Raised This Month: $ Target: $400
 0% 

Tower Defense


Post New Thread Reply   
 
Thread Tools Display Modes
stupok
Veteran Member
Join Date: Feb 2006
Old 12-20-2006 , 19:57   Re: Tower Defense
Reply With Quote #21



The plan doesn't suck, that's my highest score so far.

Putting towers on the ends really does have an odd effect. I don't know what else to do to get the edge. The flower is probably worthless. The clouds are completely worthless.

Last edited by stupok; 12-20-2006 at 20:05.
stupok is offline
sawce
The null pointer exception error and virtual machine bug
Join Date: Oct 2004
Old 12-20-2006 , 22:25   Re: Tower Defense
Reply With Quote #22

Quote:
Originally Posted by Zenith77 View Post
wow your plan sucks
??? Elaborate.

The only thing that sucks about going all green is its kind of hard to start with only a green and keep it in a spot that will be useful later on. I start here:


It works well later on, the only thing about starting there is you have to upgrade really fast (as in, right when the kill to earn the money is made) or some can slip through.
__________________
fyren sucks
sawce is offline
Emp`
AMX Mod X Plugin Approver
Join Date: Aug 2005
Location: Decapod 10
Old 12-20-2006 , 22:57   Re: Tower Defense
Reply With Quote #23

[img]http://img327.**************/img327/690/towerszo7.jpg[/img]
started with red circle. got 1 upgrade, then built the 1 all the way at the top. then worked both of the upgrades to the top, then started to build elsewhere.

the blue are where i double builded (apparently you can do that)

and the green is where one of mine disappeard :-/ (a seed was on the map instead)
Emp` is offline
Send a message via AIM to Emp` Send a message via MSN to Emp` Send a message via Yahoo to Emp` Send a message via Skype™ to Emp`
sawce
The null pointer exception error and virtual machine bug
Join Date: Oct 2004
Old 12-20-2006 , 23:05   Re: Tower Defense
Reply With Quote #24

Id avoid the green circle and one all the way to the right, and i dont try double stacking because of the vanish issue (even when you dont double stack, if its close it can sometimes vanish... it happened to me on wave 75 before)
__________________
fyren sucks
sawce is offline
BAILOPAN
Join Date: Jan 2004
Old 12-20-2006 , 23:06   Re: Tower Defense
Reply With Quote #25

sawce you should teach Tower Defense 101 at a university
__________________
egg
BAILOPAN is offline
stupok
Veteran Member
Join Date: Feb 2006
Old 12-20-2006 , 23:10   Re: Tower Defense
Reply With Quote #26

The red circle in Emp's snapshot is the best starting point if you're going with green. You can last with a single tower all the way to max upgrade without losing any civ's if you put it there.

New High Score:


With flowers, I only got to 75, but they weren't positioned quite right. Awful expensive upgrades, but they are cheaper than a fully upgraded nature tower.

Last edited by stupok; 12-20-2006 at 23:40.
stupok is offline
Liquidz
Senior Member
Join Date: Jan 2006
Location: Toronto, Canada
Old 12-21-2006 , 00:10   Re: Tower Defense
Reply With Quote #27

I went a different way and got me to level 70.



Meh highest i ever got.
__________________

Liquidz is offline
Send a message via MSN to Liquidz
Twilight Suzuka
bad
Join Date: Jul 2004
Location: CS lab
Old 12-21-2006 , 02:22   Re: Tower Defense
Reply With Quote #28

My highest was 85 or so.

The level one flower actually has more damage/range ratio per dollar than the green at max, but placement is difficult.

In order to get a third tier flower, one has to use as little as one unit for as long as possible (I can go up to level 50), and then place a flower, and hope to god you can get it before being overrun.
__________________
Twilight Suzuka is offline
Send a message via AIM to Twilight Suzuka Send a message via MSN to Twilight Suzuka
Twilight Suzuka
bad
Join Date: Jul 2004
Location: CS lab
Old 12-21-2006 , 02:37   Re: Tower Defense
Reply With Quote #29

In the interest of fairness, let me explain what these stats mean before hand.

Success in Tower Defense is limited to a certain upper bound (which I have not quite established), because two competing factors inevitably intersect:

Enemy health: Exponential (1.25^n)
Allied Money: Linear (60 + (x * n) )

Exponential growth inevitably overwhelms linear growth, so unless I am wrong about money growing linearly (it could grow based on the amount of damage dealt to the enemies, which is also exponential). However, signs and observations dictate that, at some point, money begins to run out and things get increasingly harder to build. This shows there is some upper limit.

Now, this means that to maximize how far we can go in Tower Defense, we have to maximize how much damage we do per dollar, as we start to quickly run out of dollars in later levels, and in earlier levels we don't get much to begin with anyway!

However, if we gauge purely by how much damage we do per dollar, we ignore the range factor, and get false readings. For the purposes of the statistics, I have assumed that range is a linear scale, which observation confirms. To elaborate, range is a multiplier that arises through this: (range / 100).

Of course, merely calculating the damage|range per dollar is not sufficient to explain HOW to maximize Tower Defense returns, but it does show what the resulting placements should look like.

On to the stats!

Code:
Lightning	Damage	Range	$   Lvl $   Dmg per $	BVR
Deployment	40	100	40    40       1	  1
Code:
Fire	        Dmg	Range	$	Lvl $	Dmg/$	BVR
Deployment	4	100	18	18	0.22	0.22
Level Two	12	100	20	38	0.32	0.32
Level Three	36	110	30	68	0.53	0.58
Level Four	108	150	50	118	0.92	1.37
Level Five	300	160	80	198	1.52	2.42
Level Six	2056	190	600	798	2.58	4.9
Level Max	7021	200	1500	2298	3.06	6.11
Code:
Water	        Dmg	Range	$	Lvl $	Dmg/$	BVR
Deployment	5	100	10	10	0.5	0.5
Level Two	15	150	20	30	0.5	0.75
Level Three	45	150	40	70	0.64	0.96
Level Four	135	180	80	150	0.9	1.62
Level Five	405	200	160	310	1.31	2.61
Level Six	1215	200	320	630	1.93	3.86
Level Seven	3342	200	730	1360	2.46	4.91
Level Eight	6684	215	1000	2360	2.83	6.09
Level Max	7335	300	1000	3360	2.18	6.55
Code:
Nature	        Dmg	Range	$	Lvl $	Dmg/$	BVR
Deployment	5	100	15	15	0.33	0.33
Level Two	10	100	20	35	0.29	0.29
Level Three	45	110	40	75	0.6	0.66
Level Four	136	150	80	155	0.88	1.32
Level Five	405	160	160	315	1.29	2.06
Level Six	1215	190	320	635	1.91	3.64
Level Seven	3645	200	640	1275	2.86	5.72
Level Max	9000	215	1100	2375	3.79	8.15
Code:
Flower	        Dmg	Range	$	Lvl $	Dmg/$	BVR
Deployment	9090	150	2000	2000	4.55	6.82
Level Two	15050	100	5000	7000	2.15	2.15
Level Three	909090	90	15000	22000	41.32	37.19
Damage, range, $ (cost) are simple to understand. The next column is total level damage, which is the accumulations of costs till that level. Next over is damage per dollar, raw, which is the raw damage per dollar ratio. The last column is with range factored in.

As you can see, excluding third tier flower, the highest damage per dollar is the first tier flower. Factoring in range, the nature max outruns it by a healthy margin however.

Third tier flower rapes all other values; this is the thing to aim for in the end game.

Conclusions:
Lightning outbeats all others for a significant number of levels, 3 across the board. It can be used as a quick stopgap measure, but fire outruns it as a stop gap measure, for a pretty small price increase.

Fire is the best deal in the middle game. It has a very low cost, under 1000, for a rather high BVR of around 5, and can be built quickly. However, its end game is the lowest of the three mains, and thus shouldn't be used except as a stop gap measure.

Notes: Use as stop gap in endgame

Water is the lowest dmg/$ of all in the max level, however with range factored in, is beat only by the green max. However, the additional range MAY be a factor, in placement.

Notes: Placement can render a higher amount of damage in endgame.

Nature has the highest BVR, which factors in range, and thus in the end game is the most economical. However, its dmg/$ is low UNTIL the end game; its second to last level is comparable to fires second to last level, despite such a higher cost. Nature is the way to go for the end game, but not for stop gap measures.

Notes: Best value, without factoring in placement. Try to put at end of hallway.

Flower has the highest in the end game (which is almost unobtainable), and is the only one to go DOWN when leveling up. The first tier is comparable to the max of nature, though at a reduced cost (which doesn't really matter, as it is insignificant). Unfortunately, the range means it sucks in the end game.

Notes: Get the third tier flower. WINRAR.

Next up: Placement notes.
__________________

Last edited by Twilight Suzuka; 12-21-2006 at 03:07.
Twilight Suzuka is offline
Send a message via AIM to Twilight Suzuka Send a message via MSN to Twilight Suzuka
Twilight Suzuka
bad
Join Date: Jul 2004
Location: CS lab
Old 12-21-2006 , 03:54   Re: Tower Defense
Reply With Quote #30

New: Algorithm for coming up with health and income.

_root.health = _root.health * 1.1 + _root.wave * 8.6;

_root.income = _root.income * 0.02 + _root.wave * 1.1;

Ok, so I was slightly wrong. Health grows exponentially, but income grows logarithmically.

Why?
n^8.6 + 25^1.1n = health
n^1.1 + 60^0.01n = income

That kind of exponential can be written as a logarithm, but I won't go into it. The point is, income quickly becomes bounded, while health is unbounded.

Thanks to KoST for this info. More to come.
__________________
Twilight Suzuka is offline
Send a message via AIM to Twilight Suzuka Send a message via MSN to Twilight Suzuka
Reply



Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT -4. The time now is 20:08.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Theme made by Freecode