Jump to content


  • Content Count

  • Joined

  • Last visited

Posts posted by Ubernuber

  1. Things that you might grant achievements for:

    - Defeating a level three boss within a given length of time or by a given wave number

    - Completing the game on a given difficulty level

    - Completing the game on a given difficulty level within a given length of time

    - Completing the game with a given amount of minerals unspent (on a given difficulty level)

    - Completing the game with a given amount of lives remaining (on a given difficulty level)

    - Achieving a given number of fruit kills on a given difficulty

    - Building a given number of towers of one type during a game

    - Building a given number of towers of one type over your career

    - Being the last survivor in a given number of games

    - Being the last survivor in a given number of games on a given difficulty level

    I completely agree. The long-term motivation for many maps comes from difficult challenges for even experienced players. Ele TD benefits from being a survivor style map, ie. players of different skill levels can still play a meaningful game together (compare to eg. "Protoss Special Forces" or "Mineralz").

    Additional challenges (they have got to have cool names, too - as a non-native speaker I'm totally open to better names):

    - "Twosome": Beat the game without triple element towers

    - "You can run...": Beat the game without slowing towers

    - "Sniper": Beat the game with only single target towers

    - "Look Ma, no buffs!": Beat the game without buffing towers (forge, well, trickery)

    - Stack all buffs on one tower

    - "Miser": Beat the game without selling a single tower

    - "Interest whore": Sell 1000 towers in a single game

    - "Conservative": Beat the game using only 4/3/2 different elements

    - "Tycoon": Build 10 upgraded triples in a single game (nicer than "earn X money")

    - "World traveler": Over your career, build every tower (and every upgrade!) there is in the game

    - "And the kitchen sink": Stack all existing debuffs on one creep

    - "First!": Finish 5/10/15 consecutive waves first in a game of at least 5 players

    - "Slow and steady wins the race": Finish 5/10/15 consecutive waves last in a game of at least 5 players, but do not leak while doing so

    - "A penny spent is a penny earned": Beat the game without ever reaching more than 10000 minerals

    - "Mikado": Beat the game while moving your builder a maximum of X screen units

    - "All the sizzle and no steak": Lose by leaking exactly one creep

  2. Regardless of the formula you use, you're suggesting to give different income percentages based on difficulty. Also, regardless of the difficulty level, 4% is too high, because it corresponds to the two-interest-builds that were used before the income nerf and that regularly resulted in completely filled screens.

    I'm not a huge fan of interest in general, but the idea of harder difficulty levels is to make the game harder, so it would be counterproductive to give VH players higher interest, which makes the game WAY easier.

  3. Might be from a purely DPS point of view, but a Haste tower can do a two-pass (the first of which is mainly used for charging up), Celerity can do a four-pass. Also, once the last creep is around the lower left corner, you can stop microing, because all zones it can shoot are basically at max range. Just some small micro to ensure no creeps slip through.

  4. I also consider celerity to be one of the best (if not THE best) tower - if you micro it. If you place one in the 5 area, you can effectively make a 4+ pass at maximum damage output (above and below 4, below 2, right of 6, above 8 - depending slightly on placement). Ridiculous damage. I know no other tower that takes you so far in such small numbers.

  5. Sorry, math ahead. I'm calculating an approximate function to describe the amount of money you should have at a given point in time, and then check how changing the base interest and other factors changes this amount. This helps deciding how to change the game based on how much money players should maximally have in the endgame.

    -------------------- MATH STARTS HERE ----------------------

    If t is the expired game time, dt is the income interval, b is the base interest rate in percent and is the number of interest picks, then your money grows roughly exponentially (assuming that you have to spend as much to stay alive as you earn on each wave's creeps - a crude approximation, but it works out roughly).

    The money function is given by (omitting the scaling factor 70, starting money):

    m(t) = [1 + (b+i)/100 ]^(t/dt).

    You can transform that to a base two exponential function (doubling of money after a certain time):

    m(t) = 2^[t/dt * log_2(1+(b+i)/100) ],

    transforming the power to base e yields

    m(t) = 2^[t/dt * log(1+(b+i)/100)/log(2)].

    Since (b+i)/100 is typically small (in the range < 0.05), you can linearize with fairly good approximation:

    m(t) = 2^[t/dt * (b+i)/69.31].

    To now find the time interval in which your money doubles (with the income interval dt = 15 seconds):

    1 = t / 15 * (b+i)/69.31 <=> t = 15 * 69.31 / (b+i)

    -------------------- MATH ENDS HERE ----------------------

    (currently) number of interest picks => money doubling time => number of doublings over a complete game (~3600 sec) => endgame spare money (roughly):

    0 => 520 sec (11.6 min) => 6.9 => 9k

    1 => 347 sec (5.8 min) => 10.4 => 90k

    2 => 260 sec (4.3 min) => 13.8 => 900k

    What a good change would have to achieve IMO is about 2-3 less doublings of money over the course of a whole game if you pick two interests. Then in the endgame, you'd have 100k - 200k max, which would give you a fun amount of towers to kill fruit with, but not completely spam the screen.

