iMacmatician's videos

I know that path (or half of it) but where's the trippiness in that?

Don't have to be the top time. I like seeing the full path done in top times rather than take a shortcut and skip all the interesting stuff.

I also posted some of Pascal's videos even though some of the times are slower than currently posted videos, or even older ones.

Marble Blast Levels: Compilation 25 —

So the video compilation that I mentioned would be released at around the mid-August timeframe has obviously been delayed a bit. All runs for this compilation were completed yesterday night. I still need to add bloopers, etc. before I actually release the video, but you’ll see it within the next week or so. In the meantime, here’s another video in the Marble Blast Levels series. This one is the first since MBL19 where a majority of its levels are regular levels (actually all of them are).

A lot of these runs were done to beat PRs on levels that I like doing, and shares a number of levels with MBL 18.

Notes:

Three-Fold Maze: I had a 20.76 from 2010 or 2011, I couldn’t beat it for a while and even this run beats it by only 0.04. I let go of the keys right before hitting the yellow part before the last gem, that gave me the natural bounce that makes it much easier to finish with a good time (for me, rolling on the yellow part makes it subsequently very difficult for me to get the last gem and finish).

Around the World in 30 seconds: Previous video PR: 9.81, previous PR: 9.80, this video: 9.79.

Escher’s Race: I had a 28.90 in late 2009 but my PR back then was 28.4x. I had made some attempts to beat it in subsequent years, but all without success. I finally beat it now but even this run messed up on the ending (I should be able to go 0.3-0.5 lower).

Ramps Reloaded: This run differs from my older 14.27 by rolling not bouncing on the part right before the 5th gem. That makes the path considerably harder since it is easy to hit the trim too fast after the 5th gem and go above the 6th gem.

Pathways: A very well done edge hit made all the difference. I wanted to get a 2 edge hit run, which I did while recording this level, but it was only a 24.63 since the 1st EH was nowhere near optimal.

Darwin’s Dilemma: An unexpected sub-7. My PR was 7.4x and my MBL18 video time was 7.78.

King of the Mountain: My PR is 5.17 using a great finish. There’s no realistic way I’ll beat that without using the Ctrl + P Super Speed trick, which I cannot pull off. This time is probably my 2nd best though.

Learn the Friction!: I had a 4.06 on .rec in early 2010, scheduled for MBL6, but I lost some .recs in an accident, and I only got a 4.09 while redoing the level (which you see in MBL6). This compilation finally contains a 4.06 on video.

Take a Stroll…: The ending is actually a tiny bit messed up, although it might not be easy to see. Still, it’s a nice improvement over my earlier 4.26.

Convoluted Helix: Off to get a decent time. My previous .rec time was 8.82 (late 2009).

Getting Squeezed: Only 0.08 away from my WR but uses a different ending.

Quaked Path: My first attempt at a shortcut path.

Songs: Pacific Rim - Pacific Rim, Pacific Rim - No Pulse, Half-Life 2: Episode One - Guard Down

Nice KotM and DD. Looking forward to your video compilation!

Marble Blast Levels: The Matrix Maze Series

Contents:

Overview
History and the revision model
Revision 1
Revision 2
Revision 3
Individual level notes
The scrolling “credits”
Bloopers
Figures
Concluding thoughts

Overview:

Ever since I made a YouTube account (April 2009), I have made a major Marble Blast compilation every summer. To recap, my major compilations in previous years were as follows:
2009: Marble Blast: No Gems,
2010: Marble Blast Levels X,
2011: No Gems: The Second Compilation,
2012: Marble Blast Levels 20.

This year, my major compilation is on the Matrix Maze series by CyberFox.

History and the revision model:

I probably first saw the Matrix Maze levels in the CLA some forgotten time ago, but until last December I never thought much about them and had never seriously tried them (I remember getting 2 minutes and stuff on them in the past). At around the December timeframe I was thinking about what sorts of videos I would release the following summer, and a compilation on the Matrix Maze levels seemed a nice one to have. I chose the Matrix Maze series because of 3 reasons:

C1. It consists of a number of levels under a unifying theme (I would have preferred some more variation in the levels though),
C2. I could work on pathfinding skills because the levels have numerous possible paths,
C3. They were just complex enough to avoid being dull while simple enough so I would be able to have precision runs without too much frustration.

So, I had MBL: C25, the Matrix Maze series, and a 3rd video whose theme I will not yet disclose planned for the summer of 2013. From the start, I decided to not watch any videos of anyone else’s runs of any of the Matrix Maze levels, so that I could find paths by myself. (I do not recall if I watched any back in 2008-2009 when I watched many Marble Blast videos. I know I haven’t watched any for at least the past few years.) I asked calebsoong (a.k.a. eNetro, SPIKErock) for all the times on Imperial’s Matrix Maze series video from 2008 so I could be sure to beat all those times. I also searched YouTube for more videos, and all of them had the times in the title, which made it easier for me. Since most if not all of the existing videos are from 2008 I think, I also knew that I should demonstrate considerably lower times than what was shown in the videos, and I think I have done so for all the levels except for MM 02.

I quickly decided on a revision model while doing runs for the video. The revision model works as follows. I spend a certain timeframe making .recs for all the levels in the Matrix Maze series. After all 13 levels are .rec’d, the outcome is known as revision 1. Then I redo all 13 levels for better times, the outcome is known as revision 2. I continue to do so until I reach a natural number X where I consider myself done with working on .recs for the compilation. In this case, X = 3. The reasons for a revision model were threefold:

R1. Marble Blast was no longer the only game I played regularly, I got to know more games, so I did not usually feel like dedicating 2-3 or more hours in one day to one level.
R2. I may not find a good path the first time I play a level, and if I get off the level, have it in the back of my mind for some time, and go back on later, I have more chances to find new paths. The revision model works well with reasons 1 and 2 since I can think of and introduce new paths and tweaks regularly and most of the levels are of similar difficulty so I wouldn't have one MM level top out at revision 2 and another level go to revision 5.
R3. I seem to improve my skills better in a level if I work on it over a period of time instead of all at once. A step-by-step approach can let me spend more time on a level in total, and also helps limit how frustrated I get during a session when I make too many mistakes and/or don’t perform as well as I would like.

