
Episode 19: Three Busy Bees
This is the first DFF which has been found by systematic computer search, and before starting to analyze it, I want to say a few words about this new approach to finding new, big dancing fly formations.
Prologue: Computer search - a new era in the DFF research!
After Arno had written the (temporarily) last article in this thread in 2020, stating that all known DFFs have been covered, I suggested at Krissz's online BD site to try to write an algorithm to systematically search new, bigger DFFs by the help of computers (at the same time checking if we have really found all possible "small" DFFs). I never programmed much myself (more precisely, I never did

The new DFFs are quite fascinating and quite different from those we have seen so far. But judge for yourselves - let's start with the investigation of the first DFF which has been found by computer search: the "Three Busy Bees"!
While analyzing this DFF's pattern, I found that there's some kind of magic involved, especially in the last ten phases! But first things first, here are the main properties of this DFF:
Name: Three Busy Bees
Alt. Name: Soaring-25
Discovered by: Nesdori and my brother (on my request), both independant of each other
First published: Aug 25, 2022 by me
Composition: 2FF, 4BF
Short/Long Loop: 25/75 phases
Direction: 1 Up per short loop
Speed: 25 frames per step
Remarkable: Slowest known DFF by far!
Here is the first published cave where the DFF can be seen in action:
youtu.be/yLvbaBSeHRI
Also, the Three Busy Bees are the main attraction in Dustin's BD 23, cave G (download on the Game Base)!

What do we see?
I think that by just watching the DFF in the above cave, it's nicely visible that the two fireflies and the bottom butterfly are just spinning around most of the time (these are the "lazy flies"). The fireflies form a vortex, i.e. they spin together inside the same 2x2 square. (Such a vortex is a recurring theme in many DFFs!) The remaining three butterflies - the busy bees - are working hard to hold the group together:
- First, they lift the firefly vortex up.
- Then, the three busy bees move down, two of them forming a vortex on which the third one can ride (similar to many other DFFs which have been covered by Arno).
- The three busy bees visit the bottom butterfly and lift it up.
- Afterwards, the bees lift themselves upwards in an irregular-looking pattern, until they have reached the firefly vortex again, starting the next loop.
One could identify the FF vortex with the bee's hive and the fourth BF with a flower, so we see the bees doing their daily work, flying to the flower for nectar and pollen, and the flower is lifted up when it's pollinated, and the hive is lifted up when the bees return home to eat and produce the honey before swarming out again

How does this formation work?
To understand how a DFF works, it's essential to know the precise rules of the BD engine w.r.t. fireflies, butterflies and empty space. Naturally, Arno has included these rules in episode 1 already, but let's refresh our memories before looking at the DFF:
Engine rules (w.r.t. FFs, BFs and empty space)
- During each frame, the cave scanner moves from square to square, starting at the top-left corner, then moving from left to right to the top-right corner, and then moving to the next row. The rows are scanned from top to bottom. So for a 5x5 size cave, the cave scanning order is as follows:
ABCDE
FGHIJ
KLMNO
PQRST
UVWXY
This order is very important for DFFs!
- When the scanner scans empty space, nothing happens (it just moves to the next square).
- When the scanner scans a firefly, the FF tries to move with the following rules:
1. If the square to the FF's left side is empty, then it moves there and changes its facing direction accordingly.
2. If that square is blocked by another fly, then the FF tries to move forward next.
3. If that square is also blocked, then the FF stays where it is and turns right (by 90 degrees).
- A butterfly moves similar to a FF but mirrored:
1. First priority: move to its right.
2. Second priority: move forward.
3. Third priority: stays where it is and turns left.
- Note that technically speaking, there are not two but eight different flies because each fly faces into a certain direction (left, right, up or down) which determines where it can move next!
- If a fly is scanned which has already been scanned before in the same frame, it's skipped (it can't move twice in one frame).
In the next section, we're going to apply these rules on an example, which will at the same time give us a first insight into the inner mechanics of our DFF.
How far do the three busy bees come on their own?
To understand the mechanics behind this DFF, let's take a look at the three busy bees first. It turns out that, although Arno has made very good arguments in episode 7 that 3 BF cannot form a stable DFF on their own, they can almost do it! Let's see:

Frame 1 shows the initial 3BF formation we want to examine. Two BFs (Yellow and Gray) form a vortex, marked by the black circle, while Green is moving around it, trying to find a way to break in.
Let's apply the engine rules to get from frame 1 to frame 2 step-by-step:
The scanner starts in the top-left square which is empty. Next comes Yellow. It wants to move to its right, and the space is free, so it does so (see the next diagram). The scanner moves on to the third square, again scanning Yellow. But as that BF has moved already, the scanner skips it and moves on to Green. It also wants to move to its right, but there's a problem - the yellow BF has just occupied that square! So the green BF is blocked off (marked by the red dot), and it tries to move forward instead - that square is free, alright (next diagram). The scanner moves on to the second row, and after two empty squares, it's Gray's turn. It's free to move to its right, which leads to the last diagram. The scanner moves on to the last square of the second row, finding Green again which has already moved, so it's skipped. The remaining two rows are empty, so nothing more happens, and frame 2 is reached.
Note that the actual, visible screen jumps directly from frame 1 to frame 2, the two diagrams in the middle cannot be spotted. (I just inserted those for better understanding, but from now on, I won't make any more between-the-frames-diagrams.) Now, in frame 2, the scanner starts again in the top-left corner for the next complete cavescan.
Now that we have refreshed our memories how the BD engine works, let's follow the three BF's movements a few frames further.
If you want to test your understanding of the engine rules, it's a very good exercise to verify these movements for yourself!


A star in an empty square means that two BFs are currently trying to move into that square. The star has the colour of the "winner" (the fly which is scanned first), the other fly is blocked.
We can see that until frame 7, Yellow and Gray continue to dance together in a vortex in their 2x2 grid, while Green tries to break in several times but is always blocked off (red dots). But in frame 7, Green is scanned before Yellow, so it finally breaks in and Yellow is kicked out and has to move forward. Such a break-in into a vortex is a recurring theme in many small DFFs which Arno has analyzed; now everything depends on the question if a new vortex is built. If yes, then the formation has good odds to form a stable DFF. But as we see in the next 3 frames, the three BFs do not form a new vortex, but instead they eventually face all into the same direction and will spin on forever. So this is not a DFF.
But what if our butterflies got some assistance from another fly? Well, this idea leads us directly to the "Three Busy Bees" DFF!
(Here's a short explanation for all the following maps in this and future articles; I strongly oriented myself at the way Arno made the maps in the previous episodes:
- squares = FF, cross = BF; this reminds me of the way I drew BD caves as a child: a FF was a square with a dot in the middle, and a BF was a square with a cross inside!

- arrows = current facing directions
- colours = just for orientation, to make it easier to keep track of each individual fly.
- red sign = a fly which is blocked off, i.e. an interaction between two flies. Such interactions are very important for a DFF to work because without them, each fly would only spin around without really moving.
- star in an empty square = as explained above.
- circle = marks a vortex.)

We start our examination with the situation given in phase 1. The gray and yellow BFs are dancing in a vortex, the green one is the sneaky single trying to break in. The fourth BF and the two FFs (who form another vortex) are left in white because they won't interact much with the busy bees (although the few interactions are essential for the DFF as a whole!).
In phase 1, Green is blocked by Yellow, so it tries to move forward, but is again blocked by one of the FFs. So it stays where it is and turns to its left.All other flies move into their favourite directions.
In phase 2, Green now moves into its new favourite direction and blocks the other FF off, lifting it upwards.
There're no interactions in phase 3, but in phase 4, the top FF pulls the other one behind it by blocking it off and forcing it upwards too.Green makes another attempt to break into the vortex but is again blocked by Yellow.
Actually, the FFs are done for the whole short loop - they will just keep dancing in their vortex without any further interactions for the following 21 frames! Truly lazy flies, doing no more than absolutely necessary

Now, in phase 4, the three busy bees have reached the same positions as in our pre-examinations above. So in the following phases (4-13)...

