# [FoRK] That's a bunch of Malarkey

Stephen D. Williams sdw at lig.net
Mon Jan 19 12:34:58 PST 2015

```On 1/19/15 10:58 AM, Gregory Alan Bolcer wrote:
> On Mon, Jan 19, 2015 at 10:28 AM, Stephen D. Williams <sdw at lig.net> wrote:
>
>> Inifinte payout?  Are you sure?  Any different kind of existence is at
>> least somewhat beyond an event horizon.  All the ones I can imagine are
>> qualitatively less than this one.  The most logical mapping would be to an
>> upload, but that is a cheap existence.  In any case, the new you is not
>> going to be you in any meaningful way. And we'll likely create upload
>> ourselves soonish anyway.
>>
> Yes, I'm 100% sure the game theoretic representation of Pascal's wager
> shows an infinite payout in one quadrant.  This is by definition.  [1]
>
> Again, you are arguing some different experiment and not Pascal's wager.
>
> If you change the rules or the payout, it's not Pascal's wager.

Partly, I'm arguing that Pascal's Wager is fundamentally flawed and that simplistic player logic is also flawed.  It is begging the
question.
https://en.wikipedia.org/wiki/Pascal%27s_Wager
>
> The wager uses the following logic (excerpts from /Pensées/, part III, §233):
>
>  1. God is, or God is not. Reason cannot decide between the two alternatives.
>
Weak, but we'll take it as true for now.
>
>  1. A Game is being played... where heads or tails will turn up.
>  2. You must wager (it is not optional).
>  3. Let us weigh the gain and the loss in wagering that God is. Let us estimate these two chances. If you gain, you gain all; if
>     you lose, you lose nothing.
>
You lose nothing?  Not true.  In various ways, you lose or waste your life.  Perhaps barely measurable for a few people, but
all-consuming for others.  And there are at least subtle effects. On average, I see significant effects on most people.

>  1. Wager, then, without hesitation that He is. (...) There is here an infinity of an infinitely happy life to gain, a chance of
>     gain against a finite number of chances of loss, and what you stake is finite. And so our proposition is of infinite force,
>     when there is the finite to stake in a game where there are equal risks of gain and of loss, and the infinite to gain.
>
"Inifintely happy life to gain" is a childish concept.  What life could be infinitely happy?  What would I do with an infinitely
happy life?  The sell is unattractive and not believable at all.

I would also argue that as believing may hamper a life in a way that costs the cognitive margin that may make all the difference
between a good life and a waste, and that this life may be your only shot, the loss may be effectively total loss.  I.e. infinite loss.

Believing isn't free, which seems to be the assumption.  It one of the biases in the question.

>  1. But some cannot believe. They should then 'at least learn your inability to believe...' and 'Endeavour then to convince'
>     themselves.
>

>>
>> Pascal's Wager is the game theoretic decision making of a single person,
>> not all humanity.
>
> This is interesting.  From my game theory course on finite and infinite
> games, there is no difference between a finite game played by one person or
> a finite game played by multiple persons or repeatedly if the rules are the
> same.  That's even true for multi-person dilemmas.
>
> Are you saying a prisoner in a repeated prisoner's dilemma will change his
> strategy to be less than optimal?  [2]

> One question though, I think your calculations for "an infinite amount of time" are a little off. Assuming there exists at least
> one human in each (in your words) "arbitrary life-long adherence requirements" and human life is not infinite, then wouldn't that
> number be bounded and finite?
In this case, the fact that someone else benefits doesn't help with my decision.  If all of the flaws of the question were cured and
there really was a worst-case scenario to avoid, it might make more sense.

This is the wager I hear:
"You can choose to remain free and have an unencumbered, potentially great and beneficial life.  Or you can choose to become a slave
during this life on the chance that there is another life and that by becoming a slave to the right master, you will be rewarded
with a better and longer life.  Our master, which we say is the right master, says so, according to distant ancestors."

>
> You mentioned recursion.  In repeated or concurrent games, each subgame is
> computational identical to the original game.

I meant recursion up to the question of belief at all.  Why recursion.

>
>
>> First you have to choose the right quadrant.  If you choose to act as if
>> there is a heaven / religion / God, you must first choose which heaven /
>> religion / God, among many (and potentially an infinite set).  If you
>> choose poorly, then even though you chose to believe, believing in the
>> wrong thing would have been futile (according to the beliefs of most
>> religions).  Pascal's Wager presupposes that you are already considering
>> the one true religion.  According to the rules of each of those religions,
>> that can't be true of everyone's religion.  There are a finite number of
>> religions currently, but based on their rules of evidence, over time there
>> may be infinite alternative religions with just as much evidence.
>> Therefore, the odds that you've chosen the correct religion are vanishingly
>> small.  Even if you choose to believe and otherwise invest to be proper
>> with one religion, you can't do so in many more than one religion.