From one level revision to another, I was constantly surprised at how often at least one of these two phenomena happened, even with no new path:

S1. I beat my time on the previous revision easily, given the amount of time I spent on the previous revision,
S2. I beat my time on the previous revision by much more than I thought would be realistic for me, given the amount of time I spent on the previous revision.

I chose to do three revisions for 2 reasons:

T1. 3 revisions would give me ample opportunities to introduce new paths and tweaks as necessary,
T2. If I use too many revisions, I may get to a point where I have a run in revision X that is too difficult to beat for revision X + 1 and no new path or tweak is in sight, and I wanted to avoid such situations.

Due to R2, I didn’t encounter a situation where I “needed” more than 3 revisions or no more than 2 revisions in a run. However, the total amount of time I spent on each level did vary significantly from level to level, but in the end, they all worked within the 3-revision system. I did consider a modification allowing 4 revisions for some levels, though.

I may do a revision system for future MBL: Compilation videos, but to save time, only two revisions on average per level instead of 3, and I'll only create and upload the final revision video. The revision system in this case will be more of a way to spread out my playtimes than to necessarily spend more time per level than before.

Revision 1: —

The runs were made during this timeframe: December 2012 to March 2013.

In this revision, many of my runs were done so I would have some sort of decent run that I could build upon in the future. Many runs did not beat Imperial’s times (most likely due to slower paths), but that’s okay since I had two revisions left. I also got a good feel for the levels so I knew where to focus my efforts in the second revision. I noticed that the MM Platinum levels seemed hard so I put those off until March (see Figure 1), when over half of the MM (Gold and Platinum) levels were done before 2013 started.

Revision 2: —

The runs were made during this timeframe: May 2013 to July 2013.
New paths: MM 01, MM 02, MM 07, MM 09.

Some of the runs were done in May because they seemed “natural” to do and I had faced little difficulties in revision 1. Two exceptions were MM 10, since I did not do the “obvious” edge hit at the end in revision 1, and MM P03, since the revision 1 run was not particularly well done and also much higher (42.35) than the time I saw on YouTube (34.86, WhiteBoomerang). I managed to get a lower time (34.31) but I still knew from watching the run that there would be more to do.

Revision 3: —

The runs were made during this timeframe: July 2013 to August 2013.
New paths: MM 02, MM 03, MM 06.

As you can see in Figure 2 below, I did not spend as much time on revision 2 as I did on revision 1, since I figured that I could just make up the difference in revision 3. That didn’t turn out to be a good move, since I spent more time on revision 3 than on both previous revisions combined, since I wouldn’t have a “revision 4” to fall back onto later if I had a less-than-satisfactory performance on a revision 3 run. My thought in T1 was partially subverted as it now seemed like I had not enough opportunities to work on some levels, but for most if not all levels, the path and most tweaks were finalized after at most 30 minutes to 1 hour into the session, and I just had to spend extra time if necessary trying to get a better time.

I also pushed some levels back until near the end so that I could make use of the extra experience I would get on the earlier levels. A notable example is MM P03, which is tied for 4th earliest in the revision 2 date, but ended up being the last run recorded for revision 3 [Figure 1]. If I had gone for 4 revisions for at least some levels, then MM P03 would have been one of them, and the 2 hours and 43 minutes I spent on it for revision 3 would have been broken up into two sessions [Figure 2].

Individual level notes:

Matrix Maze 01: I did the revision 2 of this run early [Figure 1] since I didn’t like how my revision 1 run of the very first level in the Matrix Maze series was slower than Imperial’s run—I was clearly starting out on the wrong foot!

Matrix Maze 02: This is the only run in revision 2 that doesn’t beat an existing video time (that I know of), in this case, Imperial’s. I had some trouble finding good paths for this level, and I doubt the path I used in revision 3 is optimal.

Matrix Maze 03: I got very annoyed especially during the revision 3 recording because after the 18th gem, I would constantly fall off the platform.

Matrix Maze 04: This level caused me some trouble in revisions 1 and 2, but I got past that and finished the revision 3 run with a nice ending.

Matrix Maze 05: I thought I wouldn’t be able to beat my 29.80 in revision 2, or if I did, beat it by 0.05-0.1 second or so, since I had high 30’s during my practice runs for revision 2 and the 29.80 seemed very well done.

Matrix Maze 06: I know the revision 3 path is 2.5 seconds faster than the revision 2 path, but I can’t ignore the edge hit in the revision 2 path!

Matrix Maze 07: I got a surprise 24.85 after all my runs for revision 2 was over while I wasn’t recording. I spent nearly two hours afterwards to get just 0.02 better on .rec. I got so many 24.95-24.99’s it wasn’t even funny.

Matrix Maze 08: The ending of this run has a strange history. I first tried to get the last gem without bouncing on the platform beforehand, but I couldn’t, hence the bounce in the revision 1 run. Then I tried again during revision 2 and I managed to get it to work using a long run-up before leaving the platform with the second-to-last (13th) gem. It was a bit annoying since half the time I still wouldn’t make it to the last gem. Later, I realized that the reason I couldn’t get the last gem without bouncing was that I was leaving the second-to-last gem platform too close to the “wedge” and the marble was hitting the grey top of the wedge, pushing it down and making it land on the last gem platform too early. So, for revision 3, I exited the second-to-last gem platform slightly away from the wedge, and then I can make it to the last gem with even a relatively short runup.
Also, my original path for this level was getting the gems in basically the opposite order to this path. That path didn’t seem to go below 38.xx.

Matrix Maze 09: One of my favorite edge hits. I didn’t know if the edge hit would really work until I tried the level for revision 3.

Matrix Maze 10: Many edge hits. The one after the 8th gem might seem like just another no-skill edge hit, but it is surprisingly easy to hit it too fast, which results in the marble having too much air time or even missing the 9th gem entirely.