...the busy bees will behave exactly as in frames 1-10 of the above "unstable 3BF scheme"! The white flies don't influence the busy bees during these phases. However, in phase 12, Yellow influences the white BF, lifting it up. White can actually move up although Green is occupying this square, because Green is scanned before White and moves away in time (that's marked by the small green star, the last symbol I'm going to use regularly for the maps! I just see that I forgot to draw another green star into Yellow's square because Yellow can move into White's square for the same reason - oopsey!)
In phase 13, there are no interactions; each fly moves into its favourite direction. Again, White does not block Yellow off because White is scanned first and moves away so Yellow can move there (this time, the green star is there again!)

During the next two phases...

...we can see how the white BF helps the three busy bees to get into action again, instead of spinning around forever. In phase 14, White blocks Yellow off, so Yellow cannot move right as Green and Gray do. However, Yellow also cannot move forward because Green blocks that square. So Yellow cannot move at all in this frame; it stays where it is and turns to its left.The spinning pattern of the busy bees is thereby broken!
Phase 15 is also very interesting. This time, Green is blocked twice by Yellow and by Gray, so Green doesn't move. Yellow, on the other hand, wants to move into Green's position. As Green is scanned before Yellow, Yellow was hoping for Green to move away in time. But as Green can't move, Yellow is blocked and has to move forward. Here, we have a rare case of mutual influence between two flies in the same frame - this is so rare because it requires a third fly (in this case, it's Gray who blocks Green so that Green has nowhere to go to make space for Yellow.)
So that's how the white BF prevents the three busy bees from spinning around forever. With this, the white BF's job is done, and it will just spin around for the rest of the short loop!
And the busy bees? Well, despite the help of the white BF, they still don't form a new vortex (yet), and in 90% of the cases, this would mean that the formation makes some irregular movements for some frames and then still ends up just spinning around. These bees, however, are special: during the remaining ten phases of the short loop, they manage to lift themselves up, eventually create a fresh new vortex and bring themselves into position next to the FF vortex again! Let's see how that works:

If I would try to explain what happens during these phases (other than just saying "well, that's what we get by applying the BD engine rules"), then most striking to me is the behaviour of Gray and Green during phases 16-21, where they perform a full "BnB" short loop (see episode 3) without being disturbed by Yellow. During this time, Yellow is kind of "freeriding" on the BnB and lifted up by no less than three steps in a row (phases 17-20, Yellow is always influenced by Gray!) Afterwards, in phase 21, Yellow turns back to its group mates and stops the BnB pattern by blocking Green off. However, if Yellow had hoped to finally start a new dance with one of the two, it finds itself disappointed: after some re-orientation in phase 22-23, Gray and Green start a vortex dance in phase 24-25. Yellow then rides on that new vortex toward the fireflies to say hello again after a long time!
If you compare phase 1' with 1, the whole formation has moved one step up, and the three busy bees have switched places. It'll take two more short loops until each bee is in its original position again, hence the long loop is 75 frames.
The last ten phases strike me as a miracle. How the three busy bees manage to move back into their original positions, without any help from outside and without really a regular pattern, is something we haven't seen in any previous DFF before!
The map of interactions
I think it's fair to say that the three busy bees are really the main protagonists of this DFF. There are 26 interactions per short loop, 21 of which occur only inside the three busy bees group. This can be seen nicely in the following map, where all the interactions (=red signs) of the short loop are listed:

