Magical Drop 3 Combo FAQ v0.3
PureDragon

Comments, questions (that are not already answered in here and pertain to this
FAQ!) Neo-Geo strategy game FAQ requests - puredragon@koreamail.com. Otherwise,
there are message boards at GameFAQs.com and Neo-Geo.com. Use them.


NOTE: If you're not all that familiar with the game and what goes on, I
recommend you play the game first. I'm going to assume you're somewhat familiar
with the fundamentals of the game (how to move the clown, that you can pick up
infinite of the same type, etc etc). There's also an excellent FAQ on
gamefaqs.com detailing the basics of the game, written by anders0n, which
should be enough to get you started and beyond.

Throughout the document, there will be small examples to illustrate points. Use
the following key to understand them: (and use a monospace font to view this,
naturally)

r, g, y, b = red, green, yellow, blue regular balls/bubbles/whatever you want
to call them
R, G, Y, B = corresponding color that clears all matching color balls from the
board
C = clown (you)
c = the clock that appears in place of whatever was just cleared
i = ice (plain)
ir, ig, iy, ib = ice (specific color)
x = bubble
o = gem (clears all of the same color that was cleared next to it)
<, ^, >, v = directional arrow balls
1 - 9 = numbered balls
O = pointed gem (clears the screen one color at a time when cleared)

I. Preface

I thought about writing this simply because everyone else I knew who played the
game seemed not to be able to put together any combos on their own.
(incidentally, I've been playing it for at least three years now) It's fun
occasionally to completely drown somebody in versus, but it would be much more
fun to have competition. To that extent, I'll help develop competition if it
needs be.

GRIPE: Magical Drop 2 had a quasi-tutorial mode. Why not 3?

This document is copyright 2002, 2003 PureDragon. Current entities with
permission to publicly host/distribute the document are PureDragon,
GameFAQs.com, fragazine.com, and Neo-Geo.com. If you would like to host it,
contact me at the e-mail address above. Otherwise, reproduction and
distribution of this document is prohibited for any usage other than personal,
and is expressly prohibited in conjunction with any manner of sale or other
profit-oriented activity.

II. Basic game mechanic notes

Some thoughts:

- The numbered balls can obviously only be cleared by bringing the number down
to zero. If you execute a combo while one is on the screen, the number
subtracted is one less than your total combo "length."

- Incidentally, if the count for a numbered ball drops below zero, that number
is used to determine how far, in all four cardinal directions, the numbered
ball will wipe out others as well. (Consider it a space-limited <^>v in one)

- To be resolved: in some cases, I've seen numbered balls actually make 1s
right next to itself, in which case these new ones are then the next reaction.
Does anyone know how to trigger it?

- If you read the secrets FAQ at gamefaqs (written by NFenwylde), the ice ball
trick seems to some like it could work wonders. Although it is an incredibly
useful trick to have up your sleeve, it's not necessary. And sometimes more
trouble than it's worth. (Not that you can't win just by doing that, though.
:p) And I don't think I need to mention that doing that against a human
opponent could get you in trouble.

- All balls except x, c, and 1-9 can be picked up by the clown. What's
especially significant is c; that means that once you clear a chain, you're not
allowed to pick up any of the balls underneath it.

III. Combo background

There are two types of combos you could aim for. One (which I'll call
"pre-made") involves elaborate setting up of almost-complete chains, clearing
one chain, and allowing the rest to complete themselves naturally. The other
(which I'll call "reacting") may involve a degree of set up as well, but
primarily consists of clearing chains while another is being cleared. Note that
they are not mutually exclusive; you're welcome to try one or both tactics any
time during or starting a combo, and you're likely to mix the two anyway.

Pre-made combos are the staple of other puzzle games such as all the Puyo Puyo
variants. You could, in fact, probably take a FAQ for that, study it, and come
into Magical Drop 3 ready to set that up. However, "reacting" not only makes
Magical Drop unique in that regard, but makes it insanely fast for a thinking
game. (Don't believe me? Talk to anyone who's tried to survive Black Pierrot.)

People have their preferred styles; find what works best for you. Typical
advice, but remember that what works for one person does not necessarily have
anything to do with someone else. (NOTE: I know only a few people will think
this, but for those few - the styles I suggest are not exclusive of each other,
and aren't the only means of classifying tactics.)