Matrix Maze P01: Sonic claimed that he spent 3 hours finding a good path and 15 minutes getting his 18.34. I spent much less than 3 hours finding this path, but much more than 15 minutes to get even my revision 1 16.52 [Figure 2]. The revision 3 run took less time than I thought given how much I struggled on revision 2, see R3, S1, and S2. 15.46 is also my best final TM catch on that level.

Matrix Maze P02: I honestly thought the combined edge/wall hit (it counts as 1 edge hit and 1 wall hit in the credits) wouldn’t work until revision 3. Even then, my previous idea of just a wall hit wouldn’t and didn’t work, and the marble has to rebound at the right angle for it to even get to the next part of the path.

Matrix Maze P03: I spent more time on this level than on any other level in the Matrix Maze series, and for good reason. This level seemed overwhelming when I first tried to even find all 10 gems and some path that would let me get all of them. It seems natural that this level saw the greatest improvements in final times (and perhaps paths too…) from revision 1 to revision 3. The edge hit after the 4th gem was surprisingly hard to do for me. A successful edge hit gives one of two outcomes:
1. The marble will bounce on the small platform right before the slope with the 5th gem, then one jumps to get the 5th gem.
2. The marble will go straight to the slope and the 5th gem.
#2 is obviously faster but also harder to do.

The scrolling “credits”:

I made the scrolling credits at the end of each revision video as a homage to the Left 4 Dead series of video games which has similar credits (and of course, I use the same soundtrack). Note: I count hits on a steep slope as wall hits, they are functionally very similar.

Bloopers: —

Here, you will see (besides the usual kinds of bloopers) two kinds of bloopers I haven't really shown before.

1. Rarely (it seems to be more common when I'm at a critical part of a run), my Marble Blast screen will skip for a fraction of a second without any warning before going back to normal gameplay. That split second is enough to screw up many runs, including the ones you see here.
2. If the marble sticks to a wall, it is not possible to make it unstick or control its left-right movement—one will just have to wait until it hits another part of the level or it exits from a boundary of the wall.

Also, the last MM P03 blooper was the last run I recorded on .rec. It was after my revision 3 run (recall that MM P03 was the last level done for revision 3), since I wanted a sub-30. Some time after the revision 3 run and before this blooper, I told myself that I would finish the Matrix Maze Series after I finished the next run with a successful edge hit, no matter what happens after the edge hit. It turned out to be this run.

Figures:

Figure 1:


Figure 2:

Note: These times are from the .rec timestamps, which means they do not include a lot of the time spent pathfinding, “trying” the level, or thinking up a new path or tweak before I get into the .rec’ing for a session. That’s partly why the times are somewhat less than I thought they would be (I was expecting closer to 40 hours for the total time for all 3 revisions), another reason is that I probably overestimate the time I spend on a level if I don’t look at a clock or something.

Figure 3:


Concluding thoughts:

I kinda wish I had spent more time pathfinding.

After my final blooper for the Matrix Maze compilation, I wrote on the .rec description,

“Well, that's the end of the Matrix Maze Compilation. I hope you'll enjoy on YouTube.”

I loved watching the revisions right next to each other because it shows the improvements clearly, but also how better you became as time passed even though recording time took way longer (I can attest to it). You can tell the execution becomes tighter every time and you're really pushing everything to the limit.

Onto the runs:

Matrix Maze 1 very very good

Matrix Maze 2 - feels improvable, esp after seeing the last revision.
- after 3rd gem drop to the side. you can gain your diagonal momentum super fast anyhow and you don't seem to require much of it to the 4th/5th gems
- Can use the SJ from 13th to 14th gem. You can easy then move left while you're in this momentum and go up for the SJ again, and then from the leftmost slope just jump to the centre. I suspect this CAN save time, but not 100%

Matrix Maze 3 - rev 3 is great but rev 2 might have the best path potential. Actually pick up the super speed instead of skipping it and instead skip the Super Jump. What you need to do is after 8th gem go mad diagonal/jump and look left to where 10th gem is. As you hit the 9th gem you need to SS towards the 10, thus you don't lose any momentum. You should not only gain more time because you're not slowing down, but you got that speed needed to the 10th gem. You can then get the SJ from the last spot (and again go mad diagonal and SS backwards into the SJ to not lose as much time) and do the end normally.
Rev 3 path might be hardest to incorporate it because it's in a bad spot where you need momentum and the SS position is therefore bad.

Matrix Maze 4 - that edge hit on rev 3 was great!

Matrix Maze 5 - tight!

Matrix Maze 6 - LOVED the new path for revision 3! Really showed the improvement there.

Matrix Maze 7 - great new path on 2nd revision. Revision 3 feels like it can go slightly more tight in execution.

Matrix Maze 8 - tight movements again, well done.

Matrix Maze 9 - Not screwing up 9th gem like that and the edge hit really brought the time down, and execution very well done again.

Matrix Maze 10 - props for managing the ending of rev 3 because WOW that last jump to the finish from last gem is actually harder than it looks afaik.

Matrix Maze P01 - God damn that edge hit was something I thought about while watching rev 2 and you did it on rev 3 and way better than I thought would be possible!

Matrix Maze P02 - Wow that rev 3 execution was excellent. Loved the additions you've done with the edge/wall hits, but it feels like the 3rd/4th gems can be improved with less height.

Matrix Maze P03 - Again in rev 2 I saw an edge hit from 4th to 5th gem and you did it in rev 3! Anyhow rev 3 is still improvable, with a number of hesitations in a few spots that shouldn't happen. Can be pushed to under 30 seconds easily.


Well done all in all, and looking forwards to improvements if you still feel like it, because I can see a few more going lower.

The bloopers? Marble Blast trolling as always. Few facepalms on some though

Sept 2, 2013, 7:41pm, admin wrote:Matrix Maze 2 - feels improvable, esp after seeing the last revision.
- after 3rd gem drop to the side. you can gain your diagonal momentum super fast anyhow and you don't seem to require much of it to the 4th/5th gemsThe drop to the side doesn't work since there is an invisible wall stopping me from dropping or jumping to the side of the walkway (I forgot to mention it in the notes).


