This page and related pages need heavy revisions

Currently, this list of random synthesis recipes is very misleading. I've done some synthesis service on Saturdays and a lot of the people asking me to synth (besides the arat crystal duping) are making me put gathering knives with kite shields in hopes of a war sword. From what I've gathered after grouping info together from my synthesis friends, the chances of that particular recipe actually producing a war sword is around 0.1~0.3%. However, a war sword seems more likely to pop out when you use something more expensive than a kite shield such as a light armor of some sort. Another downright wrong aspect of these pages is how they are still displaying what stats the result will have. It has been confirmed by someone else already that synthesis grades are completely random and not dependent on the items that are put in.

I've confirmed that the result from random synthesis is determined by the resell value of the items put in (Light Melkas have 0 resell value but and produced very low quality items when synthesized together). The rest of this paragraph is speculation, but the game seems to average the resell values of the inputted items, randomized slightly in some fashion, and then assign a "quality group" based on the result to randomly choose an item from. Then it weights the random choosing towards a specific group of items (for example the sword group if you put in swords).

Each "quality group" also seems to have its own junk items. For example, if you put in three gathering knives, you will end up with the cheapest weapons (wooden sticks, wooden blades, wooden clubs, etc) and occasionally junk (large nail, paper, small green gem, etc). This is what I believe is the lowest "quality group's" melee weapon list and junk list. If you move up a "quality group" or two, and use the kite shield, gathering knife, gathering knife recipe, your weapons tend to be a bit better (bastard sword, fluted short sword, dagger, etc) and the junk items also upgrade in quality slightly (cheap leather, common fabric, etc). In one of the highest "quality groups" with the tipping suit + kite shield + kite shield recipe), your shield results can be fairly good (hetero kite shield, heater shield, possibly even beholder/composite/dragon shields at a low rate) and the junk items will be worth a little at least (spirit restoration potion [which resells for a little under 5k at npc], finest fabric).

Anyways, with this information as well as what Temjin has above, I think we really need to revamp this whole section on random synthesis recipes. The only question is, how should we go about doing this?

This is just a spork from JP wiki. Random synthesis uses a probability calculation related to specific parameters. There is no such things as a true "recipe" but specific item combinations have a specific percent yield for all outcomes. Stat change is random though.

I agree that the table's status changes column should be removed. The rest of your post I'm not exactly qualified to comment on.

Perhaps we can use the current pages of random synthesis recipes as examples of the mechanics of random synthesis after adding information to this page on how the results of random synthesis is determined? I'm not entirely sure what you mean by "Random synthesis uses a probability calculation related to specific parameters", though... that sounds pretty vague. Also, Is it against wiki policy to put a theory on how something works on a page when it's nearly impossible to determine how it really works? Especially on something as big as this. It sounds like it'd be okay considering the theory can be constantly revised as people test new things based on it.

Well I have a lot of statistical evidence gathered. Enough to determine the parameters with some accuracy. Certainly enough to be tested by others.

Parameters:

Value - NPC value of an item. Skews probability curve towards more expensive items. Ratio appears to be 4~6:1

Class - Class of item (armor, shield, weapon, etc). Skews probability curve towards items of that class.

Type - Secondary characteristic of an item (Magical, Sword, Metalic, etc). Skews probability curve towards that type of item.

Rarity - Only applies to specific items and items with rare types or items exceptionally hard to obtain (such as Dragon and Relic types or rare drops). Skews the curve towards rare or synth only items.

using a kite shield with more expensive swords, glads maybe, should increase the success rate for a war shield, but, can a cheap shield and 2 gathering knives have the chance for a war sword? i guess theres a range of "items" or its just the % success rate which changes for every item inside the "class" we are in. Also i think, mixing different kind of items, can give as a result, one of any item of those classes, but decreases the chance of each item as well, having more items to "pick" as a result