Exp Formula's

[altomesa]

I was about to sit down and start figuring all the formula's that are used in Ryzom to determine the experience points you receive for your melee, magic, crafting and foraging actions but then realized that there's already a bunch of math guru's hanging around these forums so it makes more sense to just ask the professionals. So if you've figured out an Exp forumula that includes some/all variables, please post below. I'm sure I'm not the only one interested in this kind of stuff. :)

[iphdrunk]

http://www.ryzom.com/forum/showthread.php?t=16827
http://www.ryzom.com/forum/showthread.php?t=12333
http://www.ryzom.com/forum/showpost.php?p=182641&postcount=36

btw Berrule roolz

[iphdrunk]

I'm not the only one interested in this kind of stuff.

Some open problems in the world of ryzom that I am aware:

* "The craft formula": the holy grail of ryzom crafters. Given a database with all mats and its properties/stats, be able to figure the final stats of a crafted item knowing the pattern ant the number of mats and type it takes. The homin that figures this out... well...no words

* "The best overcraft formula": this one requires two assumptions, if we know the craft formula and the sucess probability given the overcraft delta (which are deterministic). The assumtions are a) knowing the random variable that models either a complete failure (no mats wasted) or a degrade in the event of a not perfect item and b) knowing the random variable that models the degraded quality knowing that it is a degrade (conditional probability)

Problem 1: compute the best overcraft delta that maximizes average (expectation) craft xp per crafted item (I'm not sure it's max delta)

Problem 2: assuming a crafted item can be sold for a price that depends linearly on the quality and the number of mats, compute the best overcraft delta that maximizes a linear combination of the average craft xp per item AND the economic benefit (thus problem 1 is a particular case of problem 2, or if xp is not important, the solution tells us how to get rich faster by overcrafting).


OMG... I'm a geek....don't mind me. I can't believe I posted this, now I'll go sit in the corner :p... btw, who would pay for the answer to problems 1 and 2?

[altomesa]

You rock Anissa. Thanks for digging that stuff up, and for making it easier for us all to embrace our own inner-geek. :)

[sidusar]

http://www.ryzom.com/forum/showthread.php?t=16827
http://www.ryzom.com/forum/showthread.php?t=12333
http://www.ryzom.com/forum/showpost.php?p=182641&postcount=36

btw Berrule roolz
Those formulas are all pretty accurate, but last I tested, not 100% for all level differences. Ryzom seems to have strange ways of rounding it's numbers. So there's always room for improvement :)

Oh, and my C values for the herbies that were boosted in the recent patch are obviously outdated.

[dc77066]

Those formulas are all pretty accurate, but last I tested, not 100% for all level differences. Ryzom seems to have strange ways of rounding it's numbers. So there's always room for improvement :)

Oh, and my C values for the herbies that were boosted in the recent patch are obviously outdated.
C value for steriod using herbies appears to be 6. (Oshido says so!)

-Raku

[sidusar]

Well, if Saint Oshido (in his straight jacket) says so, then it must be so :p

I think I took notes during Madakam hunting just after the patch came out but never got around to analysing them, let me check. *opens closet door and is buried under enormous amounts of paper*

[aylwyne]

Well, if Saint Oshido (in his straight jacket) says so, then it must be so :p

I think I took notes during Madakam hunting just after the patch came out but never got around to analysing them, let me check. *opens closet door and is buried under enormous amounts of paper*

I believe we tested a ploderos and a madakam and they were 6.0. We haven't tested all of the updated herbivores, however so right now we're just assuming they're 6.0.

[sidusar]

I believe we tested a ploderos and a madakam and they were 6.0. We haven't tested all of the updated herbivores, however so right now we're just assuming they're 6.0.
That's okay, I didn't test all of the nameds and bosses either before assuming they were respectively 8 and 25. I have enough faith in the consistency of the system to believe that if 5 nameds all have the same C value, then all nameds have the same C value, and thus all bosses have the same C value too. At least until I can prove otherwise ;)

If I recall, before the patch madakam were 1.0 and ploderos were 1.5, so if they're both 6.0 now I think we can safely assume the bolobi and shalah are too.

[zerotacg]

* "The best overcraft formula": this one requires two assumptions, if we know the craft formula and the sucess probability given the overcraft delta (which are deterministic). The assumtions are a) knowing the random variable that models either a complete failure (no mats wasted) or a degrade in the event of a not perfect item and b) knowing the random variable that models the degraded quality knowing that it is a degrade (conditional probability)

Problem 1: compute the best overcraft delta that maximizes average (expectation) craft xp per crafted item (I'm not sure it's max delta)

Problem 2: assuming a crafted item can be sold for a price that depends linearly on the quality and the number of mats, compute the best overcraft delta that maximizes a linear combination of the average craft xp per item AND the economic benefit (thus problem 1 is a particular case of problem 2, or if xp is not important, the solution tells us how to get rich faster by overcrafting).

since I digged that thread up again and the craftsystem changed I just did some calculations on that one
problem 1+2:
just craft at max ( Q = 50 above your lvl)
if the degraded items are equaly spread other the qualities or even when calculating with every degrade to be the minimum quality it comes to "higher = more XP"
in average u gain 122-146XP more per mat when crafting at max

for plain XP u should consider to just craft 50 above your lvl since the diggin would be faster for lower Q (so just dig mats 50 above the lvl you want to craft)
for money just take highes possible

[sidusar]

We do know so much more about the new crafting system then we do about the old one. We had no idea what exactly the chances for failure or for partial succes were under the old one, and now they've given us the exact numbers. :)
if the degraded items are equaly spread
That's the only thing we don't know for sure, but I'd assume it's a fair assumption.

I made some graphies about the new craft system too: Chance of succes (http://i11.photobucket.com/albums/a166/bryamis/craft.png) and expected experience (http://i11.photobucket.com/albums/a166/bryamis/craftexp.png), both dependant on the difference between the target quality of the crafted item and your level in the crafting skill. It shows the experience increases fairly linearly until the maximum at 50 above your level.

[zerotacg]

That's the only thing we don't know for sure, but I'd assume it's a fair assumption.

doesn't really matter
add a new graph that shows avarage XP when every degrade is at minimum quality (-20%) it still gets more xp / lvl_difference and since you know not every degrade is -20% you can assure that the average XP is higher then that

[sidusar]

No, it doesn't really matter. Only if you insist on getting the exact number, and that graph is off by a few permille anyway :)

[jared96]

I did all my crafting under the old system and don't have any significant experience with the new one other than just making sets for peeps when needed. One thing I do notice is that, with all HA skills at 250, I am getting more degardes than I did before the patch. Rubbarn boost seems far less than the 1/20 also. I got a pair of boosted sleeves on my 2nd attempt using all sups. I have made 18 full sets since (98 pcs) and not gotten a single boost. I do get boosts, almost one per set it seems, using exe or choice mats.

As for the xp, I found purely subjectively that I did best overall when getting 300xp per mat. IIRC, that required me to OC by 30 levels or so. Through the 150-200 period, best I can recall I would work my highest skill up to 35 above the other 5 skill trees, then work the 5 until OC dropped to 25 levels. After that I just watched the xp......if I stopped getting 24k (with cats) for a vest, I'd OC by a bit more.