My feelings on The Demon Rush
Game Page: Doesn't exist yet.
Verifier Responses
Enhasa and I discussed this a bit
I then asked the last verifier to give a more detailed explanation of what the runner did wrong. The following is what he gave me.
This is what the runner had to say in response
Decision: Reject
Reason: Lolo is one of those games where the solutions matter just as much, if not more than the execution. :-/
The runner is happy to improve his run though, so we're still a big, happy family.
The run is available upon request, PM me (lasts until June 15th).
Verifier Responses
Quote:
The runner did an excellent job on this run. The amount of memorization required to speedrun a game like Lolo is, frankly, insane, and the runner did it beautifully. Each room was cleared proficiently and quickly, and being only 90 seconds behind the TAS time for a twenty minute game is amazing work under any circumstances. Accept.
Quote:
Audio/video: good
Cheating: no
Fast: yes
Has some speed/safety tradeoffs, which are practically unavoidable in a Lolo run. Some individual level comments:
2-5: I'm surprised that a safe route exists without losing much time at all.
4-1: good Don Medusa dodging
5-2: It's possible (and faster) to turn the arrow while pushing a block over it.
6-1: Some ugly movement at the beginning of this room.
7-5: Good strategy, but it would be faster to push the upper block in position while going for the hearts in the middle.
8-4: This route struck me as slow for some reason.
Overall, the run was good, so my verdict is accept.
Cheating: no
Fast: yes
Has some speed/safety tradeoffs, which are practically unavoidable in a Lolo run. Some individual level comments:
2-5: I'm surprised that a safe route exists without losing much time at all.
4-1: good Don Medusa dodging
5-2: It's possible (and faster) to turn the arrow while pushing a block over it.
6-1: Some ugly movement at the beginning of this room.
7-5: Good strategy, but it would be faster to push the upper block in position while going for the hearts in the middle.
8-4: This route struck me as slow for some reason.
Overall, the run was good, so my verdict is accept.
Quote:
The sound and video are fine enough
No cheating involved
The execution of this run is excellent. Very few pauses, I seem to recall only one misstep, maybe three instances of pushing blocks too far. Very acceptable for a 23min run. However, it is route that is of most concern.
This is simply not true. Your route is very much lacking in comparison to known best routes.
Level 1
All is good. You mention in your submision text about 1-4, However If you take the bottom route as in the TAS it avoids enemy randomness and gains a frame
Level 2
2-2 is short a step for no reason
Level 3
Although I do like your 3-1 solution it is now 2 seconds slower than the known best without being any less difficult. Rest of the level looks great.
Level 4
4-1 is good. 4-3 is about a second slow. The TAS Trick for this level is much easier than it looks to do in realtime.
Level 5
This level was fine besides 5-2 it is inexcusable that you didnt realize you could push the boxes through a wrong way arrow before changing it. This knowlege would save you 2 seconds.
Level 6
6-1 I have no problem with your hesitation in this level, but your route is about a second longer again. 6-3 although long, is ok as other solutions ramp the difficulty exponentaily.The last block is pushed too far though. 6-5 Why dont you shoot down at the snake first before the left island? You avoid the pause before pushing him later.
Level 7
The easiest level. You do manage do masacre 7-5. Im not even going to count how many seconds slower about 5 seconds. A far inferior route. There's no randomness here, enough said.
Level 8
This level you lose a considerable amount of time on
8-1 Fine
8-2 Two steps inferior route
8-3 One step lost no biggy
8-4 This level you lose 4 seconds from best and this level has no enemy reaction
8-5 Is at about 4 seconds too slow.
All these solutions can be better without the risking the success
of your run.
Level 9
Your 9-1 solution looks nice, but is a second slow. Ugh, 9-3 you can wait further out in front of the medussa, anticipate the sleeper and start moving early. Plus a different shot strategy will shave off 2 seconds. Using the TAS method here will save you 1/2 second in 9-4. Your 9-5 is a second slow
Level 10
However much I hate the 10-1 solution it is safest. The Don Medussa is pushed half too far into the corner. Everything about 10-2 is off. I'm not going to bother with details but say its a good 5 seconds longer. Its longer too push the egg up then down dodge the fire in realtime by a couple of steps. Your strategy for 10-3 is 2 seconds slow. 10-4 is safe but fine and 10-5 is great.
