Probability Analysis of "code 19"

Let us try to analyze the "mathematical miracle" of the Quran from a statistical point of view aiming to be somewhat objective. The only real phenomen in "code 19" we can directly investigate using statistical methods are the counts of the initials which occur at the start of certain suras. Combining the counts of suras with similar initials we get one set of data which is checked against 19 divisibility.

Out of 14 sets we definitely have 19-div with at least 5 sets which are "Q", "YS", "KHYAS", "'ASQ" and "HM".

The counts for the "Alif" vary from Quran to Quran and no one has been able to verify the counts for these suras with any old copies of the Quran. These suras must be excluded as anyone can come up with various claims.

Basically we only have 5 sets out of n=14 sets, each set with the probability p=1/19 of being exactly dividable by 19. What is the probability that at least 5 sets would be dividable by 19 ? We can use the standard binomial distribution to investigate this:

sum (l=5 till 14) of n!/l!(n-l)! (p^l)*(1-p)^(n-l)

... this gives us 0.000541191227342443 or that there is only a 0.0541% chance of getting this combination as we see in the Quran compared with a random random selection of letters.

0.05% .. this is pretty low, in other words only 1 book out of 2000 books would give us this structure.

0.05% seems interesting but hang on a minute .. what about the method or rather methods of counting ? Have we used one consisten strategy to count the letters ? Are there other equally possible counting alternatives ? The above counts which gave us 5 sets out of 14 are based upon the fact that the letters from the "bismillah" are to be included in the counts.

In other cases, when individual words are counted, the bismillah is not included. This means that there are at least two ways the letters could have been counted: With or without bismillah. Counting without the bismillah would have made the system more consistent with other stuff but nevertheless _experiments_ led to the conclusion that the bismillah must be included.

Using both the two feasible alternatives letters could have been counted and one could simply pick the particular counting strategy which gave the largest number of 19-div sets. The probability for this is effectively double the above probability: 0.00108208956674033 to be exact and this gives us a 0.1082% chance. (The figure 0.001082... can be calculated as P(A or B) = P(A)+P(B)-P(A and B). Ie we need to subtract the probabilty that BOTH methods of counting would give at least 5 19-div sets. The probabilty for this is very low 2.9288794455242e-07, although it _is_ subtracted from the figure above the effect is very small).

With the possible choices of counting, 1 book out of 1000 could have led to a pattern similar to the one in the Quran.

One can argue about other methods of counting which could also lead to a 19-div pattern. If one was not able to obtain 19-div with/without the bismillah then for example other strategies could be tried such as excluding the counts of the letters of the initials themself, checking the 19-div in each sura instead of sets-of-suras etc.

In the latter case (19-div in each sura) we only witness that definitely 3 out of 29 suras have 19-divisibility. The suras are "YS", "KHYAS" and "Q". There is no doubt about the 19-divisibility in each initiated sura would also have generated a "miracle" .. but this only happens for 3 out of 29 cases (or 4 if we include the special case of "'ASQ").

The probability for this happening (n=29 sets, p=1/19) of at least 3 suras being 19-div is 0.194422108093635. We have a 19.44% (1 out of 5) chance that this would be the case.

But again here too we have two options: to count with or without bismillah. With the different counting strategies the exact probability for this is 35.10%.

These 4 counting methods are what I would consider "reasonable". Any one of them could have been used to show a "good-looking" "miracle":

A: 19-divisibility in blocks of suras, x out of 14 (with bismillah) B: 19-divisibility in blocks of suras, x out of 14 (without bismillah) C: 19-divisibility in each sura, x out of 29 (with bismillah) D: 19-divisibility in each sura, x out of 29 (without bismillah)

With these 4 "comfortable" methods available at hand it is very likely that a pattern could be found. In fact it would be very strange if no structure was found at all.

The "mathematical miracle" generally shown to people follows variant A mentioned above, suras are treated in blocks. However if we further carefully analyze the blocks-of-suras then it gets apparent that this is not really true, one is in fact following both A and C. There is a mix of "tactics" so to speak of, although it seems very consistent to a degree this does have a major effect.

There are only 4 initials which are common for several suras, they are "ALM", "ALR", "TSM" and "HM". Method A has _only_ effect for suras with these particular initials. For the majority of suras who all have different initials (10 out of 14) it is totally _irrelevant_ whether one goes for A or C. Selecting A or C f ex has no effect on the counts of "Q", "YS", "KHYAS" (and even "'ASQ").

