[X] My Assistant

Heh i got a weird question. I calculated my burden from items and its exactly 19.7. Will system round it up to 20 and will i get affected by that "crap"when you got >20 burden?

No you won't. Every point of Burden above 20 lowers Action Speed by 1. That's all Burden does.

Yes, Tenboro's confirmed both. I'm too lazy to search for his posts, so I'll go off my memory here.

The chance to receive a gift = (number of hours since you last checked this monster) * (number of battles won since you last received an item from this monster)

Once the game has done a check, one hour must pass by before you can do another check for the same monster.
Quality of items is a random number between 1-1000 added to the monster's power level. If the result is less than 750, then you find a low-quality item. If the result is 1250 or higher, you find a high-quality item. If the result is in between, you find a medium-quality item.

Special items, like bindings and those phazon/kevlar items can be found with any level of monster, they are just a rare occurance. Should be similar to how special items like soul stones/bubble gum are in the loot table

I think your formula is worded wrong

Gift Chance = (number of hours since you last RECEIVED a gift from this monster) * (number of battles won since you last received an item from this monster)

So checking every hour doesn't hurt your chances IIRC

EX monster has 10 wins.
Check @ hr 1: 10%
Check again @ hr 2: 20%
Check again @ hr 3: 30%
and so on, once you get something both reset

IIRC, checking resets the number of hours in that formula to 0, so my wording should be correct.


edit:

You made my actually search it, so... here you go:

Well, the way it works right now is, when you enter the monster screen it will roll on it if it's been more than one hour since you last checked. The number of hours (not rounded) is multiplied with the number of battles the monster has won since last time it brought you a gift.

So if it's been three and a half hours since you last checked, and it's won five battles since it last gave you something, your "score" is 17.5. This score then has to be higher than a random roll from 0 to 100 - in this case, a 17.5% chance. If you do not get an item, the last-checked timer is is reset to the current time, so if you check one hour later (and it didn't get any more kills), the chance would be 5%.

Note that your score can go above 100, in which case it's a guaranteed drop.

I don't think there's a statistical advantage to checking less often, but if someone wants to model it, go right ahead. There will be a disadvantage to checking too seldom, that is, having the score go above 100.
Yes.

From this post

This post has been edited by dkplee: Aug 12 2011, 06:52

I thought tens said that it doenst reset that, otherwise checking every hour would statistically hurt your chances over time of getting an item and he said that the math makes it so checking constantly doesnt hurt it

then again it may turn into something like ratings where it works but no one knows how eventually

His example makes it pretty clear cut you know. (IMG:[invalid] style_emoticons/default/dry.gif)

I thought tens said that it doenst reset that, otherwise checking every hour would statistically hurt your chances over time of getting an item and he said that the math makes it so checking constantly doesnt hurt it

I don't think there's a statistical advantage to checking less often, but if someone wants to model it, go right ahead.

And the important part :

There will be a disadvantage to checking too seldom, that is, having the score go above 100.

Let's say you have a monster. Let's name it J (for Jane, Joe, Crossdressing Joe/Jane, V-Joe or P-Jane, whatever your personal perverted preferences are). Let's assume J wins a constant 1 battle every hour over a period of 10 hours, totaling 10 wins after 10 hours.

#1 : Optimal check. You check once after 10 hours to get a guaranteed drop with a sharp 100% chance.
#2 : Gambling. You check every hour at increasing chances until you actually get something. In the worst case situation, you'll end up rolling ten failed attempts with the first one at a 1% chance and the last one at a 10% chance. In the best case situation (pretty unlikely with the current formula), you'll gain 10 materials.

Now back to reality. There is no way to exactly tell how many wins your monsters will get after X hours in order to get optimal odds. Still, I don't think Tenboro wants us to be able to easily gain at lot of materials.

This post has been edited by Mika Kurogane: Aug 12 2011, 10:31

#1 is indeed optimal. The "expected value" for hourly checking @ Mika's scenario is .01 + .02 + .03 + .04 + .05 + .06 + .07 + .08 + .09 + .1 = 0.55 Drops/10h
You are indeed being penalized on average for checking too often.