With this being true the Adventures of Lolo is about knowlege. Without the knowledge of said known routes how could your run be as fast as it should be? This is what I find most puzzling, just in the Lolo SDA forum here there are links to the best and most current routes available. Why were these ignored? I'm not sure when you produced this run but were you aware that there is a new TAS by Baxter & Zugzwang. He has been posting improvements at TASVideos for months now. Any number of individuals who have worked on the Lolo TAS would have been happy to give you feedback.....just like this.
You have gone the length to make a virtually mistake free run just clean up your route
I do have Reject your run. The true test is for each instance above I mentioned it was too slow, multiply each second by 4. Now dance around the first stage that number of steps. Would you submit that run?
No cheating involved
The execution of this run is excellent. Very few pauses, I seem to recall only one misstep, maybe three instances of pushing blocks too far. Very acceptable for a 23min run. However, it is route that is of most concern.
Quote:
In the majority of the rooms, the route i take is optimal (at least, with respect to all known tricks). Any difference between the console time and the optimal (best known) route is due to a few lost frames here and there while changing direction.
This is simply not true. Your route is very much lacking in comparison to known best routes.
Level 1
All is good. You mention in your submision text about 1-4, However If you take the bottom route as in the TAS it avoids enemy randomness and gains a frame
Level 2
2-2 is short a step for no reason
Level 3
Although I do like your 3-1 solution it is now 2 seconds slower than the known best without being any less difficult. Rest of the level looks great.
Level 4
4-1 is good. 4-3 is about a second slow. The TAS Trick for this level is much easier than it looks to do in realtime.
Level 5
This level was fine besides 5-2 it is inexcusable that you didnt realize you could push the boxes through a wrong way arrow before changing it. This knowlege would save you 2 seconds.
Level 6
6-1 I have no problem with your hesitation in this level, but your route is about a second longer again. 6-3 although long, is ok as other solutions ramp the difficulty exponentaily.The last block is pushed too far though. 6-5 Why dont you shoot down at the snake first before the left island? You avoid the pause before pushing him later.
Level 7
The easiest level. You do manage do masacre 7-5. Im not even going to count how many seconds slower about 5 seconds. A far inferior route. There's no randomness here, enough said.
Level 8
This level you lose a considerable amount of time on
8-1 Fine
8-2 Two steps inferior route
8-3 One step lost no biggy
8-4 This level you lose 4 seconds from best and this level has no enemy reaction
8-5 Is at about 4 seconds too slow.
All these solutions can be better without the risking the success
of your run.
Level 9
Your 9-1 solution looks nice, but is a second slow. Ugh, 9-3 you can wait further out in front of the medussa, anticipate the sleeper and start moving early. Plus a different shot strategy will shave off 2 seconds. Using the TAS method here will save you 1/2 second in 9-4. Your 9-5 is a second slow
Level 10
However much I hate the 10-1 solution it is safest. The Don Medussa is pushed half too far into the corner. Everything about 10-2 is off. I'm not going to bother with details but say its a good 5 seconds longer. Its longer too push the egg up then down dodge the fire in realtime by a couple of steps. Your strategy for 10-3 is 2 seconds slow. 10-4 is safe but fine and 10-5 is great.
Quote:
It is all but impossible for someone to claim that they have found the "optimal route" for any game. But with Adventures of Lolo one can come pretty close because the game's possibility space is so limited (the game is on a 21x21 grid, and input is usually only taken every 8 frames (at 60fps), or every half block of the sprite grid). To put it simply: there are only so many possible things to try.
With this being true the Adventures of Lolo is about knowlege. Without the knowledge of said known routes how could your run be as fast as it should be? This is what I find most puzzling, just in the Lolo SDA forum here there are links to the best and most current routes available. Why were these ignored? I'm not sure when you produced this run but were you aware that there is a new TAS by Baxter & Zugzwang. He has been posting improvements at TASVideos for months now. Any number of individuals who have worked on the Lolo TAS would have been happy to give you feedback.....just like this.
You have gone the length to make a virtually mistake free run just clean up your route
I do have Reject your run. The true test is for each instance above I mentioned it was too slow, multiply each second by 4. Now dance around the first stage that number of steps. Would you submit that run?