For this FAQ, I'm going to focus on "reacting", since if you were to remove
that, the Magical Drop engine would only be a less amusing Puyo Puyo clone.

IV. The technical side

So what makes a combo? The simple in-game demo just shows doing a three combo,
but it's obvious that the more you understand about how the engine treats
combos, the better off you'll be.

When a chain is cleared, the following progression of steps occurs before the
game moves on: (referred to as the "combo window" from here)

1) Are there any other balls of the same color immediately next to it? If so,
clear it and check if that one had any balls of the same color, etc etc. In
slightly more technical speak, it searches for the largest contiguous area of
balls of the same color. Once this area is found, any oO is also taken in.

2) All x that are nearby silently pop.

3) Are there any i<^>v nearby? If so, perform the appropriate action, then
repeat step 1 if a directional ball was cleared.

4) Replace the balls to be cleared with c.

5) Replace all c on the board with empty space.

6) Pull up any balls over any empty space as far up as possible. If any x
happens to be in the way, those will get popped as well.

<check>

7*) If there are any appropriate 1-9, subtract the number from the combo.

<check>

Steps 1-6 happen for every chain cleared, no matter what. Step 7 only occurs
when the game determines that you didn't make a chain before the end of step 6,
and you're carrying at least a 2 combo going into it. <check> indicates when
the game engine checks to see if, in the meantime, you've made another chain.

So what does this mean? A combo is registered if you can clear another chain
before the end of step 6. Steps 3, 4, and 6 take up time to complete, which
either gives you a breather or gives you a chance to start/continue a combo.

To wit, here's my estimations for each:

Step 3: If there's norhing nearby, then you obviously gain no time advantage.
(for the sake of argument, it's only on the order of <1ms for the check) If you
clear <^>v, it delays things by roughly 1/7 of a second. If you transform i
into something, I think it's closer to 1/8 of a second.

Step 4: When the clocks show up on the screen, you have somewhere between 1/2 -
1 (last estimate: 2/3) second where the clocks just sit there, doing their
sparkly thing. Actually, it's more of a shimmer, but who's keeping track? :p

Step 5: Instantaneous, and nothing can affect this.

Step 6: After the clocks clear, the combo window does not end until every last
possible ball that needs to be "pulled" (falls?) up is in its final spot.
Typically, you're looking at about 1/9 sec to move up one space or so, though I
think that estimate is on the short side. Yes, that does mean that if you
manage to drop a ball on a whole ton of suddenly-empty space, you'll get a nice
fat time extension.

*Step 7: Though not mentioned above, the transformation sequence from one
numbered ball to another takes up an additional 1/8 sec - the same animation as
i, incidentally.

Altogether, you have a combo window of anywhere from 2/3 to 1 + 1/3 sec between
chains. Realistically, you'll be a lot closer to the 2/3 sec most of the time.
(Simple math: if you can execute a combo that takes a minute to complete, you
should be writing this FAQ, not me.)

Since the first check happens after step 6, this suggests that there is no
penalty for earlier chains; in fact, having a chain as early as possible in the
window means you have that much more time to find and drop the next chain. But,
since it IS a check, you CAN shoot yourself in the foot and create another
chain too prematurely - if you can set up two chains before the check, then
both chains will be counted as one reaction. Usually not the effect you want.
(Also, doing two usually means you pretty much don't have time to try to go for
a third.)

If you haven't played the game much, all this really won't mean anything yet.
So go play. :p

V. Basic combo strategies

A word to the struggling newbie: it don't come easy. I'm presuming you're
looking at this because you don't seem to get it. (Or you do but wouldn't mind
making your combos longer.)

I recommend people to play through survival - easy at first. Before you can
make multiple chains, you have to be able to make chains. The concept is
simple, and its application is not exactly difficult either.

Okay, so there are only three actions you can really do in the game. Move the
clown, pick up balls, and drop balls. That's it. These are all the tools you
need to embarrass somebody.

ex1.
ry
ry
ry

C
--

Let's assume that's all you have on the field at the moment. Both colors are
already lined up and ready to be cleared. All you have to do is pick them up
and drop them, right?