Here we see all the six DFF members with all their interactions. An arrow from A to B means that A influences (blocks) B; the numbers are the phases when the interactions happen. The three busy bees are surrounded by a circle because they switch places after a short loop, and they do so in counter-clockwise direction (i.e. Green takes Yellow's place, and so on.)
We can see that the three busy bees almost move on their own; only twice do they need assistance: once when Green is blocked by the FF to "correct" its relative position to the Yellow/Gray vortex (phase 1), and once when the white BF prevents the three from spinning forever (phase 14). On the other hand, the three white flies must also be influenced once to move up with the group, which is done with 3 more interactions. The three white flies are really minimalistic - they don't work any more than absolutely necessary to call themselves members of a DFF. Each of them is influenced only once per short loop (which is necessary so they can move with the group) and each of them also influences another fly only once (which is also necessary to be part of a unique DFF, otherwise they'd only be free riders)! Together with the fact that this is by far the slowest known DFF (until now?), we can conclude that these white flies are by far the laziest DFF flies we've ever seen!

The above map shows one more important thing, namely that this is really a unique DFF. It's possible to get from each fly to each other fly by just following arrows, which means that each fly influences each other at least indirectly over a short loop. In the next episode, we wll see a map of influences for a free-riding formation, then we can discuss the differences.
The map can visually slightly change if we start our short loop with a different phase. For example, this is the map of influences for a short loop from phases 16 to 15':

For example, here we have mutual interactions between Gray and Yellow while on the previous map, there was only an arrow from Gray to Yellow, not vice versa. Still, the main information is the same - namely the fact that this is a unique DFF, as well as how busy or lazy each fly is. (Note that as the busy bees switch places after a short loop, they will eventually all have the exact same amount of interactions over a long loop!)
By the way, I chose to start this short loop in phase 16 because in this short loop, Yellow is always the sneaky single, which can be nicely seen by the fact that Yellow "loses" most of its interactions on the map (at least until the very end of this short loop, where Yellow finally breaks in and pushes Gray away in phases 10', 11', 14' and 15', just as Green did with Yellow in the original phases 10, 11, 14 and 15.)
A dancing story
Arno used to explain DFFs with dramatic and intriguing love and dance stories, and I know I've neglected this aspect so far in this article, but postponed is not abandoned, so let's see what we have here!

Phase 1-2:
Yellow/Gray: "Whooohoooo let's dance in a vortex!"
FF vortex: "Yeeehah that's really fun!"
Green (the sneaky single): "Hey, I want to dance too! That's not fair! OK, if you BFs don't let me in, then at least I'm going to drift you fireflies apart! Ha!"
Phase 3:
FFs: "Heeee Green, what's the big idea? You're just jealous! Go find your own dancing partner!"
Green: "Ah, I can't dance with you anyway. I just wanted to annoy you!"
Yellow/Gray: "Green, that wasn't nice from you! As a punishment, we still won't let you dance with us for 7 more frames!"
Phase 4:
FF1: "Hey darling, come and follow me upwards, away from that stupid BF. I'm gonna help you... ouf!"
FF2: "Thank you! Now let's dance together again!"
Phase 4-10:
Green: "Hey butteys, if you don't let me dance with you, then at least I want you to go and visit our granny Whitey with me. We haven't seen her for the last 4 seconds!"
Yellow/Gray: "Good idea, let's go!"
Phase 10-11:
Green: "Ha! You've become imprudent! Now I'm breaking in! Get away, Yellow! Down with you! And down again!"
Yellow: "Heeeeeey!"
Gray: "Green, that was very rude from you. Now I can't dance with Yellow anymore, but I also don't wanna dance with you now, you bully!"
Yellow: "And don't even dare asking me for a dance, Green! I'm willing to spin around forever rather than dance with someone like you!"
Phase 12-15::
White: "Ah butteys, so you've come to visit Granny once again? That's cute! How're you doing?" Oh, not very well, I see. Why don't you dance together? What? You fell out with each other? Oh, come on! I haven't seen you for 4 seconds now, and you don't have anything better to do than arguing? Listen, I'm gonna help you. But first, you have to help old Granny upwards... ouuuuuf... thanks! And now I can help you. That's it! Now you stop arguing and travel together through the world again!"
Three busy bees: "Thanks Granny! Don't you wanna come with us?"
White: "Ah no, cuties. I'm far too old for such adventures. Go have fun, and don't forget to visit Granny again in 4 seconds!"
Phase 16-20:
Green: "Hey Gray! Let's dance! We haven't danced the BnB for a long time now!"
Gray: "Oh yeah, good idea! Wooo-hooooo!"
Yellow: "Heeey! And what about me? Gray, have you already forgotten how rude Green was to us when we danced? And now you choose her instead of me? Hey, wait!!"
Green: "Was ist Yellie saying? I can't hear it, the dance music is too loud! Yeeeeah!"
White (from afar, sighing): Ah, they're arguing again. Will they ever learn? Well, I was the same when I was young... looks like good old granny will have to calm them down again next time..."
Yellow: "Just you wait, my friends! Now I'll move ahead of you all..."
Phase 21-25:
Yellow: "...and then I move right in front of you and stop your dance! Ha!"
Green: "Heeey what's the point of that? Stop stopping me!"
Gray: "Leave it, Greeney, Yellie is just jealous! Now let's dance in a vortex!"
Green: "You're right! A vortex is even cooler than a BnB! Yesssssss!"
Yellow: "Ah no no no, why aren't you finally dancing with me again? What have I-" *bounces into someone* "Oh, ehm... hi."
FF1: "Oh no... don't tell me it's you again!"
FF2: "No wait, that butterfly looks different than the previous one!"
FF1: "Oh, you're right. Well, that doesn't make it better.
FF2: "Well, at least they help us to move upwards, to see more from the world!"
FF1: "Yeah OK, it's not all bad indeed."
T H E E N D... oh no wait, there is no end. It's just the start again!