This post has been edited by DarkDespair5: Aug 12 2011, 11:53

The "expected value" for hourly checking @ Mika's scenario is .01 + .02 + .03 + .04 + .05 + .06 + .07 + .08 + .09 + .1 = 0.55 Drops/10h

Eh, no. Modeling this isn't trivial since the probability is variable. But you can trivially calculate the odds of getting at least one item by hour X by calculating the odds of not getting one, and you should be able to model it more thoroughly by calculating a cumulative distribution function on a gamma distribution, but I honestly can't be arsed.

If you simplify it somewhat and say that the monster kills stay fixed at say 10, you can use simple binomial probability to do it. In which case, the odds of getting exactly k drops is nCk * p^k * (1-p) ^ (n - k). Or specifically:

0: 0.348
1: 0.387
2: 0.198
3: 0.057
4: 0.011
5: 0.0015
6+: <0.2% chance

So the average drops in a ten-hour period with checking once every hour would be about 0 * 0.348 + 1 * 0.387 + 2 * 0.198 + 3 * 0.057 + 4 * 0.011 + 5 * 0.0015 + insignificant ~= 1.01, compared to a fixed 1 for just checking on the 10th hour.

You did the work for us in the end as usual, so it's all good. (IMG:[invalid] style_emoticons/default/smile.gif)

Interesting. Let's try this.

2: 0.19371.

Fixed that part. Can't really blame you with all the numbers, my head exploded when I read this.

(IMG:[img824.imageshack.us] http://img824.imageshack.us/img824/7694/msp45019ggi6e2geadf7ge0.gif)

WA is pretty helpful when you're lazy. Still technically higher than 1, so you're right. (IMG:[invalid] style_emoticons/default/rolleyes.gif)

This post has been edited by Mika Kurogane: Aug 12 2011, 14:52

what

39 1 You gain 0 points of two-handed weapon proficiency

is this normal?

Rounding error; you capped in 2H?

yes, but that went too fast. I didn't expekt building up the last 0,5 prof in just some hell arena

No you won't. Every point of Burden above 20 lowers Action Speed by 1. That's all Burden does.

As far as i know if you wear cloth and have >20 your spell have a chance to fail or something....

As far as i know if you wear cloth and have >20 your spell have a chance to fail or something....

That's interference.

what

tl; dr? "Math is hard." (IMG:[invalid] style_emoticons/default/tongue.gif)

WA is pretty helpful when you're lazy. Still technically higher than 1, so you're right. (IMG:[invalid] style_emoticons/default/rolleyes.gif)

@ Tenboro: Thanks for that (IMG:[invalid] style_emoticons/default/smile.gif) I can't believe I screwed up independent probability. >.<

@All: The problem is what if your monster gets different amount of kills at different PLs or changes killrate due to patch changes, or gets 26 one day and 3 the next...mind blown.

This post has been edited by DarkDespair5: Aug 14 2011, 00:07

What is the posting bonus in exp?
And how big is it, how long does it last, etc?

What is the posting bonus in exp?
And how big is it, how long does it last, etc?

Something like :

CODE

bonus = log(post_count)/2 * 100

where post_count represents how many posts you've made on the current day. Apparently obsolete. Sorry. (IMG:[invalid] style_emoticons/default/rolleyes.gif)

Caps at 100% for 10 posts. It's a daily bonus. It's also additive with the XP Boost ability and the Adept Learner training, so it will make those other two look less impressive if you train them.

You may want to check this thread as well. (IMG:[invalid] style_emoticons/default/smile.gif)

This post has been edited by Mika Kurogane: Aug 15 2011, 21:26

Something like :

CODE

bonus = log(post_count)/2 * 100

where post_count represents how many posts you've made on the current day.

Caps at 100% for 100 posts. It's a daily bonus. It's also additive with the XP Boost ability and the Adept Learner training, so it will make those other two look less impressive if you train them.

I'm sure you can manage to scare up ten posts a day to keep your bonus maxed even without using the Spamhaus.

So... did it really get changed to only 10 posts to max the bonus?