| Time |
Nick |
Message |
| 01:49 |
|
phantombeta_ joined #minetest-mods |
| 02:38 |
|
LazyJ joined #minetest-mods |
| 06:19 |
|
seamus_ joined #minetest-mods |
| 06:35 |
|
Taoki joined #minetest-mods |
| 06:35 |
|
phantombeta_ joined #minetest-mods |
| 11:25 |
LazyJ |
ABM chance and interval - I've seen chance greater than interval and in other mods interval is greater than chance. Why and when would you want one greater than the other or equal? |
| 11:27 |
|
Black_Kot joined #minetest-mods |
| 11:33 |
Zefram_Fysh |
there's no inherent connection of that nature |
| 11:34 |
Zefram_Fysh |
there's a bit of an *inverse* correlation, because the "chance" number is an inverse probability |
| 11:34 |
LazyJ |
Which one acts more as a throttle? |
| 11:35 |
Zefram_Fysh |
if you have, say interval=10 and chance=1, this bears some resemblance to the situation where interval=1 and chance=10. in both cases, you will on average get the ABM firing once every ten seconds |
| 11:36 |
Zefram_Fysh |
the former gives you exactly one operation every 10 s, at precise 10 s intervals. the latter gives you a more variable result, with a possibility of firing each second, and no guarantee about the interval between two consecutive firings |
| 11:36 |
LazyJ |
So if I have a low interval, the ABM doesn't fire as often but when it does, if the chance is high, then the ABM is more likely to do what it's coded to do? |
| 11:37 |
Zefram_Fysh |
the product of interval and chance gives you the average interval between firings. this is the main throttle control |
| 11:38 |
Zefram_Fysh |
low interval = more frequent, high interval = less frequent. low "chance" = high likelihood, high "chance" = low likelihood |
| 11:38 |
Zefram_Fysh |
the "chance" figure is an inverse probability, and is what controls the degree of randomness |
| 11:39 |
LazyJ |
So if there is more of a node for the ABM to work on, it may be better to have the interval larger and the chance lower so as not to bog down the computer trying to affect all those nodes? |
| 11:40 |
LazyJ |
On the other hand, if only a few nodes exist for the ABM to work on, then having a shorter interval and greater chance would get things done and over with? |
| 11:40 |
Zefram_Fysh |
no, that doesn't hold. if you want to reduce the overall rate of ABM operations, you need to increase the *product* of chance and interval, which you can achieve by increasing either chance or interval |
| 11:42 |
Zefram_Fysh |
tweak the product to affect the overall rate. tweak the chance, keeping the product constant, to affect the degree of randomness |
| 11:48 |
LazyJ |
Hmm... Okay. I'll have to experiment more with this new insight. |
| 11:48 |
LazyJ |
Thanks, Zefram_Fysh ;) |
| 11:48 |
Zefram_Fysh |
yw |
| 14:00 |
|
PilzAdam joined #minetest-mods |
| 14:55 |
|
FungusAdam joined #minetest-mods |
| 15:03 |
|
phantombeta_ joined #minetest-mods |
| 15:08 |
|
Black_Kot joined #minetest-mods |
| 16:15 |
|
LazyJ joined #minetest-mods |
| 18:07 |
|
SoniEx2 joined #minetest-mods |
| 22:54 |
|
Someguy123 joined #minetest-mods |
| 22:54 |
|
ShadowNinja joined #minetest-mods |
| 22:54 |
|
psedlak joined #minetest-mods |
| 22:57 |
|
rmilan joined #minetest-mods |
| 22:57 |
|
Ritchie joined #minetest-mods |
| 22:59 |
|
BrandonReese joined #minetest-mods |
| 22:59 |
|
catninja_ joined #minetest-mods |
