Body: A problem has persisted within the FFXI community, this is the overwhelming dependence upon the use of parsers as a finite solution to stats and hidden effects on items. Speaking as a person whose experience with phenotyping, graham staining etc, those of the microscopic level. One must really heavily upon his/her formal training as a scientist; and not forget the application of the scientific method for proving or disproving a claim.
Lets start out by first defining what a parser is and what a parser is most defiantly not. A parser is a program that searches for user defined criterion, then compiles statistics of this criterion and displays a value. The criteria (also known as "Tokens") are interpreted as best a computer can, in numerical values. Parsers are not only useful for averaging numbers they also are good on determining things like word count in a word processor, compiling shipping or spam addresses (Brogame, in all likelihood uses a text parser to and parsers FFXIAH in order to get a list of names to bot.) The main problem with parsers is it can't determine things which it can't read, (Ergo an output) and therefore somethings which we determine as pure fact are questionable because of this inherent flaw. Also in order to properly interpret the data presented a degree of personal observation must also preside. This however has the tendency to be corruptible and develop a bias. In order to properly ascertain that we have the right information and we are determining the right values we have to "observe" Initial, and Final values. However most ppl will run there parser set it to look for some value, read the final and incorrectly determine the outcome because they forgot the critical Initial values, because one has to ask how can we determining a correct and acutely accurate Average, when we so aptly dispose of our Initial values. Also in order to correctly ascertain these values you have to have a determined baseline, something you know will not influence your variables. So this includes wearing all the same same body gear if testing a weapon for instance, and using a weapon which has no known influence on values tied to what we are trying to ascertain. To clarify i have included this example. Johnny wishes to ascertain whether the Elemental staffs he has been told increase correlated magic damage by 15%. in order to most accurately determine this johnny needs to make sure that all gear other than his staff remains constant, to better improve the net result the equipment should all be a baseline source and have no determine influence on Magic dmg (i.e no int/mnd) removing all elemental bonuses and preferably all stats. For this test johnny uses a NQ Robe set even though he is 75. He unfortunately has merits in ice and lightning potency, so has determined these are not eligible for the tests. he also is unsure about mnds influence on spell dmg and has determined that the water staff might corrupt the final results of the test. Johnny also needs to select a consistent target something that doesn't change and can be used for certainty for this test. He has chosen the Tunnel Worms outside of bastok, since they have little to no resistance, and have a constant lvl of -2. He also has determined there bas hp is ~20, and def is probably relatively 10 pts at most. To determine that there is no cap on dmg to low mobs he nukes one w/ tier 4 thunder, dealing 900 pts of dmg, he nukes 4 more to increase the sample size and returns 900 values on each of them. He has chosen an acceptable target to parse, since the results on them is consistent the acc of his results will also be consistent. (This is known as isolating variables). Johnny needs a base instrument to compare against he chooses an elm pole +1 since it has the same base dmg as his NQ staffs but has no other stats. He equips the pole and nukes hitting for 400 for a total of 20 targets the avg is 400 dmg the sample size is 20 targets the sample target is tunnel worms, he then records the date, time weather effects, moon phase and weekday. This is the proper testing method. next he nukes again w/ the staff equipped his nukes do 460 dmg for a sample of 20 and records time and day etc. he can determine atm that in fact there is a 15% increase to dmg, however johnny is a man of science and knows that the other variables must be accounted for. Since there is no weather effect, he can eliminate that as a source of error. And moves on to the next possibility moon, so he comes back and retests on a full moon (for arguments sake the prior was a new moon) and notices after another 20 tests on each staff his dmg percent was up 5% accross the board. and this progresses down the line. Its boring but this is how real data is collected in real life, slow and tedious, and no true scientist would be so foolhearted as to use but one test method and present his results as imperical and undeniable proof to the fact of something.