Limbo [xbla] ss sub 50

2010-08-09 03:02:58 pm

hey guys

anybody have any strats or tricks to share about limbo?

currently 49:12 with 2 deaths. i think a perfect time would be close to 48:30

recording will start soon so im fishing for any help

edit: now 49:01 0 deaths; perfect might be higher than i thought, like 48:47 or something

Thread title:

ZenicReverie

2010-08-09 03:09:46 pm

Waiting hurts my soul...

Quote from zech 48:

hey guys

anybody have any strats or tricks to share about limbo?

currently 49:12 with 2 deaths. i think a perfect time would be close to 48:30

recording will start soon so im fishing for any help

Don't die and good luck.

LotsofLuck

2010-12-30 11:28:33 am

I would personally love to see this speedrun. I haven't finished the game yet (~40%), and there doesn't seem to be that many shortcuts you can do, but I think this run has been to done deathless. Good Luck!

feasel

2011-03-08 02:31:50 pm

difficult and stupidly random

Here is my current best Limbo attempt (45:55):

parts 2-4:

The route hasn't really changed since folks saw it back in January (thanks especially to Pootrain and Heidman for their help at the marathon). I spent some more time trying to find skips; there were some promising ideas, but nothing has come of it so far. I've been on a temporary hiatus from speedrunning since then, and I haven't done much with the game, but now I've got a little more time and I'm looking to finish up this project, which is long long overdue.

This attempt is not bad -- I got it a few weeks ago and I was seriously tempted to just submit it and be done with it all. There are no deaths, but I do make one big mistake (missing the jump at the large gears, which loses about 16 seconds), so I don't quite think it's submission-ready yet. Moreover, I'd like a chance to get some feedback from the public before I make a final version of the run.

So basically I'm looking for any kind of input: ideas for optimizing the route, the movements, the puzzles, etc. Even minor stuff is worth mentioning. Hopefully somebody will know some new tricks.

Timing: I've been timing from the beginning of character control to the moment the screen fades out after meeting your sister (technically character control ends slightly earlier when you walk past a certain point, but there is no clear indicator for the moment this happens, so I definitely recommend fade-out as the endpoint). This run clocks in at 45:55. I lose 16 seconds missing the gear jump and there are a few other minor mistakes, so I'd put my target time at around 45:30 minus whatever other time-savers folks are able to suggest.

Here is my current best Limbo attempt (45:55):

http://www.youtube.com/watch?v=7En8DlqgUU8&p=D988387254C9356F

parts 2-4:[spoiler]
http://www.youtube.com/watch?v=9vOlsNUEB0w&p=D988387254C9356F
http://www.youtube.com/watch?v=60KDCoDvaf8&p=D988387254C9356F
http://www.youtube.com/watch?v=LzJcZfArCek&p=D988387254C9356F
[/spoiler]

The route hasn't really changed since folks saw it back in January (thanks especially to Pootrain and Heidman for their help at the marathon).  I spent some more time trying to find skips; there were some promising ideas, but nothing has come of it so far.  I've been on a temporary hiatus from speedrunning since then, and I haven't done much with the game, but now I've got a little more time and I'm looking to finish up this project, which is long long overdue.

This attempt is not bad -- I got it a few weeks ago and I was seriously tempted to just submit it and be done with it all.  There are no deaths, but I do make one big mistake (missing the jump at the large gears, which loses about 16 seconds), so I don't quite think it's submission-ready yet.  Moreover, I'd like a chance to get some feedback from the public before I make a final version of the run.

So basically I'm looking for any kind of input: ideas for optimizing the route, the movements, the puzzles, etc.  Even minor stuff is worth mentioning.  Hopefully somebody will know some new tricks.

Timing:  I've been timing from the beginning of character control to the moment the screen fades out after meeting your sister (technically character control ends slightly earlier when you walk past a certain point, but there is no clear indicator for the moment this happens, so I definitely recommend fade-out as the endpoint).  This run clocks in at 45:55.  I lose 16 seconds missing the gear jump and there are a few other minor mistakes, so I'd put my target time at around 45:30 minus whatever other time-savers folks are able to suggest.

Greenalink

2011-03-08 03:38:42 pm

DS Dictator

I noticed when I played through Limbo that at the end, I lose control once the boy is starting to go through the certain light. See 8:09 of the feasel's Speedrun video (part 4).