Are there any relations to other DFFs?
Regarding the "Three Busy Bees" DFF as a whole, it is a very unique one and has little in common with any previous DFF which has been discussed so far! (However, there are definite similarities to a new 7-fly DFF which we will discuss in a future episode!)
Also there's no (p-)mirrored pair (see episode 8 for definition and examples). There are two phases where the condition for a mirrored pair is broken: in phase 15, Gray and Green compete for the same space to move. Gray wins, but with column-wise scanning, Green would win. And in phase 24, we have the same BFs involved, but the other way round; this time Green wins but Gray would have the upper hand with column-wise scanning. Also, no p-mirrored pair exists; to prove this, it is enough to mirror phases 16 and 25 (the phases after the condition-breaking ones). In both cases, the six flies will eventually spin around.
However, we can definitely find similarities to the machanics of smaller DFFs during certain phases and in certain parts of the "Three Busy Bees":
1. Similarity to a Lifted F-Vortex:
This 3-fly DFF has been covered in episode 6, but you can also see its full short loop on the following image:

During phases 2-4, the green BF and the FF vortex behave identical to a "Lifted F-Vortex" DFF. Then in phase 4, Green is influenced by Yellow so the Lifted F-Vortex pattern is broken.
2. Directly after that, the three busy bees form a vortex formation which is a vertically mirrored version of Fright-3 (episode 5):

The symmetry holds until phase 10, when the left-to-right scanning order becomes relevant for the first time (all previous fly interactions could be evaluated by only applying row-by-row scanning!). In phase 10, Yellow and Green compete for the same square to move, and Green wins because it's placed more left than Yellow. However, in phase 6 of Fright-3, Yellow is placed more left and wins, so Green has to wait one more frame before it can break into the Gray/Yellow vortex. This makes all the difference between Fright-3 (a stable DFF on its own) and our 3BFs who end up spinning around (until the "white granny" comes and helps them.)
3. As already mentioned, Green and Gray perform a full "BnB" short loop during phases 16-21. In fact, there's even more to it: from phase 15-17, the three busy bees behave exactly like Bob (Gray), Bobette (Green) and Babette (Yellow) in the free-rider formation "BnBnB" (see also episode 15)!

