale of skeptical non-conversions, and the high natural frequency of rare skeptical conversions across all religions, which together completely reverses the effect of skeptical conversions into evidence against resurrection rather than for it).

This is all just game playing. Rational people are never convinced by this. They can see immediately through every piece of this bullshit. Everyone knows these priors are ridiculous, that they are just listening to some tinfoil hatter insist celestial gremlins and faerie magic are real, “and if you would just assume they are real, we can prove even weirder things!” And everyone knows they are faking the evidence, that there aren’t “21” eyewitnesses to what they say—there isn’t even one such witness, and the one and only witness we do have, supports no such observation having occurred. And everyone knows they are leaving evidence out, and deriving bogus frequencies of the evidence they include.

This is all obvious. Except to them. And that is the only reason they still believe this. And for that very reason, because they are too delusional to listen to what we are telling them, or understand it, they’ll just respond to articles like mine with guffaws and rage and browbeating and probably a lot of insults and belittling—and ever more bullshit. Dollar to a doughnut. You just watch.