Enhasa and I discussed this a bit
Quote:
[01:01] kevinjuang: well shit seriously though
[01:01] kevinjuang: if i knew they weren't optimal i would be up for rejecting them
[01:01] kevinjuang: depends how suboptimal they are
[01:02] kevinjuang: what a great verifier i am for this game
[01:02] mikwuyma: well honestly
[01:02] mikwuyma: xyz is the most qualified verifier
[01:02] kevinjuang: if people like us can't tell, right
[01:02] mikwuyma: I can forward you his message
[01:02] mikwuyma: I don't know if he is being overly critical or just dropping lolo knowledge
[01:02] kevinjuang: yeah i would definitely cede to him since he knows
[01:02] mikwuyma: doesn't helpt hat I just skimmed his message
[01:02] mikwuyma: the only problem is
[01:02] kevinjuang: well the thing about lolo is, solutions are just knowledge not really skill
[01:03] mikwuyma: when will xyz complete a run
[01:03] kevinjuang: so if you reject a run and say "here are the better solutions so you can improve your run"
[01:03] kevinjuang: that's a lot different from saying "your run sucks, play better"
[01:03] mikwuyma: and I disagree, even though there is an 8-frame limit
[01:03] mikwuyma: you still have to be tight on execution around don medusas and such
[01:03] kevinjuang: ok that's not what i meant
[01:04] kevinjuang: i'm not saying there's not execution
[01:04] kevinjuang: i'm saying for a puzzle run rejection, there are specific improvements you can list
[01:04] kevinjuang: and then they can just use them
[01:04] kevinjuang: it's knowledge based
[01:01] kevinjuang: if i knew they weren't optimal i would be up for rejecting them
[01:01] kevinjuang: depends how suboptimal they are
[01:02] kevinjuang: what a great verifier i am for this game
[01:02] mikwuyma: well honestly
[01:02] mikwuyma: xyz is the most qualified verifier
[01:02] kevinjuang: if people like us can't tell, right
[01:02] mikwuyma: I can forward you his message
[01:02] mikwuyma: I don't know if he is being overly critical or just dropping lolo knowledge
[01:02] kevinjuang: yeah i would definitely cede to him since he knows
[01:02] mikwuyma: doesn't helpt hat I just skimmed his message
[01:02] mikwuyma: the only problem is
[01:02] kevinjuang: well the thing about lolo is, solutions are just knowledge not really skill
[01:03] mikwuyma: when will xyz complete a run
[01:03] kevinjuang: so if you reject a run and say "here are the better solutions so you can improve your run"
[01:03] kevinjuang: that's a lot different from saying "your run sucks, play better"
[01:03] mikwuyma: and I disagree, even though there is an 8-frame limit
[01:03] mikwuyma: you still have to be tight on execution around don medusas and such
[01:03] kevinjuang: ok that's not what i meant
[01:04] kevinjuang: i'm not saying there's not execution
[01:04] kevinjuang: i'm saying for a puzzle run rejection, there are specific improvements you can list
[01:04] kevinjuang: and then they can just use them
[01:04] kevinjuang: it's knowledge based
I then asked the last verifier to give a more detailed explanation of what the runner did wrong. The following is what he gave me.
Quote:
I was asked to elaborate.
This is a list of improvements to your run that require fairly little effort. The only reason why these are not employed in your run is that you didnt realize they were faster or didnt know they existed. Neither of which are really good reasons for someone who claims to have an "optimal route through a majority of levels".
2-2 1/4 sec (Found in TAS)
You can save a half push on either of the medusa's saving a step. This solution can be mirrored/varied so it looks like your own.
3-1 3 sec (Found in most current TAS)
You have a different bridge placement from the last TAS and your solution is 114 steps to the chest. The newest Improvement has the bridge at the bottom and is 102 steps to the chest. That is 12 steps = 3 seconds better and the solution is not any more difficult than yours.
4-3 1 sec (Found in TAS)
As mentioned, try this TAS strat in real time. Pause briefly (just like you do in 2-1) any time before the final trek left. Youll do it once or twice to get the feel of it. In your solution you have to wait about three steps for the don medussa to pass the block then pushing it a half step equals 4 steps = 1 sec
5-2 1 1/2 sec (You can watch the TAS)
This was the most blatent error in your run. Just push the block straight up and change the arrow when you reach it. Your little excursion takes a little less time as I originaly thought because you get the heart while going, but its still short 6 steps = 1 1/2 sec
6-1 1/2 sec (Newest solution in TAS)
I thought your route was 97 steps to the chest. It is actually 95. The newest route is 94 steps. Your misstep plus the new strat is a 2 step = 1/2 second improvement.