On part 1 @ 8:55ish it's possible to jump over the big flaming tyre but it's very inconsistent that I only could do it 1/10 tries.

feasel

Edit history:

feasel: 2011-03-10 11:29:31 am

2011-03-10 11:05:09 am

difficult and stupidly random

Character control ends when you've walked about 25% of the way into that last big tree in the background -- I see what you mean about the shaft of light; you're right, that's pretty much the spot. For doing official timings, though, you'd want a criterion that is easy to find when you're stepping through the video frame-by-frame. For the tree or shaft-of-light thing, (a) we can't be sure exactly what frame the character control is actually ending on (we're basically just eye-balling it), and (b) there's no precise visual cue even if we knew the exact spot (the shaft-of-light has a soft edge, and the tree is solid black with no features). Whereas the fade-out is a hard cut (happens in 1 frame), so there is no ambiguity.

Tire jumping -- I can not get this to work at all. It doesn't seem like I jump nearly high enough. Are you doing a standing jump or a running jump? Do you do it on the flat part at the bottom of the hill, or on the steeper slope further up the hill?

EDIT:
I got it! And I think I actually found a variation on it that's pretty reliable. I just nailed it 5 times in a row. Of course, we'll have to see if I can nail it in the context of a full run -- often things seem easier when you're playing the same spot over and over, back to back.
I do the jump standing still, facing left. Then you actually jump up on onto the tire, turn right, and jump again and you'll be flung off it to the right. I think it helps to stand in the right spot (the ground-plane is slightly wavy, and some spots seem to work better than others). I'll post a vid when I get a chance.

Character control ends when you've walked about 25% of the way into that last big tree in the background -- I see what you mean about the shaft of light; you're right, that's pretty much the spot.  For doing official timings, though, you'd want a criterion that is easy to find when you're stepping through the video frame-by-frame.  For the tree or shaft-of-light thing, (a) we can't be sure exactly what frame the character control is actually ending on (we're basically just eye-balling it), and (b) there's no precise visual cue even if we knew the exact spot (the shaft-of-light has a soft edge, and the tree is solid black with no features).  Whereas the fade-out is a hard cut (happens in 1 frame), so there is no ambiguity.

Tire jumping -- I can not get this to work at all.  It doesn't seem like I jump nearly high enough.  Are you doing a standing jump or a running jump?  Do you do it on the flat part at the bottom of the hill, or on the steeper slope further up the hill?

EDIT:
I got it!  And I think I actually found a variation on it that's pretty reliable.   I just nailed it 5 times in a row.  Of course, we'll have to see if I can nail it in the context of a full run -- often things seem easier when you're playing the same spot over and over, back to back.
I do the jump standing still, facing left.  Then you actually jump up on onto the tire, turn right, and jump again and you'll be flung off it to the right.  I think it helps to stand in the right spot (the ground-plane is slightly wavy, and some spots seem to work better than others).  I'll post a vid when I get a chance.

feasel

Edit history:

feasel: 2011-03-10 01:02:04 pm

feasel: 2011-03-10 12:52:25 pm

2011-03-10 12:52:03 pm

difficult and stupidly random

Jumpin' the tire:

Got to remember to scroll the screen far enough to spawn both objects (doesn't seem to save any time if you jump them together or separately, but it looks cooler doing them together). Also, I can't seem to stand any further to the right and still get the trick to work consistently -- I think that one spot may be special (because of the non-flatness of the ground?).

Heidrage

2011-03-10 12:56:15 pm

Willing to teach you the impossible

I really wish we were able to get that Arrow skip to work... That would be like a minute or so just with that. Found out that Carc has this game, so I could tinker around with it a bit and see if I can find anything

zech48

Edit history:

zech48: 2011-03-12 10:22:12 am

2011-03-12 10:21:50 am

i have to ask why you left the gore filter off. as far as i can tell thats just +7s

Sir VG

2011-03-12 11:01:12 am

Fucking Weeaboo

Quote from zech48:

i have to ask why you left the gore filter off. as far as i can tell thats just +7s

Well, turning it on is kinda pointless I think if you don't die. Not like it adds time, but it's totally a runner's choice.

Pootrain

2011-03-12 02:33:39 pm

train kept rollin

I've been messing about with the arrow for a bit today, gonna continue and see if I can get anything to work

zech48

Edit history:

zech48: 2011-03-12 06:26:21 pm

zech48: 2011-03-12 05:58:51 pm

2011-03-12 05:33:14 pm

Quote from Sir VG:

Well, turning it on is kinda pointless I think if you don't die. Not like it adds time, but it's totally a runner's choice.

no it saves time on the maggot kid, im curious because i already told feasel this

also theres a free-fall barrier and an invisible wall on the arrow. the wall doesn't matter (unless it goes all the way down, i cant tell) but the barrier means that even if you get a landing where you would survive you dont.
im just saying this from observation, so i cant be sure. here's as close as i got.

i was trying to be in the middle of the climbing animation to see if that made limboy invulnerable but it does not.