Sept 2, 2013, 7:41pm, admin wrote:Matrix Maze 7 - great new path on 2nd revision. Revision 3 feels like it can go slightly more tight in execution.There is no way I can improve rev. 3 by any amount of time worth mentioning, unless there is another tweak. (Although I said the same thing about MM 05 rev. 2. )


Sept 2, 2013, 7:41pm, admin wrote:Matrix Maze P02 - Wow that rev 3 execution was excellent. Loved the additions you've done with the edge/wall hits, but it feels like the 3rd/4th gems can be improved with less height.If I go lower then it is hard or impossible to get from the 3rd gem to the 4th without using extra jumps and it's easy to mess up the run at the 4th gem (so it doesn't really end up faster, although I may not be doing it right). The one in rev. 3 isn't quite optimal, but I took it since the first edge/wall hit isn't easy.

I'll definitely try your path modifications for MM 02 and MM 03 (thanks for them) sometime (I hadn't thought of those), but most likely not anytime soon because I'm done with the Matrix Maze levels for now and I have to start putting my plans for summer 2014 in motion…. Maybe I'll do an Update 1 in the line of the No Gems Compilation updates sometime.

Wow! I'm very impressed on how much work you've put in. Not only you pull the runs off, but also you came up with these paths. Besides the new paths, even the littlest tweaks impress me. You really have a good music choice for these videos. I found this series special because unlike your other MBL videos you have 3 revisions, which really surprised me. I like how you came up with the ending for the audio and video (the credit style). Also the bloopers are trollin' like Matan says. Watching the whole thing feels like it's a sage or a movie. Since this series is worth the wait, I'll give this 10/10 stars! Way to go! (Liked and favorited!).

Good to see someone that created videos of this series which I loved really much. The third revision was really astonishing, the new paths that you came up with were really good. I bet you beat most of my times (forgot what they were as the last time I played the levels was back in 2011), but I'm sure that some can still be optimized with better paths and tweaks, like Matan said. Anyhow, I really appreciate your effort in compiling this spectacular series and I look forward to the major compilation next year.

Roll Only: Ramps Reloaded (19.08, 18.08) —

Notes:
My Ramps Reloaded (19.31) from July 2009 seemed to be anomaly, even 3-4 years afterwards, I still couldn't seem to beat it. Three days ago, I decided to make a larger effort to do so, and after some practice and two tweaks (spot them), I finally got a better time of 19.08. Some bloopers during these attempts led me to think that an edge hit at the very start of the yellow part of the run would push the marble up to what is normally the 6th gem, and could get a significantly faster time. That idea ended up true with my follow-up 18.08 run.

Song: Nightwish - Rest Calm

It's nice to see a video from you again Nice edge hit on the second run!

Marble Blast Levels: Compilation 26 —

Marble Blast Levels: Compilation 27 and 28 are already being worked on and are likely to be released in Q1 2014.

Alice-TC's Mountaintop Return: Two time savers over my previous run of this level. One of them I didn't really bother doing before and the other is obvious in hindsight.

Arrowhead: Long level, can go sub-1:40 with this path but I didn't bother to. This path is not necessarily near optimal, but at least the start is cool. Gems 48-51 were actually the hardest to get, as not getting good spin control will result in a phail. That's why I didn't jump directly to #49 after #48—it's easy for me to mess up there.

Batlements II: An edge hit is obviously possible, but previously I didn't know how to make one that would save time, or at least be worth the effort. The one I did works well. One can also not jump on the blue platform below the first trapdoor, but at least in my experience, that gives much less control over the marble afterwards. The end result is a cost of *greatly* increasing the probability of failure on the edge hit. Thus this tweak may not save much time overall, especially considering the fact that one would have to roll a bit longer on the preceding blue platforms beforehand.

Bridge Over Marble-ed Water: I watched Pascal's run of this level, and I was fairly sure I got a 7.6x or 7.7x back in the old days, but I couldn't remember. I then looked up the level on YouTube and saw that Matan had posted a video in June 2007 with a 6.68 second time. I quickly saw the difference in paths and applied the path Matan used for this run.

Checkpoints 2: A level that one would call a somewhat luck-based level. Contains one obvious mistake but still not a bad run. I bet Pascal could get around 30 seconds.

Orange Obfuscation: One run is a long path, one is a short path, and one is a demonstration of what happens if you follow the helpful tip early in the level.

Red Fold: Two paths, both should go sub-13 with better skills. There's a potential combo path where you edge hit and use the rebound speed to get on the higher part of the level, but I wasn't able to pull it.

The Tower Challange: It's a good thing the Time Modifier is so long!

2 Islands: The jumps to the tele porters on the yellow part of the level need to be done with precision or else will end in failure.

Song(s): I don't know, from a CD.

Gotta love these compilations. Nice job!

Great job with all levels, good fails too!

Marble Blast Levels: Compilation 27

As you can see, I have reached the custom level packs with the Pokko's level collection in them.

Notes:

Agility Island: One needs a high degree of precision to get a low time in this level.

Combo Trick: You can get the TM using the bounce-jump-bounce-jump-bounce-jump start pattern, but then it's slower to get to the finish and the time restarts longer.

Construction Ascent: Obviously a sub-3 is possible. This level annoyed me the most due to three main parts of the level:
1. The OOB trigger that is placed so close to the TM that I'm convinced the level creator put it there specifically to screw up attempts to catch the TM. Many runs would have worked, in that I landed on the trim after catching the TM, if it weren't for this trigger.
2. The sideways Gravity Defier that messed up a lot of runs.
3. I would get the wrong amount of velocity or something at the end, so the Gravity Defier at the end flings me OOB instead of to the finish pad. The finish I did in this run is better than the one I normally do (just jump on the GD) but even this has room for error.

First Level of 2011!!: The windowed videos are made that way using Keynote. I found out that the new version of Keynote does not let any slide be longer than 60 seconds, so I had to break up the video into two parts and join them in iMovie. (Too bad I didn't think to use my copy of Keynote '09.) The teleporter after diamond #4 requires some precision. One has to roll off the castle into the platform with the teleporter because if one jumps or bounces off, the marble will rebound on the platform just high enough to briefly put it out of the range of the teleporter, and one will waste time.

Galaxy Invader's Quick Course: Late in my time of attempts, I realized that one does not need to bounce more than once on any platform in the finish area.

Island Challenge: There are (at least) three ways to do the double Jump Boost part; this is the fastest way I know. The ending isn't quite optimal but I'm not too picky.

Jump!: One can edge hit the platform with the Jump Boost, but it's harder to do and I don't think it's any faster. I actually failed a lot of times between the landing after gem #2 and collecting gem #3.

Jumping Tutorial: I missed the finish pad a few times. This time I didn't since after landing on the finish platform, I rolled instead of jumped.

Out and Back: None.

Pokko's level 3: I like a level where you do a blind gem catch and then finish right afterwards.

Pokko's level 5: Originally I just had the 9.00, but after watching a video on YouTube by Alexis (I think), I learned of this new path (which I previously thought might not be possible).

Pokko's level 16: The ending is pretty fast, reminds me of the ending of Pitfalls.

Pokko's level 20: Imperial (I think)'s path plus an edge hit. Half the reason why I added the edge hit was because I couldn't seem to beat his time otherwise. There are two ways of jumping to wall-hit and getting the second TM, the "fast" (TM catch average 2.8x) and the "slow" way (TM catch average 2.9x-3.0x). This video features the "slow" way but the TM catch is more like the "fast" way.

Pokko's level 21: I didn't know if the edge hit was possible, but I tried a bit and saw that it was.

Pokko's level 24: Probably one of the sneakiest edge hits I've ever done. Yes there are two but I think you'll know which one I'm talking about.

Small loop: Surprisingly frustrating. Many runs went OOB with 1 or 0 diamonds remaining, and even this run nearly failed at the end. (I actually had a run that finished with an edge hit, wasting precious time.

Tasty Level: Wall hits everywhere! I counted 11, which is one wall hit for each diamond and Time Modifier in the level.

Treetop Platformer: The ending seems kinda luck-based, since I could easily fall into a gap instead of successfully jumping to the finish.

Upper Swivel: None.

Your First Roll: None.

Songs: Don't know, they're from the same CD I got the songs used in MBL: C26.

Amazing job as always!

Marble Blast Levels: Compilation 28

I haven't really been working on my upcoming summer 2014 video, instead you get more Marble Blast Levels.

Notes:

Castle Raid: I tried to do an edge-hit on the gem platform then wall-hit on the faraway wall to get on top of the structure without using any Jump Boosts other than the one I collected at the start. Unfortunately I couldn't make it, but I came somewhat close.

Champs Challenger: I really like this path, or at least the first few seconds. I think I missed some 0.86-0.88's (successfully made it to the main part of the level, then went OOB somehow).

Consturves: None.

The High and Low Road 2: Just for the start pad trick.

-Martin's Mission No. 1-: Kinda tricky level but it's not too hard once I got the hang of it. The Gyrocopter parts aren't the most optimal. Notice that I nearly messed up the ending.

Martin's Run 1: None.

Pokko's level 17: None.

Quick-Path 6: Best TM catch 14.4x, I'm not sure how. Also it's possible to finish without the time restarting but I didn't make that on this run (or most runs). This level is quite annoying since one has to gain a lot of momentum on the slopes and when I jump to the 1st slope after the 1st gem, I have to get on the slope properly or I will not move up as high as I should or I will fall off the side of the slope.

Rooftop Travel: The first (alphabetically) of three levels in this compilation that let me practice the double Super Jump. Can go sub-14, maybe I will try it later.

SkyClimb: The second such level, the first 20 seconds weren't done very well but the marble movement during most of the rest of the run was done better than in any of my other practice runs, so this run was my best time by quite a bit.

SkyScraper Hop: The part after the "roads" can be challenging because one can easily waste air time after the Super Jumps, and there seems to be no easy way to solve that. I ended up getting the last gem first and do a precise double Super Jump to the finish, which means less air time overall.

Up the Way High: One needs to get all 6 TMs to finish without time restarting. On many, many runs, I missed one or more of the TMs, and most of the time it wasn't possible to get the missed TM(s) without wasting even more time.

Songs: From the same CD as for MBL: C26 and C27.

Great stuff as always Imac! Some insane EHs there, and good music, keep it up! And hi, been a while.

dushine wrote: Great stuff as always Imac! Some insane EHs there, and good music, keep it up! And hi, been a while.

Thanks and hi Dushine! (I haven't check this thread since I last posted.)



Marble Blast Levels: Compilation 29

I went through some more custom levels. Some longer custom levels as well as regular levels are planned for subsequent compilations.

Comments:

Bounce and Roll: I had a 14.10 in the compilation and I actually uploaded it and was writing the comments when I thought that I could edge-hit up to the bouncy floor 1 higher than the one I landed on. I tried it and it was a bit harder but very doable, and saves around a second.

The Bounce The Eazy Way: One can do an edge hit on the last platform before the second bouncy floor, but I couldn't properly get to the finish afterward (it always seemed to take me 2 bounces, which wastes time).

Castle Cross: Can be annoying since one needs good spin to stay on all the catwalks.

Checkerboard: I tried to get a low edge hit and good jumping on all the squares.

Dynamincs: Apparently it is faster to not jump after the 2nd through 4th gems—it looks like one gets better marble control that way.

Invertigo: I've known this trick for years, but this is the first level I found that had an obvious shortcut using the trick. It’s not easy to do right, especially when it seems like there’s no way to know how high the marble will go until afterwards.

Koopshi's Skill Tower: It's a good thing those mines are there.

Marble Mountain: I would constantly go too high on the fan portion or would go off to the side.

Planks: Kinda like Castle Cross, also I think the planks after the 2nd platform could be done faster, but that’s harder.

Roll and Jump: This run features a jump and a roll.

