Game Mechanics Quiz 1 - How much do you really know?

[thepenguin]

it's the table

EDIT: makes a little more sense when I remember europeans use 1,0 instead of 1.0

[UpperKEES]

Basically what hes saying is that as the game mechanics
are now, there is such a thing as too many speed nodes.  :P

Yep, but of course you can build too much of anything when trying to set a good time, but the above data shows that certain amount of speed nodes hardly gain you any speed. When building even more of them you don't even gain speed at all (which is comparable to building more reactors when you are already collecting 32 energy).

The most efficient numbers of them at the moment
(according to KEES) are NOT these: 3, 5, 7, 8, 10 or 12,
so build 1,2,4,6,9 or 11 nodes for the most efficient use
of energy, and speed over long distances (when using a
fractal collector grid).  :D

Please note that these number only apply when you build your collectors in a square shaped form, like some people do, causing their distance always to be exactly 5 blocks. In reality most people will have a variety in spacing between their collectors and using the average is of no help in this case. When you build all your collectors in diagonals (i.e. 5 blocks to the right and 3 up, resulting in a spacing of 58 pixels) you can see in the table that your loss with 7 or 8 speed nodes isn't that bad.

The most important thing however is realizing that an additional speed node does not have to improve the speed. Look for instance at collectors spaced 5 blocks apart (= 50 pixels; the maximum horizontally and vertically). When you have 6 speed nodes it takes a packet 10 frames to reach the next collector (= 10/36 = ~0.28 sec.). When you build another speed node this trip still takes 10 frames! When you build speed node number 8 you gain 1 frame, so the trip takes 9/36 = 0.25 seconds. This means you have to build 2 speed nodes (total cost 70 energy!) to gain less than 0.03 second per node travelled! Would you do that if you knew it?

(@ KEES, I believe he doesn't understand what your trying
to communicate, because you use paragraphs that are too
long and he doesn't take the time to read them.)  ;)

Hmmm, I think I use about 3-4 sentences per paragraph, which is less than most books do. I guess I have to research these statistics as well now! ;D

it's the table

Consider the table something you only use when you like to check my findings, when you are interested in doing some research yourself or when you like to use this info while playing or creating a map.

EDIT: makes a little more sense when I remember europeans use 1,0 instead of 1.0

Good point. Yep, we do, like almost every country except for native English speakers. I would have to adjust my regional setting to make Excel display it differently....

[UpperKEES]

Question 9 has been posted. :)

[Colin]

The answer to Question 9 is:

14,706 frames, Or 6:48 + 1 half second.

I came to this conclusion with a test map that I made with a Creeper clock in addition to waiting until the wall broke then I clicked the score button to save my time, I also started the emitters on a 1 second delay and took away 1 second for accuracy. If I didn't get it right, I must not understand why this is a question that UpperKEES posted.  ;)

[UpperKEES]

14,706 frames

Incorrect.

[Nate Dog]

14,688 frames

[UpperKEES]

14,688 frames

Incorrect.

Edit: correct! This turns out to be the right answer, see here. My apologies!

[Karsten75]

(@ KEES, I believe he doesn't understand what your trying
to communicate, because you use paragraphs that are too
long and he doesn't take the time to read them.)  ;)

Frankly, I don't understand it either.  :P

[UpperKEES]

Just ask if you're interested; I'm happy to explain some more like I did here.

[mopa42]

It looks like the game only checks for wall decays about every 7 frames, and that a wall could last anywhere between 14670 and 14676 frames. If I can only give one value not a range, I'll split it in the middle and say 14673 frames.

[Blaze]

9 frames

14698 frames.
An educated guess....
If I win or am close, I will be very surprised...

[UpperKEES]

14673 frames.

Incorrect.

14698 frames.
An educated guess....
If I win or am close, I will be very surprised...

Correct! :) Your guess indeed seems very educated, so you can be surprised now! ;) One point for you.

a test map that I made with a Creeper clock

That's indeed the way to test it, see below.

It looks like the game only checks for wall decays about every 7 frames

My findings are that the decay is checked for every frame, but the creeper only updates every 8 frames (actually ¹/₈^th of the creeper updates every frame, see here), so that may have influenced your results when you also used a creeper clock.