[quote="Sir VG"]Well, turning it on is kinda pointless I think if you don't die.  Not like it adds time, but it's totally a runner's choice.[/quote]

no it saves time on the maggot kid, im curious because i already told feasel this

also theres a free-fall barrier and an invisible wall on the arrow. the wall doesn't matter (unless it goes all the way down, i cant tell) but the barrier means that even if you get a landing where you [i]would[/i] survive you dont.
im just saying this from observation, so i cant be sure. here's as close as i got. [hidden]<object width="640" height="390"><param name="movie" value="http://www.youtube.com/v/MBsgP2kSYUY?fs=1&amp;hl=en_US"></param><param name="allowFullScreen" value="true"></param><param name="allowscriptaccess" value="always"></param><embed src="http://www.youtube.com/v/MBsgP2kSYUY?fs=1&amp;hl=en_US" type="application/x-shockwave-flash" allowscriptaccess="always" allowfullscreen="true" width="640" height="390"></embed></object>[/hidden]

i was trying to be in the middle of the climbing animation to see if that made limboy invulnerable but it does not.

Sir VG

2011-03-13 01:33:23 am

Fucking Weeaboo

There's a rope on the left side near the bottom. What happens if you jump over to that and cling on?

Caveberry

2011-03-13 03:21:25 am

trepidation

I want this game so bad, but I have a PS3. Sometimes, I think of getting a 360 just for their XBLA games. I'm kicking myself for not having Super Meat Boy right now.

I can't wait to see how this run progresses....you lucky bastards...

zech48

2011-03-13 10:39:43 am

Quote from Sir VG:

There's a rope on the left side near the bottom. What happens if you jump over to that and cling on?

i dont think you can grab ropes at semi-high velocity, which you would probably have if you were all the way to the left. but regardless its not a real rope since limboy doesn't even reach for it like he does for real ropes when you fly past them.

feasel

2011-03-13 06:48:30 pm

difficult and stupidly random

Like Zech said, the only place where the gore filter affects gameplay is when it skips the animation of the guy with the brain-slug walking down the hill and drowning. Roughly 7 seconds. Other than that one spot, it's just a bunch of cosmetic differences (e.g. hung bodies replaced by hanging tires).

So the reason I haven't been doing this is that I think it takes away some of the aesthetic appeal of the game (no mangled corpses, no drowned bodies, no dead children in cages). And also, gore-enabled is the default setting so it seemed reasonable to just keep it that way. The differnce is a fixed amount of time, so if you're comparing two runs (one with and one without the filter), you can subtract this out. However, I suppose it may be difficult to compute what that time-difference is down the frame level of precision (or at least, it would take a little extra work to do so). So I could go either way.

It would be nice to know what folks think about this issue. Is there precedent for dealing with these situations? I'll do it whichever way people think is best.

Like Zech said, the only place where the gore filter affects gameplay is when it skips the animation of the guy with the brain-slug walking down the hill and drowning.  Roughly 7 seconds.  Other than that one spot, it's just a bunch of cosmetic differences (e.g. hung bodies replaced by hanging tires).

So the reason I haven't been doing this is that I think it takes away some of the aesthetic appeal of the game (no mangled corpses, no drowned bodies, no dead children in cages).  And also, gore-enabled is the default setting so it seemed reasonable to just keep it that way.  The differnce is a fixed amount of time, so if you're comparing two runs (one with and one without the filter), you can subtract this out.  However, I suppose it may be difficult to compute what that time-difference is down the frame level of precision (or at least, it would take a little extra work to do so).  So I could go either way.

It would be nice to know what folks think about this issue.  Is there precedent for dealing with these situations?  I'll do it whichever way people think is best.

ZenicReverie

2011-03-13 07:01:45 pm

Waiting hurts my soul...

I'd say it's like a cutscene, and should be skipped if possible. I do agree though, those changes, while cosmetic, would detract from the aesthetic. Since it's only 7 seconds, I'd say it's your choice, but better to confirm with Mike.

feasel

2011-03-13 08:53:05 pm

difficult and stupidly random

Regarding the Arrow Skip:
I too have encountered this "invisible wall" on the right-side rope. The kid actually reaches out for the rope, but you are stopped before you get to the rope and just drop straight down.