The Mountaintop: 2 edge hits, good enough for me.

Uvodd's Super Short Quarked Course: If one is not careful at the part around the ice, one will jump/bounce more than once on the ice (or fall off entirely), wasting time.

Songs: From the same CD as in the last 3 videos.

Also, here's a chart showing all my Marble Blast Compilation videos. The image is too big to neatly post in this thread, but it can be seen here . The templates listed are just examples, so compilations like the Matrix Maze series are not included.

Technical stuff below:

Let us make a note about the "f(L) Gems" template. In the future, I may end up making more videos based on gem modifications (stuff other than just no gems is much more complicated, that's why I haven't done many so far). Let L be a level and let g := g(L) be the number of gems in L.

Let n is some nonnegative integer. Then No Gems and 1 Gem are special cases of n Gems, where n = 0 or 1. It is understood that a level with 2 gems in a "4 Gems" compilation will have its gem count unchanged (although there is no reason why we cannot add gems).

Let h be a function that depends only on the number of gems in L. Then n Gems is a special case of h(g) Gems, where h(g(L)) = max{n, g(L)}.

Let f is a function that depends on the entire level L, including gem placements and Powerups. Then h(g) Gems is a special case of f(L) Gems, where f(L) = h(g(L)). So we could have a function that gives us the number of gems that are not floating in the air on a non-moving level.

In the above scenario, we would probably want to remove all the floating gems, but the function only tells us how many gems are to be removed, not which ones. That's not a problem since we can pick which gems to remove, but we can define a new function F which tells us exactly which gems to remove. In practice I'll probably end up using F rather than f….

iMac i am a big fan of yours and I hope you do much more!

Thanks! Expect to see my next video about two or three weeks from now.

Marble Blast Levels: Compilation 30



So I wouldn’t say that this video is anything “special” like Marble Blast Levels X or Marble Blast Levels 20, but that’s because I’m going to have a triplet of videos, MBL: C31, C32, and C33, instead of one video. These videos will contain the following kinds of levels:
MBL: C31: Long levels
MBL: C32: MBG/MBP regular levels
MBL: C33: Old levels

Notes:

Anna & Susan: What I like about this path is the wall-hits after the last two gems. P.S. This level would normally be considered long for my compilations, but MBL: C31 will have longer levels than this.

Bridge of Night: Originally I switched the order of the last 2 diamonds and did not have an edge hit. The path tweak in this video saves a bit over a second.

A Change of Friction: Apparently jumping on the slope right before the finish will cause me to go too far left to hit the finish. The wall-hit near the end saves me from having to slow down after the previous slope, which I often did not slow down enough on.

Detours: I think there was an edge hit in this run at the maze part. Normally I do not get enough height after the pipe bounce to reach the finish.

The Enjoyable: I haven’t played Tightropes in years, but it seems like my tightrope skills have improved in the meantime. The first Super Jump is quite difficult—it is easy to not get enough height. Also on most of my runs, I would miss one of the Super Jumps in the second set of double Super Jumps for some weird reason.

Marble Speed: Two paths for a quick time. The second path is possible without the lucky edge hit at the end, you have to do the right jumps before then. I got a 4.13 or 4.14 using that path but did not get it again while recording.

Quick-Path 10: This is a path that uses the Gyrocopter twice.

A Random Level: Pascal’s advice to me years ago about how letting the marble bounce without pressing any direction keys will stop skidding really helped me in this level.

Strain: It’s possible to hit the edge of the bouncy pads to reach the Jump Boost very quickly, but that’s hard to do and I didn’t get it enough times to learn exactly how to control the marble after the edge hit.

123zc1’s first constructor level!: It’s possible to bounce at a different place on the bouncy floor to activate the Jump Boost without hitting the platform it’s on. Then one will get Jump Boosted just above the next platform (right before the next diamond).

(totally not late) Lovely video, I loved the paths you came up with. Bridge of Night is probably my favorite of the bunch.

No Gems: The Second Compilation — Update 3



This update includes a good helping of MBP Advanced and Expert levels. It was originally scheduled for late 2013 or even slightly earlier but got pushed back. I do have plans for a 4th update but that’s much further in the future.

Notes:

Spork in the Road: The 3.40 path modified so that it will work with a start pad trick. Notice that the marble cuts the corner where the flat part meets the slope, and it doesn’t land on the slope until it’s about halfway up. The 3.40 path doesn’t cut the corner. It’s possible to cut the corner and still make it to the Time Travels without the start pad trick, but it’s a lot more difficult.

Three-Fold Maze: Back in 2009 or so, Sonic or KidPlayer44 (can’t remember who) used a path using no Gravity Modifiers. I used the Gravity Modifier path for NGC1 and NGC2 because that path was faster for me (and I got better times than what they did). Recently I tried the non-GM path and if it is done right, it is faster than the GM path.

Twisting the night away: Pascal’s path isn’t quite as difficult as I thought it was. In practice runs, I tried to get before the sequence of moving blocks at the end (I’m sure he tried too), but I wasn’t early enough.

Learn the Wall-Hit: The start pad trick again.

Slippery Steps: Every time I try this path I’m amazed at how much time I can chop off my previous run of this level (NGC1: 3.86, NGC2: 3.59).

Frictional Ascent: My NGC2 run was slow on the landing after the loop exit, which this run has fixed.

Morph: You can just press the keys right after you jump on the square, it will work (although sometimes you won’t go high enough).

Rolling to Eternity: In my NGC2 run, the path contains certain places where jumps need to be made, which means that one cannot accelerate as much as one would like to. This run is an (not completely successful) attempt to minimize these slowdowns.

Trigonometry: I got the Jump Boost earlier compared to my NGC2 run. It’s possible to go from the second trigger to the finish essentially in a straight line (so very quickly), but then one runs the risk of making it *too* early or the line might be in the wrong direction.

Catwalks: This path is much easier than my 32 second path that I unsuccessfully tried for NGC2. However, the 32 second path does not use the edge hit on the stairs (since the other edge hit seems almost impossible) and so theoretically it is still better.

