Vendor buy back ratios.. any one have Data..

[lethaldose]

Does any one have any Vendor buy back ratio Vs. the vendor paid for items used in the combination.

(buy back price / total cost) x 100 = %

250pp / 500pp x 100 = 50% <- not dada, just an example.

this would be helpfull info for my data sheet, when deciding what item to use to skill up. I like to factor in Cost, success% of 20, and buy back price.
some times its worth making a more costly item when you know your % of success is great and sell back to merchant to minimize overall cost of skilling up.

Lethaldose

Note* I read that the success ratio is (1.15 x skill)-(trivial - 70)
I use it becase it works well with trivials near our skill level (it gets way-off the greater the diffrence of the trivial and your skill becomes)
I dont think its accurate, because there is always a 5% chance of fail. thats means the equation can NEVER equal 100% or greater, and the above equation offers an OVER 100% possibility. The true equation must be an asymtope (means it will never reach a number, but it will get close)
The above equation is linear, the true equation is not linear but asymtropic.
Just my 2cp

Lethal, I'd say the equasion is accurate, only that when it reaches 100% substitue that to mean "Max", that being 95% in this game.

I haven't had much exp. myself in buyback ratios, but I know it varies greatly with the type of item you are selling.

Try selling/buying a FS weapon, then do same with a Peridot.

Someone else may have something specific on a few skills though.

[lethaldose]

You said it best

Lethal, I'd say the equasion is accurate, only that when it reaches 100% substitue that to mean "Max", that being 95% in this game.

if it was accurate there woud be no "sustitutions"

5=5 not 4.67.... this equation requires us to say 4.67 is close to 5, and it is.. but its not exact.

so the eqation is not accurate. we do not know the unknown variables VI/ Sony uses -so possibly this is the best quation we will ever have to forcast the outcomes.
All i did was point out the huge % error of this equation. The % error of this equation would cause a professional to lose there job, or get people killed.
Sorry im a Engineer/mathmatician student.. :twisted: when i see a bad equation I have a tendancy to attack it from all sides to find where it fails..

Lethaldose :twisted:

Ok...

Then how do you make a simple equasion that says everything over 100%=95%?

If you are at 200 when something trivials at 50, you'll have a 5% fail rate... thats understood.

Why do you show a problem without a solution? If you have a solution, wouldn't that be a more complex equasion than just understanding that you never have better than 95% success rate?

[Twistagain]

Actually, the true equation could be something very similar to the above as below (in pseudocode):

Chance = Max(5,Min((1.15 x Skill)-(Trivial - 70),95))

So that it would calculate the formula, then select the minimum of that result and 95, then select the maximum of that result and 5. Any values outside the 5-95 range would be discarded. There is nothing wrong with the formula's form. We know it to not be completely accurate based on the plotting of extensive amounts of data, but it is fairly close. But, there is plenty of precedence in math/computing to allow for formulas which are discarded once a min/max has been reached.

[Zallarenya]

Definitely an engineering student because a math student realizes that functions need not be continuous or defined on all points

In truth, the equations theorized are very approximate. It's hard to get enough data to fully describe what's happening. It's hard to write equations to describe the behavior, but fairly easy to code.

For the equation you mention, you might see something like this notation in a textbook describing basic function theory (f is successrate, T is trivial, S is skill):
[code:1]
/ -|
| min {1.15S - T + 70, 0.95] |
f(S,T) = <| |
| max {0.05, 1.15S - T +70] |
\ -|

[/code:1]

I say might see something like this, because most of the discussion of functions of this type are limited to single variables in textbooks

At any rate, don't expect to see one continuous function to describe successrates in EQ. (doh, second edit: Actually, the function described would be continuous where it is defined 0<=S<=252, just not wholly differentiable. Whether the actual success rate is continuous has yet to be determined.) An asymptotic equation just doesn't fit the observed data, and it would be harder to program for a game maker. The easiest to program is linear, and when the value is out of bounds, just chop it off at the limit.

Edit: changed wording and format of the equation for display

[Zallarenya]

Oh yes, you had a question to start off with Here's my answer:

People have collected buy/sell data for lots of things, and few generalizations can be made because the ratios differ from item to item (or groups of items).

Gems have a decent buy/sell ratio, for example, while research materials and fine steel weapons have horrible ones. Small money items like cork, spices, and bottles are 1:1 -- crab crackers, too, oddly enough at 9gp and change (I ran out of inventory room and sold my crab cracker to the nearest vendor, only to buy it back for my sell price moments later).

In the upcoming database, many of the items will have both sell and buy prices listed. While this isn't a priority and the number of items with these fields populated will be low on launch, if folks are interested in collecting and reporting the data, it can grow to be a useful resource.

Small money items like cork, spices, and bottles are 1:1

I do not think this is exactly true. I will check later and apologize if I am incorrect though.

I have noticed that it does a fair bit of rounding off when buying/selling. Thus an item 5cp or less seems to buy/sell for the same amount.

A stack of 1 cp items goes for 2 silver, 1 copper almost universally. Selling a stack back will net me 1 silver 9 copper I am almost positive.

So while it may appear to be 1:1, in quantity it doesn't hold up.

[warrenn]

I believe the only tradeskill which has a "formula" for sell price of completed items is jewelcraft. In that, the completed items sell for just about what it cost to make. The other tradeskills it seems pretty random. Like there might be a baking item that needs flour, water, and milk and it sells for 1 gold. Then something else that needs flour, water, milk and spices and it sells for 1 cp.

[Zallarenya]

You're right, delfontes - I oversimplified a bit.

I've only personally noticed this behavior with things in the price range of a water flask and up. I think a stack of cork (1cp each) is still 1:1 (and you can't buy more than a stack at at time), but couldn't tell you for sure.

Still, as you pointed out, it varies greatly I wouldn't attempt to make any sort of generalization, but rather take verified numbers and calculate cost/profit/loss on whatever you're interested in.

[Kaysha Soulsinger]

I suggest to head over to the pricing section on the forum, there has been a detailed discussion on Vendor buyback and Vendor sell prices, as well as the influences of Charisma, Faction and Race.

If we assume best Faction and Charisma to hit the best prices possible on selling and buying, the formulas are simple.

IV = Item Value (in copper pieces)
SellFactor = item-dependent Sell factor

Offer = IV / 1.05
Ask = SellFactor * IV * 1.05

Most (if not all) gems and precious bars of metal have a Sell Factor of 1. Items like Fine Steel weapons and Research items (Runes, Pages, etc ...) have a Sell Factor of about 9. It's going to be a hell of a lot of work to manually figure out all the IV's and Sell Factors, while this data should all be readily found in the item data. I believe the IV's are, in any case. I have yet to see if the same applies to Sell Factors, or not.