So what would be the best list of actions to do this? (again, back to compsci -
we're trying to find the optimally shortest sequence of events such that we
also minimize the number of events needed inside all combo windows)

IMO - pick, drop, [move, pick, drop]: 5 total, 3 in 1 window

NOTE: 1 window means there was one combo window, meaning a combo of 2.

We don't care about the combo window until you make the first drop. As soon as
the 3 x r touches the top of the screen is when your combo window begins.
Incidentally, you shoot balls up to the top quite a bit faster than if they
fall up on their own.

So you have 2/3 sec to move, pick, and drop. Moving is instantaneous (again,
see Black Pierrot - the only time you lose moving is because of your fingers),
so we care about the times for picking up and dropping back down. Both, you'll
see, are just plain fast. But, being on the cautious side, assume you have
about 1/3 - 1/4 sec leeway for analysis on the fly. Of course, in a case like
this, you have time to figure this out before the combo starts, since it's
considerably intuitive.

One pattern in survival - easy asks you to continue this idea ten more times.

ex1a.
r
r
r
y
y
y

C
--

Possible: pick, drop, pick, drop: 4 total, 0 in 0 windows

IMO: pick, right, drop, [left, pick, drop]: 6 total, 3 in 1 window

This example is to illustrate a sticking point that novice players have to deal
with. When you clear a chain, the clocks that replace it are unmovable. That
means that you cannot pick up any balls under it. (You can put balls on *top*,
however.)

ex2.
gg
gb
bb

C
--

IMO - pick, move, drop, [move, pick, move, drop]: 7 total, 4 in 1 window

I thought about ways to manipulate the board to make it easier, but I guess
not. In this case, there's three actions you commit before the combo begins,
and four that needs to be done during the combo window.

Same analysis as above, except now that you have one more action to perform
during the window, you have slightly less time to analyze things. Mental
preparation beforehand is key until you can get better at finding them on the
fly.

ex3.
yyyr
 rr

 C
----

Possible - pick, right x 2, drop, left x 3, pick, right, drop, (very similar to
ex2) right x 2, pick, left, drop, [left, pick, right, drop]: 19 total, 4 in 1
window

IMO - pick, right, pick, right, drop, [left, pick, left, pick, left, drop]: 11
total, 6 in 1 window

This is where it becomes important to know your own abilities, even as you're
trying to develop them. Obviously, the second way (the one I use) is a lot less
work, but jumping to six actions in the same time frame starts to affect
people. I'm confident that, with practice, most people would be able to do my
way easily, but sometimes, to minimize actions in window time costs more setup
time. This may seem trivial, but there are some situations where it pays to set
up much and draw out a longer combo than salvage maybe a two out of it.

ex4.
gbg
bgb

C
---

Possible - pick, right x 2, drop, left x 2, pick, right, drop, right, pick,
left x 2, drop, right x 2, pick, left, drop, [left, pick, right, drop]: 23
total, 4 in 1 window

IMO - right, pick, right, drop, left x 2, pick, right, drop, right, pick, left
x 2, drop, right x 2, pick, left, drop, [right, pick, left x 2, drop]:  24
total, 5 in 1 window

Both suggested solutions are ugly, but for a reason: I intentionally only let
myself use three columns. This is a typical newbie mistake - for the most part,
you will have access to the entire board at any moment, so you should
definitely use it.

To demonstrate the difference one column makes, let's redo ex4. with one more
column on the side.

ex5.
gbg
bgb

C
----

IMO - pick, right x 2, pick, right, drop, left, pick, left, pick, left, drop,
[right, pick, right x 2, drop]: 17 total, 5 in window

Either way, it's still ugly, but you have a bit of breathing room, and thus
less forethought than originally planned.

ex6.
rygb
rygb
rygb

C
----

IMO - pick, drop, [right, pick, drop], [right, pick, drop], [right, pick,
drop]: 11 total, 9 in 3 windows (avg. 3/window)

Simple extension of ex1, and again, something you'll see in survival - easy.
What's important to note here is that, for the first combo window, you can
actually complete right, pick, drop, right, pick if you choose and buy yourself
some time. The general principle is: you want to execute as many commands as
you can up to the final drop of the next combo window in the current one. That
means you just stop short of making a second chain until the first window is
finished.

ex7.

ryb
ryb
yyy
ryb
ryb

C
-----

Possible 1 - pick, right x 3, drop, left, pick, right x 2, drop, left x 4,
pick, right x 2, pick, ... (point is, there's a waaay too long way for a three
combo)

Possible 2 - right, pick, drop, []: 3 total, 0 in 1 window

IMO - right, pick, drop, [right, pick], [drop]: 6 total, 3 in 2 windows (avg
1.5/window)

Possibility 1 is only to let you know you're thinking waaay too hard.