6-3 1/3 sec (This strat is off-hand observation)
Only half block the medussa with the block in this stage will save a step. while you wait for the sleeper to come around at the begining wait in the grass half under the bricks half under heart, saves a half a step. Total about 1/3 of a second
7-5 6 sec (Follow the TAS strat)
Go counter clock-wise just like you do. After all hearts are collected in your same order the skulls are running wild. Walk up change the lowest arrow and move into the center. Yeah, there is no randomness. Furthermore you must have not realized that enemies dont hurt you unless they run into you. You can run/press against them all day without dying. So I didnt count my first go around, but all those steps arranging the blocks are a waste. It ended up being 24 steps = 6 seconds longer
8-2 1/2 sec (TAS Video)
The key to the TAS video is pushing one of the blocks up from the bottom and walking 2 steps less around to push it back down. Take note of the top right move order.
8-4 4 sec (Most Current TAS)
The most current solution to this level is 110 steps yours is 126. 16 steps = 4 sec. As I mentioned there is no enemy interaction in this level, so no excuses for an inferior route.
8-5 4 sec (Zugwangs early 8-5 vid) 4th Clip
Here is the safest tricky 8-5 route. This Earlier TAS improvement is surprisingly easy to accomplish in realtime. It saves steps because it only pushes one don medussa partaily down and does the block pushing much more efficiently. Your route 187 steps this 173. 15 steps = 4sec
9-1 1 sec (Follow the TAS timing.)
The key here is getting the roller to follow you around the bottom circle, doing so keeps you from waiting 4 steps = 1 sec at the top for him to return.
9-3 2 3/4 sec (The last clip in previous link is the safest/fastest route.)
As I mentioned you can anticipate the sleeper and move earlier (just after he turns left) plus you get the heart above the chest while the sleepers are dancing around. You dont wait for the sleeper to block the fire guy, or wait after the sand. 11 steps saved = 2 3/4 sec
9-4 1/2 sec (In TAS Video)
While you wait for the snake to reappear run around and push the block in front of fire guy. Saves 2 steps = 1/2 sec
9-5 1 sec (TAS video)
You can just rearrange your route to save 4 steps. Yours is 35 to the chest The best is 31
10-2 2 1/2 sec (FCM) http://dehacked.2y.net/microstorage.php/info/258544469/Adv-Lolo%2010-2%20Realtime%20strat.fm2
This may be the trickiest but again its not any more difficult to do than yours. I did this in realtime in a single attempt. I counted yours at 139 steps this 129 Aprox. 10 steps faster = 2 1/2 sec
10-3 2 sec (Use the 9nth TAS version of this level)
Follow the TAS route and youll save 8 steps. This room can be a pain to compare so I count how many times the Don Medussa hits the rock you break. Each cycle is four steps yours was 36 times, Best known is 34, the difference of 8 steps = 2 seconds. Again this solution isnt any harder than yours.
Approximately 31 seconds saved total
Note:
All these different routes don't count the handful of solutions that are too difficult to do in realtime.
Again, you executed your run very well, it is your route that is lacking.
Also the Adventures of Lolo topic in this forum has links to all the additional info you should need.
Good Luck!
This is a list of improvements to your run that require fairly little effort. The only reason why these are not employed in your run is that you didnt realize they were faster or didnt know they existed. Neither of which are really good reasons for someone who claims to have an "optimal route through a majority of levels".
2-2 1/4 sec (Found in TAS)
You can save a half push on either of the medusa's saving a step. This solution can be mirrored/varied so it looks like your own.
3-1 3 sec (Found in most current TAS)
You have a different bridge placement from the last TAS and your solution is 114 steps to the chest. The newest Improvement has the bridge at the bottom and is 102 steps to the chest. That is 12 steps = 3 seconds better and the solution is not any more difficult than yours.
4-3 1 sec (Found in TAS)
As mentioned, try this TAS strat in real time. Pause briefly (just like you do in 2-1) any time before the final trek left. Youll do it once or twice to get the feel of it. In your solution you have to wait about three steps for the don medussa to pass the block then pushing it a half step equals 4 steps = 1 sec
5-2 1 1/2 sec (You can watch the TAS)
This was the most blatent error in your run. Just push the block straight up and change the arrow when you reach it. Your little excursion takes a little less time as I originaly thought because you get the heart while going, but its still short 6 steps = 1 1/2 sec
6-1 1/2 sec (Newest solution in TAS)
I thought your route was 97 steps to the chest. It is actually 95. The newest route is 94 steps. Your misstep plus the new strat is a 2 step = 1/2 second improvement.