The Ultimate Friction Challenge: I added an extra edge hit and caught the 0.5 second TM that I missed in NGC2.

Mastering the Marble: My path plus Pascal’s start.

Songs: Guess where they’re from….

Wonderful compilation, great paths. Catwalks edge hit was insane!

Marble Blast Levels: Compilation 31



So earlier this year I provided a roadmap for the next 3 Marble Blast Levels videos:


MBL: C31: Long levels

MBL: C32: MBG/MBP regular levels

MBL: C33: Old levels



This plan has changed slightly and now C31 contains MBG/MBP regular levels. The other two will still happen, don't worry.

This video features many paths and tweaks which I had never attempted before in a serious run.

You can see that the video is in native 1080p, this resolution is the highest I have ever used in a Marble Blast video (previously it was 1440x900 game in 1728x1080 export).



Notes:



Take the High Road: The second gem was fairly tricky to reach in a good time.



Upward Spiral: I didn't realize the importance of a good edge hit at the start until now. The one in this video isn't optimal, but it's as good as or better than most other ones I got.

As for the bloopers, I recall Pascal posting on the forums some time ago that he got a 9.47 or so TT catch.

Money Tree: A long time ago I got a 41.0x and this was my successful attempt to beat that time.

Stepping Stones: None.

Twisting the night away: The ending was better than what I expected.

Siege: This path has been a bit of a pain, for years I could do the path normally but I would invariably make some mistake or other that led to higher times than what I wanted. This is the first run where I did everything "correctly" plus the time saver after the first gem.

Mudslide: Oo Wo0dY oO's path in MBG.

Ordeal: My PR before this run was probably a 9.73, although I might have gotten another 9.67 in the past.

Tightrope: Remember in my notes for MBL: C30 when I said "I haven't played Tightropes in years, but it seems like my tightrope skills have improved in the meantime"? Before the preparations for MBL: C31, my best time in this level was something around 35 seconds. Back then I also did not jump on the third tightrope. I switched camera control from the keyboard to the mouse for the fourth tightrope, and then back again afterwards.

Natural Selection: I tried to reduce the time it took to get to the moving platform.

Darwin's Dilemma: The 6.xx path (which is "my" path plus only one path modification near the end) was actually not that difficult after all this time. I also made sure I did an edge hit, which makes sense because even my first video of a shortcut path for this level featured an edge hit.

Busy Bee: I remember when I struggled to get Platinum Time (15.00) on this level.

Bumpy Highway: My previous best time was a 14.9x I think, so I was just trying to optimize my run.

Songs:
My Chemical Romance - Jet-Star and the Kobra Kid / Traffic Report
How to Train Your Dragon 2 - Where No One Goes
Sia feat. The Weeknd and Diplo - Elastic Heart

So many perfect edge hits! Ordeal was just... oh

Amazing vid!!!! So awesome you are uploading again

Marble Blast Levels: 32



My plan was to have MBL: C32 and MBL: C33 be of old levels and long levels respectively, but plans changed and now the new Marble Blast Levels: 32 (more concise name) combines both of those planned videos into one long video.

I have known each of these levels from the 2008-2010 timeframe (pre-MBLX) in one way or another, and most of them I have not played seriously from that timeframe until when I started making this video. Thus, I will list, beside each level name in the notes, the estimated top time I achieved in the level in those days, if a reasonable estimate can be made.

The marble skin is representative of this video. It is the "x^x spindle" (Google it) which I saw in around 2004 and forgot its name. Thus I couldn't look it up anywhere, and in time I wasn't sure that I had even seen it properly the first time. But one day about a month ago I was looking at some links and a paper called "The x^x spindle" caught my eye. At that point I knew that I had found it again. This skin was made in the middle of making this video, but I recorded all .rec's using this skin. It is also not the Marble skin 2014-2015, that's a yet-to-be-made skin.

Alleyway Onslaught (?): It was this level that taught me how to jump up from one narrow protrusion to a higher one (in this case, the ladder at the start). Later on I would use the same skills on levels such as Stairmaster. The irony is that now I use a path (thanks to Imperial) that doesn't require this type of jumping. In the old days I could not go from the end Gyrocopter to the finish pad with no extra bounce in between, but I can do it now.

Block Chaos (~55): The last part has lots of Gyrocopters so I looked for any type of shortcut I could do. I thought of some involving difficult edge hits on the space catwalk but then I found this ending path, which is easier and faster.

Boxed In (31-32): I could never do a proper edge hit back in the day, and in my runs of this level I never did one. Even now, a good edge hit is surprisingly hard, as a high edge hit wastes time and a low edge hit is usually too fast for me to move the marble right (or left) towards the Super Jump in time to collect it.

The Circus (30-31 in mid-2012): This level's description stuck in my head for years (I think I played it in the old days but not seriously). It was considered for MBL20 but I didn't put it in because I didn't get a particularly good time then. I liked the path I used though, and I quickly found it again and improved it for this video.

Flipsie Doodles (6.04?): Kind of an annoying little level but a good run makes it worth it. I'm not sure if I'm remembering my old top time correctly because I'm not sure if I used the start pad tweak back then and I couldn't have gotten such a time without that tweak. In any case I'm pretty sure that I got sub-7, and I remember getting some very good bounces on that bumper.

Gem Stack (~23): Back in the old days, I used a lot more jumps than in today's run. Also I often used the Super Jump, but in a different point in the level.

Grasp (14-15): I don't remember if I got sub-15 or not many years ago, but in any case I should have lowered my PR by at least 1 second now.

Group Level (10.1x-10.2x): This run uses a different Super Jump pattern than the one I used in the old days, and in this run I pressed diagonal left the entire time after the first Super Jumps.

Hall Of Fame (~7.10): I got a 7.10 on this level in around April/May 2009 and I told Sonic about it in his "MBG: Random Custom Levels" video. Sonic replied that he got sub-7 "two days ago." Unfortunately for me, I wasn't able to figure out the path I accidentally performed to get my 7.10. I decided to try this level again for MBL: 32, and soon enough I found the path and joined the sub-7 club.