This time I used a digital creeper clock that was able to display minutes, seconds and frames:



At the top left you see a small basin with a wall element surrounded by 7 crazonium wall elements. The 8^th neighboring cell is taken by an emitter that emits almost directly at the start of the game (in frame 3). On top of the wall element I placed a survivor so the game will stop immediately when the wall dissolved.

The row to the right of the basin indicates the minutes passed (from 1 to 10). The section below indicates the seconds passed (1-60) within that minute and the bottom section indicates the amount of frames passed (1-36) within that second.

Now let the testmap run for a couple of minutes. The clock is not as accurate as you may hope, because the creeper is only updated every 8 frames as I described above. For that reason you have to use the O = 2 x P script with a delay of 27 milliseconds (which is slightly less than 1 frame). This enables you to advance the game frame by frame. In theory that is, because it also depends on the speed of the process scheduler of your operating system. In reality about 8 frames are processed for every 5-6 times you tap the O-key, even when the script and CW processes have been set to real-time (don't try this at home kids ;)).

Let the game run until you arrive at 6:47 and pause the game. Now keep advancing by tapping 'O'. About every 5-6 taps you will see the emitters in the frame section of the creeper clock get updated (8 emitters at once), see below:



Keep counting your taps after an update of the creeper. The last update will be when the top row of the frame section is completely updated as you can see above. The last 4 emitters only have creeper in the center cell which is noticeable by the square around the emitter. Two more taps with the O-key and you'll see the "SURVIVORS DESTROYED" dialog, which is at 6:48 + 14 frames. Subtract 3 frames because of the initial delay of the emitter dissolving the wall and you get to 6:48 + 11 frames. This adds up to a total of 14699 frames (~408.3 seconds). There is a slight error margin (of 1-2 frames), because the script does not always advance exactly 1 frame per tap, so I count any answer within that margin as correct.

From this we can deduct that:
A wall element with 2 sides exposed to the creeper will decay in ¹/₂ x 14698 = 7348 frames = ~204.1 seconds.
A wall element with 3 sides exposed to the creeper will decay in ¹/₃ x 14698 = 4900 frames = ~136.1 seconds.
A wall element with 4 sides exposed to the creeper will decay in ¹/₄ x 14698 = 3675 frames = ~102.1 seconds.
A wall element with 5 sides exposed to the creeper will decay in ¹/₅ x 14698 = 2940 frames = ~81.7 seconds.
A wall element with 6 sides exposed to the creeper will decay in ¹/₆ x 14698 = 2450 frames = ~68.1 seconds.
A wall element with 7 sides exposed to the creeper will decay in ¹/₇ x 14698 = 2100 frames = ~58.3 seconds.
A wall element with 8 sides exposed to the creeper will decay in ¹/₈ x 14698 = 1838 frames = ~51.1 seconds.

I have tried the same procedure within a window of 11 frames for every single starting frame of the emitter (ranging from frame 0 to frame 10) and the results came to the same thing every time. When starting 1 frame later, the survivor would die 1 frame (1 tap) later, thus proving the check for a dissolved wall element is done every frame.

An interesting side issue:
When processing the game frame by frame you'll notice another remarkable thing: when an emitter emits, the creeper from that cell will spread to all neighboring cells 8 frames later, but no creeper is left in the center! You can clearly see this in the creeper clock for frames 2 to 8. Again 8 frames later the creeper flows back from the neighbors to the center cell. You would never see this behaviour with heat or fluids spreading, but it explains why you can see the emitters flashing during the game (I always wondered why).

Okay, that was a long story for an answer to a short question. :P All clear?

Edit: see here for a slight correction of the wall decay times and the answer to this question!

[Colin]

I think it is funny how close I actually was.  :P

The answer to Question 9 is:

14,706 frames, Or 6:48 + 1 half second.

[UpperKEES]

Yes, you were damn close. With a little more accurate creeper clock you would have won! (I briefly considered giving you the point after posting....)

[Colin]

Well, this is why you state rules for yourself before you post:

There is a slight error margin (of 1-2 frames), because the script does not always advance exactly 1 frame per tap, so I count any answer within that margin as correct.

You have to stick to your own rules.  ;)