Most people would take possibility 2 off the bat. It's an easy 2-combo that's a
gimme right there. And when pressure comes, sometimes you just have to do.

The difference between possibility 2 and personal is only a matter of
philosophy. I personally think it worth it to add three more moves and a little
more timing issue to get a third chain out of it, but some value clearing the
board a little more. Either way, I'll admit mine reeks of unnecessary effort.

Suggested practice:

1) Re-do all the examples as fast as you can. Play them out in your mind. Don't
worry if it's the best choice, or if it's a solution I did/didn't write about -
just as long as you can come up with a solution and feel somewhat confident you
won't take forever (= more than five seconds) to do it.

2) Survival - easy may seem like an insult to some, but it's a good place to
get exposure without beating yourself up.

3) Be creative. Don't look for memorized patterns; try to train yourself to ask
"okay, so where can I put three of a color together?" Try to come up with
possibilities that you might see (there's no way I could ever imagine trying to
detail all of them for even a two-combo) and go for it.

3a) At the same time, don't beat yourself up if you do come to realize you're
seeing patterns and solutions. The nature of the game tends to support the
statement "chaos is organization beyond our recognition" - you'll realize
everything is patterned. It's just that if wrenches get thrown, you can't
afford to be helpless.

4) Don't be afraid to stop and analyze. You're not a master or anything, so you
still need time to train yourself to see things.

VI. Intermediate combo notes

Truth be told, this is not a very deep FAQ, and most higher-level thought
heavily relies on fundamentals and your experience with the game. If people
find this helpful and want a more detailed look at tougher combos, let me know
and I'll consider writing an intermediate FAQ. I'm not sure I consider myself
advanced - not until I can consistently break into the royals on the Japanese
skill chart. :p

That said, let me make some notes here about taking your play up a higher
level. I personally think that, if you're willing to take the time to establish
good basics, the rest will come as you play. (incidentally, you're likely to
find the game that much more fun/interesting than just watching World bounce
:p)

It's important to know the effect of all the balls in the game. Some are
automatic two-combos, which means that you have two combo windows to find and
drop a chain to continue. Or, if you like, you can actually cut in between the
windows with a chain (or six) of your own, to delay the secondary effect. And
once you let that effect happen, you have another window (while the effect is
carried out) to chain off of it as well.

In survival, an O = *six* windows, five of which can be delayed. Just be
careful to play only colors that have not already been wiped. For the most
part, unless you've thought ahead, you probably won't be able to get any red
chains off in time.

Take advantage of the clocks! Not only does their presence buy you the most
amount of time in a combo, but they can also be used as a buffer zone. In that
respect, you can actually skip ahead a step or two (or at least disregard my
suggestion about overlapping windows before).

ex8.
T=0     T=1    T=2            T=3    T=4     T=5
rrg     r g    r              c      g       c
r g  -> r g -> r           -> c   -> g    -> c
y g     c g    r              c      g       c
y       c                     g
y       c                     g   (meantime, you're off
        r                     g    finding more chains)
C         C      C (g x 3)    C      C

If you can, incorporate board maintenance into your combos. For instance,
suppose you're playing survival and you notice that the sides have been
consistently creeping downwards while you've carved out the middle. With some
setup, you should be able to chain through those sides and hopefully save
yourself some stress as well.

To extend combos, train yourself to look at the entire board. Some people only
focus in one direction, or worse yet, one 4x4 area or something. These are good
handicaps to start, but not sufficient to step up. Nearly all of your chains
will start from the bottom of at least one of the columns, so if nothing else,
continually scan the bottoms for opportunities.

Sometimes the tendency to continually go after the same color shows too. That
doesn't work nearly as well as being flexible in what color you can do next.