6-3 1/3 sec (This strat is off-hand observation)
Only half block the medussa with the block in this stage will save a step. while you wait for the sleeper to come around at the begining wait in the grass half under the bricks half under heart, saves a half a step. Total about 1/3 of a second
7-5 6 sec (Follow the TAS strat)
Go counter clock-wise just like you do. After all hearts are collected in your same order the skulls are running wild. Walk up change the lowest arrow and move into the center. Yeah, there is no randomness. Furthermore you must have not realized that enemies dont hurt you unless they run into you. You can run/press against them all day without dying. So I didnt count my first go around, but all those steps arranging the blocks are a waste. It ended up being 24 steps = 6 seconds longer
8-2 1/2 sec (TAS Video)
The key to the TAS video is pushing one of the blocks up from the bottom and walking 2 steps less around to push it back down. Take note of the top right move order.
8-4 4 sec (Most Current TAS)
The most current solution to this level is 110 steps yours is 126. 16 steps = 4 sec. As I mentioned there is no enemy interaction in this level, so no excuses for an inferior route.
8-5 4 sec (Zugwangs early 8-5 vid) 4th Clip
Here is the safest tricky 8-5 route. This Earlier TAS improvement is surprisingly easy to accomplish in realtime. It saves steps because it only pushes one don medussa partaily down and does the block pushing much more efficiently. Your route 187 steps this 173. 15 steps = 4sec
9-1 1 sec (Follow the TAS timing.)
The key here is getting the roller to follow you around the bottom circle, doing so keeps you from waiting 4 steps = 1 sec at the top for him to return.
9-3 2 3/4 sec (The last clip in previous link is the safest/fastest route.)
As I mentioned you can anticipate the sleeper and move earlier (just after he turns left) plus you get the heart above the chest while the sleepers are dancing around. You dont wait for the sleeper to block the fire guy, or wait after the sand. 11 steps saved = 2 3/4 sec
9-4 1/2 sec (In TAS Video)
While you wait for the snake to reappear run around and push the block in front of fire guy. Saves 2 steps = 1/2 sec
9-5 1 sec (TAS video)
You can just rearrange your route to save 4 steps. Yours is 35 to the chest The best is 31
10-2 2 1/2 sec (FCM) http://dehacked.2y.net/microstorage.php/info/258544469/Adv-Lolo%2010-2%20Realtime%20strat.fm2
This may be the trickiest but again its not any more difficult to do than yours. I did this in realtime in a single attempt. I counted yours at 139 steps this 129 Aprox. 10 steps faster = 2 1/2 sec
10-3 2 sec (Use the 9nth TAS version of this level)
Follow the TAS route and youll save 8 steps. This room can be a pain to compare so I count how many times the Don Medussa hits the rock you break. Each cycle is four steps yours was 36 times, Best known is 34, the difference of 8 steps = 2 seconds. Again this solution isnt any harder than yours.
Approximately 31 seconds saved total
Note:
All these different routes don't count the handful of solutions that are too difficult to do in realtime.
Again, you executed your run very well, it is your route that is lacking.
Also the Adventures of Lolo topic in this forum has links to all the additional info you should need.
Good Luck!
This is what the runner had to say in response
Quote:
Ok. Thanks for the details -- you've clearly put a lot of thought into the game as well. (I corrected him about this)
My goal was to plan and execute the route without looking at the TAS (until i was finished), since it kind of takes the fun out of speedrunning if you just watch somebody else's route and copy it -- especially in a game like this.
But i guess if the aim is to have the best possible end-product, without concern for how it was obtained, then it makes sense to crib all the moves from TAS (except those few that are not reasonable for an unassisted run).
I've been working on Lolo2 (using the same general approach), but I will go back and make a better Lolo1 route and re-submit if that's what it takes.
Thanks for the comments
J
My goal was to plan and execute the route without looking at the TAS (until i was finished), since it kind of takes the fun out of speedrunning if you just watch somebody else's route and copy it -- especially in a game like this.
But i guess if the aim is to have the best possible end-product, without concern for how it was obtained, then it makes sense to crib all the moves from TAS (except those few that are not reasonable for an unassisted run).
I've been working on Lolo2 (using the same general approach), but I will go back and make a better Lolo1 route and re-submit if that's what it takes.
Thanks for the comments
J
Quote:
Actually, i just noticed that there was a new Lolo run posted to tasvideos.org several weeks ago -- after i had completed my Lolo run, watched the (now-outdated) TAS, and wrote up my submission notes.
So i guess that explains why there are so many more sub-optimal rooms than i had thought.
So i guess that explains why there are so many more sub-optimal rooms than i had thought.
Decision: Reject
Reason: Lolo is one of those games where the solutions matter just as much, if not more than the execution. :-/
The runner is happy to improve his run though, so we're still a big, happy family.
The run is available upon request, PM me (lasts until June 15th).
Thread title:
