Challenge #5

[Karawasa]

All challenges must be played with the 3.0 version. They must be completed under normal game conditions. Very hard difficulty only.

Compete against fellow players (on this forum) in trying to beat the game the fastest.

[Cisz]

Is this challenge 7 V2.0?

The fastest what? Player in the same game?

[Guest Ryukira]

I assume that you will be playing on race mode and try to get done as quick as possible, atleast thats how i interpret it

[Karawasa]

Must have been tired when I typed this. Edited to make sense grammatically. Not race mode, just normal mode.

[Guest ibeyuranh]

most players are new to the game when i host, there no challange, once in a while there good players. Anyway, im usually playing DOTA. but if anyone is ever up for a game,whos good, whisper me ibeyuranh. USA west

[Cisz]

In the old challenge 7 the contestants agreed upon any speed challenge to be single player only, because the best times will come from a team of players that specialise in different damage types.

So this challenge is not meant to test a single player but a team, right?

[Guest kokuril]

I think it mean "Beat the other player fastet"

It is difference From Old Chanlenge 7.0 because This new Challenge want to beat the other player fastest.

So If the player's skill is different . Some player 'skill is Expert . Some player is Rookie . If play in the same game. The Expert player will beat the game very fast

[Karawasa]

I meant that you all are "fellow players," and thus are competing against each other.

[Guest Sancdar]

If only there was a single-player mode where you could play multiple lanes at once, you could get insane times. You'd need korean micro to do it, but still.

[Guest GipFace]

To those who did not pay attention to the previous thread, the current world record stands at 28:01. (This time should be stickied on the first post) If you want to beat it, you'll need quick hands and the ability to sell and rebuild almost constantly in order to match elemental cannons (the 29:18 and 28:57 replays had significantly less micro). The build order is WFLD NWLF DWP. However, I made an argument that P (pure) could be replaced with a second N because if you're fast enough, you won't have enough money for a third pure.

The realistic minimum is around 27:30, so there isn't much more room for improvement. The theoretical minimum of 26:05 (consistent 23 second clears) is approachable if a whole team dedicated themselves to clearing levels ASAP.

The main problem in the current WR speedrun is the long clear times for levels 44-51 (L51 took 32 seconds to clear!) . Clear them quickly and you'll beat 28:01 for sure.

PS - I got my new computer, so I can finally play War3 lag-free. But until I get a gaming keyboard, I cannot attempt this again. A natural keyboard is terrible for gaming. >.<

[Guest GipFace]

I have now beaten the 28:01 time by a significant amount (over 30 seconds) using a different strategy. However, I think my first run is a little inefficient, so I will try this again later in the week and show you the new time.

I won't spoil what I did yet, but here's a hint: the old set of WFLDN WLFD WP is obsolete! Who would have thought?

[Guest GipFace]

My first run was 27:29 (which I didn't save). Then I tried it again and got 28:02?! I didn't do anything different!

Is it a timing issue? I can't imagine how THIS replay is slower than the 28:01 replay. Can someone watch this and tell me what went wrong?

The beginning has been optimized even further, and I got the Monsoon at 36 instead of 41. By all means it should be a heckuva lot faster than the 28:01 replay unless there was something I didn't do here.

http://rapidshare.com/files/44881560/El ... s.w3g.html

I believe this is the optimal strategy, so I'm very surprised that this happened.

[Guest Cataras]

Since i don't know the 27:29 replay it'S hard to tell if something went wrong.

Anyways there is something i noticed when i watched the replay and the 28:01 replay both twice! You say that you optimized the beginning but when you compare the times when you get lumber at lvl 10, 20, 30, 40, and 50 you will also notice that your new "beginning strategy" is remarkably slower:

Lumber @ Lvl. 10: ~ 14 seconds slower

Lumber @ Lvl. 20: ~ 23 seconds slower

Lumber @ Lvl. 30: ~ 35 seconds slower

Lumber @ Lvl. 40: ~ 25 seconds slower

Lumber @ Lvl. 50: ~ 8 seconds slower

Obviously you catch up with your old speed only from the 30th wave onwards which is most probably due to the Monsoon Tower at level 36.

