Blaise starts another chaining thread with a question...

[Blaise]

Greetings,

This week I broke three chains in the space of a minute in Jade Noids.

I pondered, once the cursing stopped, what Cosmic Forces might be working here. My chains don't break often... if chain breaks are a mostly random effect, it feels like I should see single breaks more often. Twice now I've had multiple breaks in the space of a minute.

Anything that causes chains to break, other than random chance?

DETAILS: I have 50 Loovma. The breaks happened in the top third of the tunnel. There were no webs visible, but I was running by several Jade Noids. I have more good Karma than bad. I don't remember saying anything bad about DT in a public venue....

[Himitsu]

Luck is a harsh mistress.

Sometimes a fighter with 300 Atkus whiffs a rat.

Sometimes a giant vermine whiffs a mystic who's on red as she's running towards town.

Sometimes a person breaks three chains in a row.

While Loovma reduces the chance of chains breaking and increases the chance of getting metal back when the chain does break, the odds of the former isn't close to 0% and the odds of the latter isn't close to 100%.

So it's all a matter of luck. This probably won't make you feel much better but I don't think the universe is conspiring against you.

[Lorikeet]

I agree it's a matter of luck. I have had chains lasting almost 200 pulls, and ones that broke on 0. The only thing I've noticed is that if one breaks and you try the next one too fast, there seems to be a high probability the second will break as well. I try to wait a short pause before chaining with a second chain after one has broken. You said you broke three in the space of one minute, and that could account for it.

[Brune]

Lorikeet wrote:

The only thing I've noticed is that if one breaks and you try the next one too fast, there seems to be a high probability the second will break as well.

Ya know, it's funny but that seems to hold true of several things in CL.

I notice this particularly in combat. Things seem to be "streaky," i.e. if you swing at something you hit rarely (say, 1 in 10), once you DO hit, it SEEMS like your chance of hitting again right away is much higher. My fighter cousin actually tries to take advantage of this in the tests.

Maybe it's a misperception, but then again, I don't know how CL works under the hood.

[Luminary]

(OOC & math geek mode on)

While much of this sort of 'repeated failure' may be attributed to random chance, or a mysterious game mechanism, some of it may be the result of the way that random numbers are determined by the Clan Lord system.

I am pretty sure that CL uses a random number generator that is seeded from the system clock. While this is a good method for selecting a single random number, using the built in generator will cause random numbers to 'group' over time. You may have heard that computer generated random numbers aren't really random...this is the main reason.

As an example, when a character performs an action that requires a random determination, and then performs the same action a short time later, the server must create two random numbers.

If the server is using the built-in randomization functions, it is (indirectly) using the system clock for a seed. Since the characters actions are separated by only a short time, the two seeds will be fairly close in value. The computer then performs its randomization calculations, based on the two similar seeds. Most of the time, it will result in sufficiently different numbers that they appear random. However, a disproportionate amount of the time, the two 'random' results are fairly close in value.

There have been several books, and a number of theses and dissertations written on creating more random numbers with a computer, but all of them require the use of more calculating power or computing code...

(OOC & math geek mode off)

[Brune]

Heck, and here i thought things were just "streaky."

;-)))

[Hidden]

(OOC & programming geek mode on)
Originally posted by Luminary
As an example, when a character performs an action that requires a random determination, and then performs the same action a short time later, the server must create two random numbers.

If the server is using the built-in randomization functions, it is (indirectly) using the system clock for a seed. Since the characters actions are separated by only a short time, the two seeds will be fairly close in value.
Only if the system random number generator is really really poorly programmed. I can't speak for any other system, but at least Classic mac apps (and by extension, OSX apps) use a single seed specified (usually by system time, as you say) at app launch, and then generate a new seed every time based on the random number given.

[Steady Foot]

Blaise, do you need more coins to make up new chains?[:)]

[Taryn]

Originally posted by Brune
Lorikeet wrote:

The only thing I've noticed is that if one breaks and you try the next one too fast, there seems to be a high probability the second will break as well.


Ya know, it's funny but that seems to hold true of several things in CL.

I notice this particularly in combat. Things seem to be "streaky," i.e. if you swing at something you hit rarely (say, 1 in 10), once you DO hit, it SEEMS like your chance of hitting again right away is much higher.

Lorikeet's statement seems to be true even if the next chain attempt is by a different person. It's an awful sight when several people are trying to chain and they all fail.

And I've experienced the same streakiness in my studies of weapons (swinging thousands of times at the monastery dummy). I have to use a higher sample than I ought for the accuracy I'm trying to achieve.

But maybe this is intentional, just to frustrate people trying to figure out the numbers by statistical studies. Suppose the random number generator was actually the sum of two digits of a clock.

Taryn.

[Blaise]

Originally posted by Steady Foot
Blaise, do you need more coins to make up new chains?[:)]

My practice has been to keep my "last" chain repaired by the Clan, and maintain the others myself. I had enough money on hand to repair two of them right away, and I'm close to having enough for the third after a generous tip from Gunslinger (so a stroll with one of our Fighters will top that off [;)] ). I think I will hit Kiri or Mehan up for funds for the final chain (which I had broken the day before in the same place).

Lost a piece of metal because my pack exploded with the three-in-one failure... but I had a spare in storage. :cool:

Speaking of generosity... Kiri's been helping me get a Catsbane Necklace together... and Lex was kind enough to supply the last part! [:D] Now I can go and fall in a REALLY bad place! Thanks Ladies....

Blaise

[Farhope]

I think (but I could be wrong) that you have more chance to break a chain
- if you are running while attempting to chain
- if you are not very very close of your "target"

I like to think that the more Loovma training you have these 2 things are easier to succeed...

[Lorikeet]

Originally posted by Farhope
I think (but I could be wrong) that you have more chance to break a chain
- if you are running while attempting to chain
- if you are not very very close of your "target"

I like to think that the more Loovma training you have these 2 things are easier to succeed...

I don't know how much Loovma you think is required, but I haven't noticed any difference with 30. I use the same chain macro and I am often running while chaining. When it breaks I'm sometimes stopped.

Lorikeet (who figures it's just a matter of luck)[;)]