For 10 out of 14 initials one has NO choice but is forced to count the initials in each sura and check for 19-div within the sura to obtain a consistent method.

For the other suras with "ALM", "ALR", "TSM" or "HM" one can investigate for 19-div EITHER in individual suras or combine the counts across the suras. The choice is simple: pick the method which gives the largest number of 19-div sets. The postulates are fine but the probabilities of finding a miracle are hence also increased dramatically.

By deciding to lump the counts an alternative counting strategy is created which works perfectly for "HM" suras but has no influence on the other suras. Reliable counts for "ALM" and "ALR" are not available, "TSM" suras break the pattern regardless how they are included while the rest of the suras have specific initials which can't be combined across suras anyway.

>From this perspective we must look at the probabilities in two parts. "5 out of 14" actually turns over to "1 in 4" and "3 out of 9" (excluding the double case of HM/ASQ).

Along with the two counting methods (with/without bismillah) we get 35.11% chance of that at least 1 block-of-suras out of 4 will be dividable by 19 (standard binomial calculation with subtracted possibilty of both methods giving us the same result).

What about the rest "3 out of 9" ? In this case it should be noted that for sura 38 "S" and sura 50 "Q" wheather one decided to include the bismillah or not is effectless as "Saad" and "Q" are not part of the bismillah and this does not give us two counting options.

This imples that "3 out of 9" again is in reality "1 out of 2" and "2 out of 7".

For "1 out of 2" there is a 10.24% chance that we would get at least one sura which is divisiable by 19.

While the chances for getting at least 2 hits out of 7 individual suras with the two counting methods (with/without bismillah) is 9.51%.

Unfortunately there is nothing special about 35% and 10% lucks at all. The results are "lower" than what one would expect but not significant at all. With limited amount of data there will always be stuff which slightly deviates from the standard probabilities but there is nothing spectacular about what we have found so far. This by no means can be advocated as the "greatest mathematical miracle". When all the data is shown together it can look quite impressive but behind the curtain clever methods have been used to combine the stuff and make it look like something fascinating.

--Short summary:

o Starting with "Saad" and "Qaf" initiated suras we have no options but to count as it is. Here we get 1 sura out of 2. (10.24% chance).

o For the other suras where all the initials are different one can try counting with or without bismillah. It turns that if we include bismillah then we definitely get 2 hits out of 7. (9.51% luck).

o We have some suras where we did not get 19-div but as the initials are common one could try to lump together the counts. This gives another us 1 out of 4. (35.1% chance)

The above list shows the total decorrelation of the real effects behind the "simple" counting strategy: to include the bismillah and to treat certain suras in blocks. (Notice that if there _was_ something special in the structure then it is not excluded and would anyway have come forth as f ex 1 out of 2, 7 out of 7 and 3 out of 4 ... At the very best we would get 2 out of 2 (0.27%), 7 out of 7 (0.00000011187287%) and 4 out of 4 (0.00076733603947%). _That_ would indeed have been something special ...)

Additionally there are also choices of different Quranic scripts which can have certain effects.

With all these possible alternatives: 1: Quranic script 2: with/without bismillah 3: how to treat the suras (blocks/individually) 4: and even "correcting" the Quran it is highly likely that we will find at least something of interest, which may look like being out of the ordinary.

"N"/"NwN":

Let us go back to the case where the suras are treated in blocks, just to take the side of those who promote "code 19":

By modifying the initial "N" in sura 68 the number of sets which are dividable by 19 can further be increased to 6, this is however a very nonstandard way to write a Quranic initial and it has not been proven that this initial has ever been written using the new "NwN" way.

The probability that at least one set out of 14 would have a count exactly one less than required for perfect 19-divisibility is in fact the same as "the probability of at least one set being dividable by 19". This figure can easily be found (using the above formula if we let l run from 1 to 14). It turns out to be 0.530900836937515. So there is roughly a 50-50 percent chance that we would end up with at least one case where the count would be just one short of being a multiple of 19. As the probability is slightly greater than 0.50 it would be unnormal not to find at least one "N"/"NwN" case!

It must be emphasised that if one is willing to "correct" N into NwN then one can also expect these kinds of spellings for other suras. For example writing "ALM" as "ALF-LAM-MYM". If we also start taking such modifications into account then this leaves no doubt that the "miracle" is nothing but a man-made invention.