    Changing the income interval to even 20 seconds would pretty much completely destroy interest builds (even though it doesn't sound like much):

    (with interval time 20 sec)

    0 => 693 sec (11.6 min) => 5.2 => 2.5k

    1 => 462 sec (7.7 min) => 7.8 => 15k

    2 => 347 sec (5.8 min) => 10.4 => 85k

    I suggest changing the income interval to 18 seconds and changing nothing else.

    (with interval time 18 sec)

    0 => 623 sec (10.4 min) => 5.8 => 4k

    1 => 416 sec (6.9 min) => 8.7 => 25k

    2 => 312 sec (5.2 min) => 11.6 => 190k

    This would leave income builds largely intact, but take the completely ridiculous edge off the endgame. Other builds would not be strongly affected.

    edit: It is similar to change the income bonus from +1% per pick to +0.5%:

    (with interval time 15 sec and bonus 0.5%)

    0 => 520 sec (11.6 min) => 6.9 => 9k (unchanged!)

    1 => 416 sec (6.9 min) => 8.7 => 25k

    3 => 347 sec (5.8 min) => 10.4 => 90k (old one-income version, total interest 3%).

    This variant nerfs you stronger if you take more income, zero-income builds are unchanged. Similar to the 20 seconds interval, this makes interest builds almost unviable for the average player, in my opinion, but it would probably also prevent the high-end pros from reaching totally absurd amounts of end game money.

  6. It can absolutely be a problem in multiplayer.

    I just tried a double income FNW build on very hard (II FNW FNW FNW, public). If you survive until wave 25 (not that hard), you can easily beat the rest of the regular game without problems (Vapor, Infrared, Celerity with some Well sprinkled in). The build is not _that_ strong against fruit since the mainly AoE damage stops being sufficient around 100 kills and the low FPS makes is impossible to micro, but I was able to fill the entire buildable area with fully upped towers and had 100k+ spare money.

    I know of no other 3-element build that does it that well, but there might well be (EFW, EFL?).

    Lowering the base income (or alternatively making the interval longer) limits the viability of most builds... basically all fruit hunting builds rely on income. It doesn't strongly affect "noob" players, since they mostly don't have spare money anyway. So changing that would reset the fruit kill highscores, but I don't think that's a critically bad thing. I'm actually not that fond of the completely filled screens that make the game unbearably slow.

  7. Update: Version 3 (release candidate)

    Added functionality:

    - Added drop-down boxes with all possible 3-, 4- and 5-element builds. Resets previous choices.

    - Tower icons and names can be clicked to increase or lower their must-have level(left and right click; equivalent to raising or lowering counter; does not necessarily change current level; works in upper as well as lower area).

    - Element icons can be clicked to increase its picks by one (within the limits of available picks and must-have tower restrictions, i.e. can't lower element picks beneath minimum requirements).

    Use cases:

    - "I absolutely want tower(s) X at level(s) Y. What other towers do I have automatically, and what can I get with the remaining picks?"

    - "What towers do I have access to with DELW? What are the possible variations?"

    The two above use cases can be mixed, but the "must-have" towers always override the rest, possibly resetting other choices made beforehand to the minimum reqs for the currently selected must-haves.

    You find attached a ZIP file containing a standalone .jar application (requiring a Java Runtime Environment to run) as well as the applet source code. You can edit the source to your heart's content, but please credit me as Ubernuber.664 (EU) if you redistribute it. Consider the terms of my sharing this software as a Creative Commons nc-sa-by. (If you just want to use the applet, you can forget this blather immediately.)



    Element TD Build Selector.zip

  8. Version 2: prettified with graphics, eradicated bugs.


    - Select the towers you absolutely must have (in the upper list)

    - fine tune via clicking the element icons (lower right)

    The lower list shows you the levels of all towers you have with these picks.

    Known problems: selecting picks via the element buttons and then changing a must-have-tower in the upper list resets element picks to those selected via the upper list.



  9. Hey all.

    I was registered at the old forum way back in FT-time. I was delighted to see one of my favorite custom maps back in action in SC2 and played some. At some point I got back into my old habit of drawing up huge lists with possible builds and the towers one gets access to with certain combinations. After a while I thought "this should be pretty easy to program into an applet" and thus was born the Tower Selector.

    I wrote the code over the last two days. It's not so pretty, but mainly functional with one or two bugs relating to resetting some settings.

    I'm posting here to get some ideas for how to make the program better (both from a usability as well as a prettyness perspective). The current usecase is "I absolutely want towers x at level 2 and y at level 2. What other towers do I get automatically?". Towers in red can't be taken at a higher level (due to element pick restrictions or because they're at maximum already).

    Please don't hesitate to tell me what would work and look better or is more sensible.



    P.S.: Once it's in a usable state, I'd like to get permission to post the runnable jar here, or alternatively to host an applet on the official site.


  • Create New...