Okay, I know how to generate tables. I know the purpose...and I don't want URL's to any sites on the web, because I've read them all, seen the PDF's, downloaded the tutorials...and I'm still f***ing clueless.

I just want an explanation of the variables in the table names...and what they do.
Example: What is the "index", and how does changing it do anything? How do I know how many tables to generate to complete a full set.

Barring all the garbage I get about "that's impossible", or "that takes too long", here's what I am doing:

Using Winrtgen v1.3 from http://www.oxid.it I am generating tables for LM hashes.
I am doing it for a custom charset:
ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789!@#$%^&*()

46 characters. I don't know what to set the chain length to, or what to do for index and chain count. Those make no sense to me. By leaving everything default, and increasing the tables...I slowly but surely get a better success percentage. I don't know why more tables increases this...mainly because I'm still not sure what a chain is, and what is stored in a single .rt file. I apparently have to generate 100,000,000,000,000 tables to get a 100% success rate according to this.

It won't tell me how large it is for anything more than 10,000,000,000 tables (1665497181 e86 at 610.35MB/table)

I have it generating 175 tables right now for a grand total of ~114GB. Each table takes less than two days on my AMD Sempron 2600 @ 1835Mhz.

So what I want to know:
What is in one chain?
Does it matter how long the chain is?
How many chains should I have in a table?
What is stored in one .rt file?
Why do more tables mean better success rate?
If I'm only doing 175/100,000,000,000,000 tables, does that mean I only have 0.00000000000175% of the full table?

A_T