I like your method of jumping down onto the arrow head (triggering the fall-event while you're still in mid-air) and then jumping right. I hadn't found that one. You know about the "climb down, climb up, jump right" method. And there's another one where you can start by hanging down off the "elbow" of the arrow and letting go (which will cause you to land on the fringe of the arrowhead, triggering the fall-event), and then jumping to the left at a certain moment so that you grab back onto the elbow again, and then jumping rightward to the rope. Using that technique I've managed to get some good solid hits against the invisible wall.

As far as the left-side rope, I haven't found a way to get there. The closest I can come is to do the above-mentioned method to grab back onto the elbow and then climb up it and try to jump leftwards. But there isn't quite enough time to do this.

Regarding the Arrow Skip:
I too have encountered this "invisible wall" on the right-side rope.  The kid actually reaches out for the rope, but you are stopped before you get to the rope and just drop straight down.

I like your method of jumping down onto the arrow head (triggering the fall-event while you're still in mid-air) and then jumping right.  I hadn't found that one.  You know about the "climb down, climb up, jump right" method.  And there's another one where you can start by hanging down off the "elbow" of the arrow and letting go (which will cause you to land on the fringe of the arrowhead, triggering the fall-event), and then jumping to the left at a certain moment so that you grab back onto the elbow again, and then jumping rightward to the rope.  Using that technique I've managed to get some good solid hits against the invisible wall.

As far as the left-side rope, I haven't found a way to get there.  The closest I can come is to do the above-mentioned method to grab back onto the elbow and then climb up it and try to jump leftwards.  But there isn't quite enough time to do this.

zech48

2011-03-13 09:06:37 pm

if you cant grab the left rope before the arrow i see no reason why you should be able to afterwards.

as for the filter, yeah it kind of subtracts from the game but thats what we're doing here. its not like we can maintain the atmosphere with you jumping everywhere anyway. on that note, why were you jumping all the time? did you time it to be faster or are you just playing around? i cant tell.

i have some notes on your run but rather than try and describe parts ill give time-notes later when i have time

feasel

Edit history:

feasel: 2011-03-13 09:27:38 pm

feasel: 2011-03-13 09:26:51 pm

feasel: 2011-03-13 09:11:48 pm

feasel: 2011-03-13 09:11:16 pm

2011-03-13 09:10:21 pm

difficult and stupidly random

~~Because i'm a spaz.~~ Running jump is about 20% faster than running.

edit: Turns out, even up-hill it is faster to do running jump (though not by much).
The one exception is extremely steep up-hill slopes (basically only occurring during the rotating-world parts) where just holding the jump button is the fastest (e.g. 7:33 of part3).

zech48

2011-03-16 07:23:21 pm

part 2- 4:05 hang off the edge so you can drop on the dog more precisely
part 3- 3:20 (the fly) you dont have to sneak till almost exactly halfway through the pool
5:05 pull the tire dont push it
part 4- 8:10 this:

regarding timing. there is no visual cue for beginning of character control so i would use fade in

feasel

Edit history:

feasel: 2011-03-17 01:38:43 pm

feasel: 2011-03-17 01:28:54 pm

feasel: 2011-03-17 01:25:46 pm

2011-03-17 01:20:29 pm

difficult and stupidly random

Thanks Zech48. Lotsa good stuff. Itemized below:

Quote from zech48:

part 2- 4:05 hang off the edge so you can drop on the dog more precisely

Good call. Don't know how i missed that.

Quote from zech48:

part 3- 3:20 (the fly) you dont have to sneak till almost exactly halfway through the pool

Yeah I know where the cutoff is -- I just kinda spaced out there. Smiley

The thing i really need to work on, though, is getting a sense of exactly how fast you can walk without triggering the bug. The threshold is somewhere in the middle of your walking range, so it's possible to lose time by walking too slow (which i often do).

Quote from zech48:

5:05 pull the tire dont push it

Ah. Took me a while to figure out how to do it with your method -- got to deal with the box differently: you jump across the gap, pull the box left *just slightly* (can't spend too long pulling else the door comes down) and then there is just barely enough time to get up on top of the box and through the door before it closes for the first time. (after that point it's the same: you're on the right side of the door, and the box is caught in the door). It's quite a bit tougher this way, but it does save 4 seconds, so i'll definitely incorporate it. Thanks.

Quote from zech48:

part 4- 8:10 this: ...