Comments:

* Can we conclude that there is a pattern with the initial letters as they are used in the Quran _if_ we treat suras in blocks _and_ do count the bismillah ? The probability could have been there but suras with initials TH, TS and TSM actually give us the answer: they do NOT follow the pattern at all, nor are these letters "difficult" to count. If there was a pattern in the Quran then along with these sets (8 out of 14) the chance of occurring that would have been 0.00002654016599% (with or without bismillah).

In fact looking it the proper way around we would have (2 out of 4) and (5 out of 7). Respectively 3.07% and 0.0015507757604201% chance (with or without bismillah).

Now; here too we do get some low numbers ... but this does NOT happen. The pattern is broken.

Using arbitrary methods one can come up with ways to make these counts dividable by 19 but that would not be a pattern set by the Quran .. rather a man-made attempt to create a pattern. As the human mind always keep looking for structures one can easily make it look very impressive, especially for those who know nothing about combinatorial probabilities. It will no doubt be something "wonderful" and "miraculous" for them.

* If "code 19" is a sign of code then it should be very clear,simple,consistent and really follow a system which does not break.

The Quran tells us that in the creation of heavens (astronomy) there are signs for those who believe:

"And one of His signs is the creation of the heavens and the earth and the diversity of your tongues and colors; most surely there are signs in this for the learned." 30:22

"Most surely in the creation of the heavens and the earth and the alternation of the night and the day there are signs for men who understand." 3:190

But what kind of pattern do we see in these signs ?

"Who created the seven heavens one above another; you see NO INCONGRUITY in the creation of the Beneficent Allah; then look again, can you see ANY DISORDER? Then turn back the eye AGAIN AND AGAIN; your look shall come back to you confused while it is FATIGUED." 67:3-4

Total flawless. No disorder. No breakdown. This is not the case with "code 19" which even requires the removal of two verses from the Quran; like God could not protect His Book in a flawless way.

* At best 1 book out of 1000 (or even 2000) does not give a vary large probability. Millions of books have been written and quite a lot of books will hence show the same pattern we see in the Quran if we would just add some non-changing phrase at the start of certain chapters.

The universe we live in is not chaotical. There are patterns and correlations in virtually everything we witness or do. Human speech and writing are no exceptions. With 30+ years of work in "pattern finding" anything can be found.

A probability at highest in thousands does not seem to justify the removal of verses and even "corrections" in the Quran. That would warrant probabilities in at least millions. With the current findings no "sceptic" will ever get convinced that the Quran is from God. With a 35% chance for "HM" set being dividable by 19 and 10% lucks for the other suras being dividable by 19 people will only laugh at you.

"Code 19" with the current probabilities is nothing but an insult to God as the ultimate mathematican. Is that all what God is capable of ?! If you want to see real probabilities then take a look at your own DNA .. . compute the probabilities that a human would be created out from the soup of particles which existed after BigBang .. ..

"What! do they not consider how Allah originates the creation, then reproduces it?" 29:19

"We will soon show them Our signs in the Universe and within themselves, until it will become quite clear to them that it is the truth." 41:53

"And how many a sign in the heavens and the earth which they pass by, yet they turn aside from it." 12:105

Notice that even though the probability of humans beings created out from the "soup" is like 0 _vastly_ exceeding the current probabilities of the structure in the Quran still those who reflect upon these things keep asking for forgiveness from God, they do not forget the fire:

"Those who remember Allah standing and sitting and lying on their sides and reflect on the creation of the heavens and the earth: Our Lord! Thou hast not created this in vain! Glory be to Thee; save us then from the chastisement of the fire." 3:191

Believers in code19 however don't want to see the hell in sura 74 with it's 19 guardians. They prefer to make the honorable messenger of God a disbeliever..(Check the article Was Muhamad disbeliever? ) They even accuse others of seeing "too much fire".

* Does God want us to be statistican and mathematicans to be able to find out the probabilities and base our faith upon that ? I don't think so.

There are many problems in probability theory whose solutions totally defy our intuition. How many of believers in "code 19" are aware of the fact that is we pick 15 persons at random then there is 25.3% chance that at least 2 persons will have the birthday on the same day ? Or with 50 persons selected at random there is a 97% chance!?! Here we are not talking about 1/19 but 1/365s!

Those who limited understanding of probabilities are the only ones who will fall into this 19-hoax.

* Last but not least: I have not verified the counts of any sura/set at all and hence I find no reason to accept them anyway. Please see 17:36.