Saturday, 1 November 2008

the probability of god

I recently read an interesting and fun book. It was The Probability of God by Stephen Unwin, and it attempted to show how one could use various facts about our world to compute the probability that God exists.

Unwin works as a risk assessor for large US companies, and uses the mathematics of Bayesian probability in his work, so he thought he would try to use the same ideas to examine God. He considers factors (or "evidentiary areas") that suggest God might be more likely to exist, such as how we know what's right and wrong, miracles and religious experiences, and other areas that might suggest God is less likely to exist, such as the problems of natural and human evil in the world.

Unwin assesses the evidence and ends up with a probability of God 's existence of 67%.

Of course we shouldn't take it too seriously. The probabilities have no factual basis, and are simply his subjective assessment. But I found the exercise interesting, because it allows everyone to use their own estimates of probabilities and come up with their own answers. And allows each of us to test our belief against something that at least tries to be even-handed.

In fact I found it so interesting that I wrote a program that allows the test, upgraded with a few more questions, to be taken on-line by anyone who wishes. Testing the program indicates you can get any answer, from 0 to 100%, so it doesn't seem to be biased. So if you're interested, you can make your estimate of the probability of God on my website. I'd be interested to hear what you conclude.

For the record, I calculated the probability of God to be 98%, which I guess tells you something about my answers.


  1. Ah. They didn't split out the idea of a creator-God from a benevolent manager-God who answers prayers. So my result isn't valid, as I would get sharply different answers depending on whether you're talking about a creator or a manager.

  2. I'm interested to know in what ways that would make a difference. I would have thought either would fit the test.

  3. Ok, here’s my breakdown of the questions, with explanations of whether I think each one treats deists the same as theists (as it should since they believe in God), or scores them more like atheists.

    1. How likely overall do you think God is?: Makes no difference whether it's a creator or a manager.

    2. "Big bang" There may be a flaw in this question, I think. You say "If there is no God, how likely is it that the universe would have begun out of nothing" but "If there is a God, how likely is it that the universe would have been created" without specifying "out of nothing" or "as proposed by the big bang. “What is the probability the universe was created" conjures images of non-big-bang creation for lots of people, so they're going to say that's higher than "the universe came from nothing". Perhaps the question needs to be reworded, as I notice for the rest of them the two questions basically mirror each other. 3. "The universe seems designed".

    3. This one hinges on belief in a designer God. So my answer here probably pushed up my overall score just as much as it would for a theist.

    4. "Natural Disasters". This one is a wash, as believing in either a creator-god or manager-God would require you to justify natural disasters.

    5. "conscious beings". Theists would score higher on this one. I think it's about equally likely that human beings would arise under the laws we have, regardless of whether God made them or not. But theists would probably say that conscious beings like humans are less likely to arise under a Godless universe. It appeals to a notion that human beings are special, consciousness makes us special, and it's divine in origin. Not believing in a direct divine origin for humans doesn't mean you don't believe in the possibility of a divine origin for the universe.

    6. "Logical thinking creatures" The same basic rationale exists for this as for #5. If you think the human mind is divine in origin, which is linked strongly to the "God made us in his image" idea, you're going to answer more highly on this, in that it’s unlikely a Godless universe would contain thinking beings.

    7. "Creatures with sense of right and wrong" Same as 5 and 6. Basically what this boils down to is "Do you agree that humanity, and not just the universe in general, was created directly and intentionally by God?"

    8. "Evil acts of humans in the world" I’m not quite sure how theists would answer this. To me it would seem intuitive that evil is more likely under a Godless universe, but somehow theists don’t see it that way, I guess, if the Bayesian revision works like it should.

    9. "Science can't detect God?". This would be answered about the same by any kind of believer, regardless of what kind of God you believe in.

    10. "Miracles". A theist would answer that these experiences are very likely if God exists, and very unlikely if he doesn’t. A deist would answer the same as an atheist, that they’re about equally likely regardless of whether there is a God, because God’s not doing them anyway.

    11. “Religions” A theist would answer that religions are likely under God but not likely otherwise, whereas both a deist and an atheist would tend to answer that it’s equally likely regardless of whether God exists.


    1. No effect
    2. Possibly flawed question
    3. No effect
    4. No effect
    5. Theists score higher
    6. Theists score higher
    7. Theists score higher
    8. Unsure, possibly favours deism
    9. No effect
    10. Theists score higher
    11. Theists score higher

    Theists score higher than deists: 5 (45%)
    Unsure: 2 (18%)
    Treat deism the same as theism: 4 (36%)
    Rounding error: 1%

    My total score was 26%, while I would put my level of certainty at somewhere around 60-75%.

  4. Myron

    Thanks for your feedback. I will have another look at question 2 to make to improve it as you suggest.

    But I'm inclined to think that the test fairly assesses the differences between deism and theism. A deist doesn't believe in religion or religious experience, so questions 10 & 11 don't add to the evidence, and a deist may end up with a lower score than a theist.

    For other readers' benefit, we have transferred further discussion to the reason & belief forum associated with the probability of god test. Hope to see you there!


Note: only a member of this blog may post a comment.