That is neat! I had no idea you could slip in one last jump there, well AFTER the point where we thought character control had ended. Do you know if it actually causes the screen to fade out any sooner? If so, that is another good reason to use fade-out as the end of timing.

Quote from zech48:

regarding timing. there is no visual cue for beginning of character control so i would use fade in.

There is totally a visual cue for the beginning of character control. When the boy is standing, he is straight up and down. When character control begins and the boy starts running, he is inclined forward. The transition between straight up and down and inclined forward happens in a single frame and is really really easy to spot.
Using fade-in as the startpoint would make it consistent with the endpoint (fade-out), but using the beginning of character control would make it consistent with most of the other runs on SDA. So...

Thanks Zech48.  Lotsa good stuff.  Itemized below:

[quote="zech48"]part 2- 4:05 hang off the edge so you can drop on the dog more precisely
[/quote]
Good call.  Don't know how i missed that.

[quote="zech48"]
part 3- 3:20 (the fly) you dont have to sneak till almost exactly halfway through the pool[/quote]
Yeah I know where the cutoff is --  I just kinda spaced out there. :)
The thing i really need to work on, though, is getting a sense of exactly how fast you can walk without triggering the bug.  The threshold is somewhere in the middle of your walking range, so it's possible to lose time by walking too slow (which i often do).

[quote="zech48"]
5:05 pull the tire dont push it[/quote]
Ah.  Took me a while to figure out how to do it with your method -- got to deal with the box differently:  you jump across the gap, pull the box left *just slightly* (can't spend too long pulling else the door comes down) and then there is just barely enough time to get up on top of the box and through the door before it closes for the first time.  (after that point it's the same:  you're on the right side of the door, and the box is caught in the door).  It's quite a bit tougher this way, but it does save 4 seconds, so i'll definitely incorporate it.  Thanks.

[quote="zech48"]
part 4- 8:10 this:  ...[/quote]
That is neat!  I had no idea you could slip in one last jump there, well AFTER the point where we thought character control had ended.  Do you know if it actually causes the screen to fade out any sooner?  If so, that is another good reason to use fade-out as the end of timing.

[quote="zech48"]
regarding timing. there is no visual cue for beginning of character control so i would use fade in.[/quote]
There is totally a visual cue for the beginning of character control.  When the boy is standing, he is straight up and down.  When character control begins and the boy starts running, he is inclined forward.  The transition between straight up and down and inclined forward happens in a single frame and is really really easy to spot.
Using fade-in as the startpoint would make it consistent with the endpoint (fade-out), but using the beginning of character control would make it consistent with most of the other runs on SDA.  So...

zech48

2011-03-17 02:13:22 pm

dont forget that you have to wake him up

feasel

Edit history:

feasel: 2011-03-17 10:02:02 pm

2011-03-17 10:00:01 pm

difficult and stupidly random

I think "character control" implies that you actually have control of the character and can move him around. You can think of Limbo as being analogous to a game that begins with some text boxes you have to mash through before you actually gain control of the character. The precedent in cases like this is to use the point where you start moving the character, not the point where you have to press a button to start paging through the text boxes.

For people not familiar with the game: After selecting "new game" at the menu screen, the screen goes black and then fades in to show the boy unconscious. (fade-in is zech48's startpoint recommendation.) Then the game will sit there indefinitely until you press a button (for which there is no visual cue -- so nobody is recommending this as the startpoint). Then it plays an animation of the boy getting up (maybe 8 seconds?) and finally the boy becomes controllable (this is my startpoint recommendation).

Honestly I could go for either one. It just seemed to me like the point where the boy starts moving is more consistent with SDA standards. (And admittedly I like arguing rules trivialities.)

Regarding the suggestion about pulling the tire instead of pushing it (part3, 5:05):
Zech48: Does what i described in the last post sound like what you had in mind? I'm having a hard time consistently getting the box wedged in the door and then climbing over it in time. I'm wondering if i'm not doing it quite the same way you are. For one thing, i'm not sure exactly at what point it's best to let go of the tire. Would it be a lot of work for you to make video? Or even a verbal description of how you do it would help.

I think "character control" implies that you actually have control of the character and can move him around.  You can think of Limbo as being analogous to a game that begins with some text boxes you have to mash through before you actually gain control of the character.  The precedent in cases like this is to use the point where you start moving the character, not the point where you have to press a button to start paging through the text boxes.