>> Therefore, as an individual, there's no way to cover all of your bases
>> short of infinite time.
>>
>
> Again, change the labels.  You can avoid all the rest of your philosophical
> discussion simply by doing that.

You have a good point.  Let's try:
http://rationalwiki.org/wiki/Pascal%27s_wager
>
>
>     Reversing the wager
>
> One way to counter the wager is to replace Pascal's Judeo-Christian God with a perverse god that punishes those who believe in him
> without evidence, and rewards those who don't. Note that this doesn't even presuppose that the Bible and other holy texts are not
> divinely inspired: this god could have authored them to serve as a test of credulity. Importantly, because Pascal's Wager can only
> work if you improve the prior probability of any one god's existence over the others (which generally doesn't happen in /any/
> theological argument) this wager is exactly as valid as Pascal's original formulation.
>
[table]
> This theoretical belief system presents a win/win scenario for atheists and a lose/lose scenario for those who believe in God.
> Since the two contrasting ideas of a specific god are logically equivalent in likelihood, atheism is shown to have the greatest
> potential for gain, completely negating and effectively reversing Pascal's argument. Remember, this is equally empirically
> provable as Pascal's Wager, and so we now have to factor in a 50:50 chance of Pascal's Wager being true or this one being true.
Then the next step, assuming that you can decide to believe:
>
>
>       Further reversal
>
> It is dubious whether people can force themselves to believe whatever they want. Instead of considering whether one should believe
> in God, this reversal considers whether it's a good strategy for an otherwise good, socially responsible person to pretend to
> believe in God:
>
[table]
> As before, feigning belief is never a better strategy than being honest about your disbelief, unless the punishment for feigned
> belief (trying to trick a god) is milder than the one for disbelief. True believer is left out, because you're arguing from the
> viewpoint that belief is something that takes place deep down inside, and is not a facade, and that you simply lack it.

>
> As to the model, logic, and math, didn't you ever take set theory?  Number
> theory?  Finite and infinite equivalence classes?  Maths is maths.   The
> smallest finite percentage you can imagine times infinity is still
> infinity.

Infinity * zero = zero, not infinity.  If there is a long chain of probabilities, all approaching or at zero, it may be that it
becomes likely that one of them is zero.
Depending on what determines the probability, it may involve an infinite term in the denominator, which reduces to a zero
probability term.  Given that the infinity conjectured here is probably not infinite, the result doesn't look valid.

>
> If you don't like Pascal's model, then change the model, but don't call
> your new model Pascal's Wager.   I think you are talking about the mapping
> of the model to your belief system and not the actual model.

Pascal's Wager is flawed and overly simplistic, even though it is easy to understand.  We're left with the situation where a fatally
flawed argument is used usefully by anology:
http://rationalwiki.org/wiki/Pascal%27s_wager
> While Pascal's Wager is a theologically flawed argument, the principle has applications in other aspects of life, such as
> investment and betting - purely because the game theory aspects are reasonable - and Pascal's Wager is brought up by name as an
> analogy where such gain/loss comparisons can be made. In short, it presents a way of dealing with uncertainty by assessing not
> /probability/ but only the magnitude of potential gains against losses.

>
>
>
>> Most people rely on various fallacies to unconsciously justify their
>> assumptions.  Principally, that they are part of a large group of
>> like-minded believers, so theirs is likely the one true religion. In honor
>> of the biggest mistake that lasted the longest even in the face of evidence
>> obvious to everyone, let's call that the Flat Earth fallacy: If everyone
>> believes it, it must be true.
>>
>> Besides, everything related to this fails Occam's Razor.  The resulting
>> probability is the result of the product of many fractional probability
>> terms, many of which are likely 0.
>
>
> Occam's razor is a scientific principle, aka a philosophical heuristic to
> problem solving.   Pascal's payout matrix is solved using a computation.
Pascal's Wager centers on a computation where the assumptions and values of the terms, let alone the very axioms of the question,
are religious dogma.
Religious mumbo jumbo, even when expressed as a computation, is still religious mumbo jumbo.

>
> Greg
>
>
> [1] Are you sure you are arguing apples to apples?  I'm having a laugh
> poking fun at your definitions; you are arguing philosophy.
>
> [2]
> http://www2.isye.gatech.edu/~jswann/teaching/6230/6230%20Swann%20Infinite%20repeated%20Friendman%20Cournot%20Wage%202010%20%28L8%29.pdf
>
sdw

--
Stephen D. Williams sdw at lig.net stephendwilliams at gmail.com LinkedIn: http://sdw.st/in
V:650-450-UNIX (8649) V:866.SDW.UNIX V:703.371.9362 F:703.995.0407
AIM:sdw Skype:StephenDWilliams Yahoo:sdwlignet Resume: http://sdw.st/gres