In phase 2' of the BnBnB formation, Babette is blocked off by another fly which isn't present here in the "Three Busy Bees". So from then on, both formations differ: while in the BnBnB, Yellow breaks into Green's and Gray's "dancing area" already in phase 5', the same BF needs a lot longer to do so in the "Three Busy Bees", first moving far up alongside Gray until phase 20 and then coming back to disturb the dancers from above.
So all in all, the "Three Busy Bees" are like a big puzzle which is set together of several small pieces from formations we have already seen in this thread. This makes me feel that even without the help of computers, we might have found this DFF sometime by just toying around with the known formations!
Also in the following episodes, we will see a lot of such "puzzle pieces". So although the DFFs get more and more complex with more and more flies involved, we don't have to "start from zero" in our understanding of their inner mechanics, and I think that alone will make it fun to analyze them!

How to launch this DFF?
If Rockford wants to launch such a big and stretched DFF in a cave with 2FFs and 4BFs inside, it's a good idea to first look for a phase where he can set as many flies as possible into place so that they don't influence each other. For the Three Busy Bees, phase 14 offers itself for that:

(By the way: yes, I think this is a very nice Rockford and yes, I know that a three-year old can draw such a human; in fact, I drew humans the same way when I was three, although I didn't know about Rockford yet...)

If you think the white BF away, then all the other flies are just spinning around without influencing each other. So Rockford can can calmly set one fly after the other into place with the right timings, and then bring the white BF to the group lastly to launch the DFF.
In the following video, we see this in practice:
youtu.be/Rxayrwwi7uI
Rockford first brings the gray BF into place (I marked the looping path of Gray and Green on the image by additional circles for better orientation), then the FF vortex follows, then Green, and finally the yellow and white BFs (which are grouped together in the cave because this turns out to work nicely). The thin strokes on the image mark Rockford's path after releasing the last two BFs (White and Yellow); he must be careful not to get crushed by Green or Gray along the way. The (dirt) square must be snapped free in time so White can move down in the following frame.
Note that in the cave in the video, the flies are already placed, grouped and timed in a way to make the whole launch as easy as possible; if you just have 2FF and 4BF randomly placed somewhere in a cave, the launch would become still a lot harder to do! Also, I think it's hard (if possible?) to create a cave where the launch of the Three Busy Bees DFF is strictly forced; for example, Nesdori solved the above cave by launching 2 independant Lifted B-vortices instead.
Free riders: the "Soaring Kite"
The Three Busy Bees can be extended with a free-riding BF. (Remember: a free rider is an extra fly which can move with a DFF without influencing it.) It is the brown BF in the following image.

This extra fly will just spin around most of the time inside the 2x2 square marked with the circle. It only interacts with the rest of the group once per short loop, namely in phase 20 when it's lifted up by Yellow. (This interaction is of course necessary for the free rider to move with the DFF.)
As you can see at the bottom half of the image, arbitrarily many more free riders can be added to the DFF! This is another analogy to the Lifted F-Vortex DFF, where the same thing is also possible. Below are two videos: the first shows the above formation in action, while the second video shows Arno's older cave on the analogous Lifted F-Vortex formation for comparison (the "Flying Kites").
youtu.be/OV_F6MOaQHY
youtu.be/OyFkcdsNVhE
Stabilizing small "almost-DFFs" - a new class of (big) DFFs!
The "Three Busy Bees" are clearly based on the idea to make a 3BF-formation work with the help of minimalistic/ "lazy" flies (the white flies on the images) who spin around most of the time, but then have one decisive moment when they interact with a "busy" fly. As we will see, similar principles can be found in other big DFFs as well. It will often be the case in the following episodes that we will see some flies of a DFF doing almost all the work while others are there for assistance. This is quite opposite to the smaller DFFs, most of which are based on vortices where every fly has to work almost equally hard to make the DFF work as a whole!
What's up next?
I think it makes sense to examine a bunch of new free-rider formations on old DFFs next. This will include the (probably) last 6-fly DFFs. Then we can conclude the topic "DFFs with up to 6 flies" with a complete overview and move on to really big DFFs with surprising and unfamiliar inner mechanics! Stay curious...!