For people not familiar with the game:  After selecting "new game" at the menu screen, the screen goes black and then fades in to show the boy unconscious. (fade-in is zech48's startpoint recommendation.)  Then the game will sit there indefinitely until you press a button (for which there is no visual cue -- so nobody is recommending this as the startpoint).  Then it plays an animation of the boy getting up (maybe 8 seconds?) and finally the boy becomes controllable (this is my startpoint recommendation).

Honestly I could go for either one.  It just seemed to me like the point where the boy starts moving is more consistent with SDA standards.  (And admittedly I like arguing rules trivialities.)

Regarding the suggestion about pulling the tire instead of pushing it (part3, 5:05):
Zech48:  Does what i described in the last post sound like what you had in mind?  I'm having a hard time consistently getting the box wedged in the door and then climbing over it in time.  I'm wondering if i'm not doing it quite the same way you are.  For one thing, i'm not sure exactly at what point it's best to let go of the tire.  Would it be a lot of work for you to make video?  Or even a verbal description of how you do it would help.

Limbo fan

2011-03-18 03:18:14 pm

This was a GREAT run - I notice your small but not unimportant details and timings you have memorized... very good!

Here are my small contributions - some people might think I'm nitpicking but trust me, I'm only doing this out of respect for feasel and a (close to) perfect run:

Part 1

01:08 - I feel there is a bit of excessive hopping right here, just drop over the edge and grap the rope instead of jumping over the edge.

02:21 - Do you really need that much momentum to jump over the trap?

Part 2

11:52 - I notice that you somehow skip the "zombification" animation when you're being "possessed" by the larva (that little dance he does) - how do you do that and why don't you do it any other time?

Part 3

02:46 - Perhaps you can time the swinging a little bit better? I actually don't know what the best scenario is here -
1. to land as close to the edge as possible or
2. get as much momentum as possible by swinging to "boost" when you release.

03:46 - I forget at the moment but it's not possible to release the insects leg? You have to wait for it to break off?

04:05 - Again, excessive jumping - try not to jump over the edge, and if you have to, jump so that you don't hit anything on the way down :).

05:01 - I just might contradict myself on this one but isn't it faster to jump over the ladder all the way down instead of climbing a bit down the ladder first? Or do you die from the fall?

05:25 - Damn edge grabs Smiley

06:08 - Can't you use the brilliant tactic of pulling one box while pushing the other? (You do this in part 4 - 03:23... it's a GREAT tactic, something I would have NEVER figured out).

13:37 - You jump away from the box instead of just dropping.

Part 4

02:17 - Do you think it is possible to get on the right side of the box before it's shot down?

03:09 - Isn't it possible to stand a little bit to the right of the magnet switch, jump left and trigger the switch at the same time? I would think you can gain a couple of footsteps or about 0.5 seconds if it works.

Again, good job on the run - damn good job!

This was a GREAT run - I notice your small but not unimportant details and timings you have memorized... very good!

Here are my small contributions - some people might think I'm nitpicking but trust me, I'm only doing this out of respect for feasel and a (close to) perfect run:

[b]Part 1[/b]

01:08 - I feel there is a bit of excessive hopping right here, just [u]drop[/u] over the edge and grap the rope instead of [u]jumping[/u] over the edge.

02:21 - Do you really need that much momentum to jump over the trap?

[b]Part 2[/b]

11:52 - I notice that you somehow skip the "zombification" animation when you're being "possessed" by the larva (that little dance he does) - how do you do that and why don't you do it any other time?

[b]Part 3[/b]

02:46 - Perhaps you can time the swinging a little bit better? I actually don't know what the best scenario is here -
1. to land as close to the edge as possible or
2. get as much momentum as possible by swinging to "boost" when you release.

03:46 - I forget at the moment but it's not possible to release the insects leg? You have to wait for it to break off?

04:05 - Again, excessive jumping - try not to jump over the edge, and if you have to, jump so that you don't hit anything on the way down :).

05:01 - I just might contradict myself on this one but isn't it faster to jump over the ladder all the way down instead of climbing a bit down the ladder first? Or do you die from the fall?

05:25 - Damn edge grabs :)

06:08 - Can't you use the brilliant tactic of pulling one box while pushing the other? (You do this in part 4 - 03:23... it's a GREAT tactic, something I would have NEVER figured out).

13:37 - You jump away from the box instead of just dropping.

[b]Part 4[/b]

02:17 - Do you think it is possible to get on the right side of the box before it's shot down?

03:09 - Isn't it possible to stand a little bit to the right of the magnet switch, jump left and trigger the switch at the same time? I would think you can gain a couple of footsteps or about 0.5 seconds if it works.

Again, good job on the run - damn good job!