Monday, 14 May 2012

Butterfly Milkers

First of all, Diablo 3 comes out tomorrow so I don't expect to be playing much Glitch for at least a few days, maybe even a few weeks.  As a result I probably won't have a lot to blog about, though I think I'll have to post something about the incredible cost of restoring objects in your street/house in the near future.

Until I get to that, though, here are some interesting things about butterfly milkers which probably apply to meat collectors as well.

Butterfly milkers tick every 24 minutes.  The timer starts when you put them down or when you collect from them.  That means that if the number of milks increased 12 minutes ago and you collect then you reset the tick time and you have to wait another 24 minutes, not another 12, before more milks go into the milker.

Filling Milkers
When collection time comes the leftmost non-full collector activates and gets one milk from each of your butterflies.  The collector can only fill to 80, however, and the next collector does not activate to pick up the extra milks.  Thus, if you 6 butterflies in your yard then after 13 activations (5 hours and 12 minutes) you will have 78 milks in your leftmost milker.  The next activation will generate only 2 milks rather than 6, filling your leftmost milker completely.  The activation after that will put six milks in the second milker from the left.  For the purposes of determining the lifespan of your butterflies, the extra butterflies are not milked - it is not the case that they are milked and the milks go to waste.

For example, if you start 11 butterflies at the same time then they will fill a milker in 8 ticks, but the last tick will only use 3 of the butterflies.  After 400 ticks 3 of your butterflies will expire leaving behind a small amount of milk (or rarely a butterfly egg) but the other 8 will still be there and it will take another 50 ticks (20 hours) to make them expire.

Because of this the marginal value of a butterfly varies depending on whether it brings you up to an even divisor of 80.  The 21st butterfly actually has virtually no impact on the rate at which you collect milks over the first 20.

There also appears to be a bug that when you have exactly two milkers they both fill up on every tick so you get twice as much milk as you should.  This is very silly but not really that abusable.  Even with a modest number of butterflies you fill up two milkers very quickly so you aren't benefitting unless you are checking them all the times.  For example, with 5 butterflies it takes only 16 ticks to fill up, so you'd need to check your milkers every six hours to benefit from the doubling effect.  With larger number of butterflies, like 20, you might need to check them every hour and a half.  Probably better to just get more milkers so you aren't wasting a bunch of milks while you are sleeping unless you are logged in all day.

Too Many Milkers
Milkers and meat collectors warn you that if you have a large number they may work more slowly.  This does not increase the time between activations - that remains at 24 minutes.  Instead it gives a chance for each activation to fail.  If an activation fails the entire activation fails, it does not determine the chance separately for each butterfly.  Thus, if you have 5 butterflies the leftmost non-full milker either gets five milks or it gets zero.  Either way it will check again after 24 minutes.

The difficult question to answer is exactly how much it slows down your progress to have high numbers of milkers.  Unfortunately because it is random we need to collect a large amount of data in order to find the answer.  With only 60 data points being collected each day this could take a while.


  1. Yeah, I'll be interested in that data. I often don't log in for long periods of time, so I've found large #s of collectors work well, but I'd be curious as to how many days is the break-even point for a given # of collectors.

    Have I missed a post definitively figuring out how the # of animals depend on each other? The relationships seem . . . complicated.


  2. I haven't done any research into how many animals you can have on a street. I'd like to provide data on ideal numbers of milkers, but I only have any data on five milkers, and it's not enough to draw conclusions (up to four doesn't seem to go slower, five is around 95%-sh).

  3. It may well slow down, but I usually use more like 10 per back yard/street. My criteria is -- are they all full @ the end of the day? (I usually only get to play once an RL day). Lots of trial and error (and a little noise in the measurement), so I'm not overly sure I'm anywhere near optimal-- sometimes I still have space in them, and sometimes they're all full, but my schedule isn't consistent enough to be sure...