The most important ability is the ability to predict. That, unfortunately,
comes either from analysis, experience, or (ideally) both. As you learn to
predict behavior (again, stemming from following the simple rules of the game),
you will also unconsciously begin to plot moves ahead. Believe me, I can't look
a single move ahead in chess, so don't let that discourage you. :p

Final note: as you improve your abilities, you'll find yourself setting up for
different reasons (and probably less setup as well). Honestly, at no time
during any Magical Drop 3 play do you ever encounter any truly random patterns
(though until you learn what the patterns are, it's better you learn to analyze
the board on the fly - and it also helps when you suddenly can't recognize the
pattern). It may be impossibly *fast*, but it's never really random. Spotting
patterns in a hurry makes you much faster on your fingers.

VII. Differences between versus/survival, adventure

Versus mode requires a slightly different mindset than the other two modes, and
so I'm going to list a small section of the differences.

First, the versus window is SMALL. It's even smaller than survival - easy. You
have a total of 7 (count 'em) columns, so you don't have much room. On top of
that, there is also no wraparound at the borders, meaning you'll have to
manually cross the board and back.

Certain balls never show up in versus. Namely, Ox<^>v1-9. (Though if someone
could confirm that arrow balls never show...)


In a match against somebody, you'll definitely need to adapt your combo
strategy. If you're on the offensive, it becomes critical to consider that the
method for determining numbers of lines sent over is not solely based on the
length of your combo, but also involving the number of balls you clear. If you
can make one chain of 12, for instance, that's usually 2-3 lines right there -
at least a two or three-combo. Since "damage" is capped at eight lines at once,
balance becomes everything - you don't want to drop a massive bomb on your
'friend', clear your screen, and leave him alive. By balance, I mean how many
balls per chain, how many chains in the combo - even how many combos.
Considering that no lines get sent until your combo is over, you can opt to
play for pressure instead - constantly dump if you feel so inclined. :p

Evidence to suggest that more combos may be preferable to more balls - if
you've played versus enough, you realize that every character sends in a set
pattern most of the time. Most of the time. The higher your combo is when you
finish, the more chaos the computer will insert in the lines you send over.
Take Devil, for example. He just sends in swaths (3-wide?) of colors, so you
can usually just send everything right back at him. Fool acts in the same
manner. But how can they actually beat anybody, you ask? Consider Devil who
just dropped 8 lines after a 7 combo. All of a sudden, there's no order, a good
half of the balls are frozen to a specific color, and where do you start
chipping away at this thing, let alone combo it? More combo = more chaos for
your opponent.

On the other hand, if you were on the other end of something like that and need
to come out of it, you don't WANT to give him time to send over something. No
lines are transferred while a combo is in progress - that is to say, you won't
send OR receive any lines until the combo is finished. Therefore, it'd be in
your best interest to only send over eight but clear up as much of the board as
you can in the process.

During a combo, the quota is not updated. If your match comes down to quota,
and you have a combo, just stop when you think you've finished.

VIII. To do

- I still need to work on refining tactics and speed myself, but I don't
anticipate I'll find something unrelated to what I wrote about. Like I said, I
still need to be able to consistently break into the "royals" on the Japanese
chart. :p

- Get actual times for the combo window, though I suspect that may not be all
that important.

- Speed training tactics? That seems like overkill. Then again, I have hit
level 64 in survival - normal, so I guess that'd be pretty far. :p

- Stop neglecting reality

IX. Thanks and credits

- Data East for the software, Neo-Geo for the hardware... it LIVES, i tell you,
it LIVES

- anders0n for writing an excellent survey of the game, and NFenwylde for
translating the Japanese secrets page. I refer to both of their works as
required reading before this. :p

- Quannie G for being the first victim (ahem) :p

- LeShadowElf and AnChi for being next ;)

- CJayC and the entire gamefaqs.com community

- Keith and the rest of the Neo-Geo.com crew (for what? who knows - but they
deserve a mention nonetheless for their efforts)

- All the Magical Drop players out there - keep the game alive

X. Version History

0.1 - initial release. Wrote down everything I thought of at the moment.

0.2 - fixed some typos and general formatting mistakes, added permission for
fragazine.com (since I'm writing for them now)

0.3 (4/11/03) - more formatting issues, small pointers. Bragging rights. :p
Added permission for Neo-Geo.com, and am considering rescinding fragazine.com
future permissions due to lack of communication and other personal issues.