adenine C5H5N5
uracil C4H4N2O2
cytosine C4H5N3O
guanine C5H5N5O
Ribose C5H10O5
Phosphate: PO4
thermodynamic favoratism of nitrogenous bases and ribomes
SpringerLink - Journal Article
self replicating RNA with provolution
Self-Sustained Replication of an RNA Enzyme -- Lincoln and Joyce 323 (5918): 1229 -- Science
synthesis of full nucleotides via natural abiogenic methods:
ScienceDirect - Advances in Space Research : Abiogenic synthesis of nucleotides in conditions of space flight of the biosputnik “BION-11”
So here we have all the building blocks of life forming via abiogenic means all the way up to basic nucleotides (a nucleotide consists of a nitrogenous base, a ribome, and a phosphate), and a research paper I can't read fully that describes a sequence of events that would create long polymerase chains. The only thing we're missing is a minimum length and number of self-replicating chains within that length.
A length I arbitrarily found was 114:
Nucleotide sequence of microvariant RNA: another small replicating molecule.
Only problem is I can find nothing saying how many combinations within that are self-replicating, it could be anywhere from 'all of them' to 'none' but more than likely in between. If we start plugging in random value we get some figures:
4^114 = 1 in 4*10^68 chance a single reaction will produce a self-replicating strand
2^114 = 1 in 2*10^34 chance a single reaction will produce a self-replicating strand
1^114 = 1 in 1 chance a single reaction will produce a self-replicating strand
Now let's have some math fun, let's arbitrarily assume that the number of reactions taking place at any moment on a prebiotic earth is equal to the number of grains of sand upon the earth (10^24) and lets assume that a reaction takes place every minute, we get:
(10^68 / 10^24) / (364 * 24 * 60) = 7.6*10^38 years
(2*10^34 / 10^24) / (364 * 24 * 60) = 38156 years
(1 / 10^24) / (364 * 24 * 60) = 10^-30 years (let's just assume 1 second, shall we?)
So the top and bottom ones seem very unlikely, and the middle one seems really low. We can't put a date on the start of the RNA World Hypothesis if it's correct because our only figures (4 billion years ago for life) go back to when early cells existed and RNA World does not rely strictly upon cells- at best ligand globs as stated by an earlier poster or simply free floating. Let's not let reality stop us though! Now we get into pointless speculation.

(moreso than this post already is.

Seriously, I'm all but pulling figures out of my butt here.) Let's assume 4 billion years ago RNA World began, that gives us 400 million years of chemical processes to produce it. Reversing the equation:
Chance = Time(total) * Time(reaction) * Number(reactions per unit time)
So plug it in for 400 million (age of earth - 4 billion give or take a bit)
Chance = Time(total) * Time(reaction) * Number(reactions per unit time)
400000000 * (364 * 24 * 60) * 10^24 = a 1 in 2*10^38 chance per reaction
So there you go, the chances for life to arise on a given planetary body fitting the requisites for life are very, very good even if the per-reaction chance is crazy mofo with a nuke low.
N.B. Most of this is whipped out of the aether of my swiss cheese, take it worth the value of a fart, or the value of my dog's fart.