So i don't know why the new beginning strategy is supposed to be faster because i checked the times twice. The only other thing that is different between the replays is that you don't build a nova tower at all but this is very unlikely to save you another 30 seconds....

Hope this helps...

Cataras

[Guest GipFace]

Ugh, I thought putting a super cannon in the T-spot would be faster for sure. I guess I'll go back to the good ol' advanced cannon spam

I knew I would lose some seconds on level 30-35 because I lost Nature and Nova, but this certainly proves that Freezing is slower than mass elemental cannons. Actually, maybe it's because I didn't make Acid early enough...

And I thought I could beat my time using half the micro >.>

Thanks for the help!

UPDATE: I tried going back to the mass advanced cannon start. It doesn't work because I consistently leak level 2 and 3. Something with the timing has changed from 1.20e to 1.21. Can someone verify this? I cannot view the 28:01 replay as Evercraft does not work with my No-CD loader. (Yes I own a legit copy of War3, but I can't be bothered to have my CD inside my drive all the time)

If no one can duplicate the old strategy, then we'll have to invalidate all 1.20e speedrun times and start anew, with my 28:02 being the "new" benchmark. (I just realized that the <28m run was done on my old computer which still uses 1.20e)

(Just for shits and giggles, I used an APM calculator on that replay. 335 APM, which is actually more than the 28:01 replay [276]. Ouch...)

[Guest TuN]

28 Minutes? That's amazing...I've downloaded it, gonna check it later, busy atm, and I'll probably report back when I'll check..

Cisz aren't you gonna chase that record?

[Guest Cataras]

I tried this on single player always and it's pretty hard. Now i tried it once on Bnet and it's much much easier in my opinion because the creeps are a bit slower....

So is this supposed to be played on single player or on Bnet in a game where you play alone?

[Cisz]

I'd expect the b-net times to be longer, so you will probably get poor times online.

[Guest Sancdar]

Unless you use teamwork

Maybe you and GipFace should try getting the fastest possible game together.

[Guest GipFace]

For this to happen, you'd need 6 other players (one for each element while the speedrunner uses Super Cannon as a compromise). Teamwork is only needed for levels 6-35. Once Monsoon is up, the teammates can all die and watch the remainder of the speedrun at work. According to Cataras' post, levels 36 to 60 are virtually optimized to the maximum potential.

The only question is whether this is faster than single-player.

[Guest GipFace]

The creeps move at the same speed in singleplayer and multiplayer, but the rate of fire in multiplayer is faster.

If anyone wants to try a co-op speedrun, my suggestion follows for the fastest times:

Speedrunner

1-5

19-28

36-60

Goon 1 - Speedy Undead [NLDWF]

6 - Nature Cannon

11 - Light Cannon

13 - Nature Cannon

16 - Darkness Cannon

18 - Double Death

29 - Undead

30 - Undead + Nature Cannon

Die at 31

Goon 2 - Heavy Hitting [FELWDN]

8 - Fire Cannon

12 - Lava

14 - Amped Earth

17 - Double Lava

31 - Metal

Die at 32

Goon 3 - Jack of Trades [WFLNFN]

9 - Water Cannon

15 - Water Cannon

32-35 - Solar + Storm + Ice

Die at 36

Goon 4 - Darkness Whore [D]

7 - Darkness Cannon

Die at 8

Goon 5 - Earth Whore [E]

10 - Amped Earth

Die at 11

[Guest Cataras]

I wasn't talking about coop but if the tower really fire at a faster rate - then you should always try the game on Bnet/multiplayer.

[Guest GipFace]

Dammit, multiplayer isn't faster at all. My best time in multiplayer is 30:03.

http://rapidshare.com/files/49173467/El ... s.w3g.html

So I guess the barriers are 28 minutes in singleplayer and 30 minutes in multiplayer.

[Guest Jacob]

I dont get it, beat the game the fastest on normal mode ?. Then you must be able to skip the timer, is that even possible ?.

[Cisz]

We have a new version, the old challenges are closed now.