Incognitus III: The Final Chapter (~35): My best times used to be in the 35-40 seconds timeframe, but with additional tweaks I was able to go sub-25. The marble went too close to the edge / finish pad before and after going in that little hole for the 5th gem, wasting a bit of time, but it's minor given that the collection of the last 2 gems and finish was very well done.

Looks are Deceiving (~55): Well, I have a lot of shortcuts now, including one usage of getting to the finish soon enough after going out of bounds such that the game doesn't restart. It is possible to finish the level in like 4 seconds by Ctrl + P usage, but I didn't know of that tweak back then so I did not use it now (I knew of the OOB tweak though). Also it's more fun to do it the "long" way. The Super Speed part is seemingly a bit random, sometimes one will go high or low. This run has a higher Super Speeding than what I would have liked, but I often mess up on the TT catches afterward, which did not happen in this run. Also, it might be possible to Super Speed directly to the blue part above the Shock Absorber, but I didn't try it.

MBU Tower (30-31): This was also a level I considered for MBL20. I think my best time was around the 30 second mark but in the 2008-2009 timeframe it was probably more like 35 seconds. There seems to be a lot of shortcuts for this level but most of them don't seem to work or are annoying. I found one that gives a very low time but it is still annoying.

Mission Impossible (N/A - I didn't play it in the old days): The long level of this video. I watched Matan's video of this in the MBG Movie many times (mostly due to the song ) and gradually I became aware of possible improvements to the path. Eventually I put these tweaks into place a few months ago and got a decent time. I improved my run further and the result is what you see here.

Mkbul's Skill Course (don't know): I don't remember when I started doing the "reverse" path for the middle of the level. But before that, you will see the tweak that is the reason why I put this level in this video.

One fold battlecube ( > 30 I think): Did you know that I didn't get Marble Blast Platinum until around September 2008, about 5 months after I started to play Marble Blast Gold again? At that time I had already heard of the Battlecube levels in MBP, so I played this level in MBG as a way to play a Battlecube level without MBP.

Pyramid (3x): The first edge hit is not intentional, I just went with it. After the second edge hit I was fairly cautious for the rest of the run. The last two edge hits are tricky because the edge that you hit is lower than the surface the marble rolls on, which is not the case with other edge hits of this form. Thus I needed to adjust my marble movement so I roll off the surface earlier than usual.

Mbg's Race Against Time (20-21): I completely forgot about this level for years, it was also one of the last levels I decided to add to this video. For some strange reason I find it very difficult to Gyrocopter up to just above the gem platform, I almost always go much higher. Thus in most cases it is actually faster to hit the side of the platform, which reduces height.

Turbo4000's Race Against Time (above 20): Notice that I get the rear Super Speed before getting the gem. That wasn't optional in this particular run due to the angle of the edge hit + Super Speed. Since the rear Super Speed and gem are "behind" the edge, any run that catches the gem first either has a slowdown after the gem catch or requires a higher Super Speeding (which also slows the run down) to catch the gem, not fall off, and move forward to collect the Super Speed. This edge hit + Super Speed gives one of the lowest heights in my attempts, and if it were at an angle that would collect the gem first, some slowdown would occur. So while catching the gem first is almost certainly faster overall, the path I used in this run is not as much slower as one might initially be led to believe.

Razac (~30?): I didn't like that "staircase" to the top Gyrocopter back then and now, and I think that in the old days I found some path that skipped the staircase.

Rooms 2 (8.0x-8.5x): A nice little level that I played often in 2008-2009, although I get the feeling that the black stuff is some special surface which I don't (and didn't) have. The difference today is that I don't land on the part with the gem, saving a second of time.

Rotational Hallways (~8.0): The ending always messed me up back then and it still does now. I just couldn't get to the finish soon enough after getting on that ending incline. This run was very lucky in that I made probably the best ending in all my attempts.

Simple Course (2.3x lol that's pretty much all I remember): If you don't make correct moves after the TT catch, you risk the time restarting before you reach the finish.

Sky High (something really large): I didn't know about the double Super Jump when I first watched Matan's video, so I was always amazed at the boost he did on his second run in that video. Fast forward some years and that technique is common in my runs. The ending wasn't quite satisfactory, but I would often be hit by the Gravity Modifier on the last platform so better safe than sorry.

Slalom (20 something): I thought of this path 4 years ago and actually mentioned it on Matan's Slalom 1-4 video. At that time I had no idea if it was possible or if it saved time (although if it were possible I thought it would save a bit of time over the WR path), since I couldn't properly pull it off. I tried it again for this video, and I got a bunch of failed attempts before this lucky 16.15 (you can tell I was surprised; I misjudged the movement for the wall at 12 seconds) which was basically the path I was thinking of. The path for the second run was also by surprise, I thought it was also possible but I also thought it would be really hard to jump on all the tops of the walls. It's actually not that hard once you get the first few down.

Speed Demon (27): I really like the ending. I made sure that I got all 4 "groups" of TTs at the base of the tube and still have time to jump to the finish.

Speed Gems (6): I played this level a bunch of times in the old days and now I further tweaked my run and optimized the time.

Spiral (5.0x): I wanted to get sub-5 on this level back then, I don't think I ever got it. This level taught me the "jump around a wall" trick.

Super Castle (?): The level description made an impression on me. I'm not sure when I first played it though. You have to be careful on this level, there are many places where your marble can stick to a wall and screw up your run! Usually the time restarts by around 0.15-0.2 seconds at the end, but for some reason that didn't happen at all in this run (9.81/10.00).

Super Simple Level 3 (~7.5): This path is harder than it looks because if you go too fast on the green stuff, you go too far to the left with no way of getting the marble to the finish.

Trapgems (27-28): I couldn't finish with just one bounce on the platform with the last gem.

Tricky Trapdoors (~40): I didn't even need the Super Jump at the end, unlike in those practice runs where I would bounce 2 or more times on the trapdoors after the Gyrocopter returned me to normal gravity.

Up the Mountain (~12.3-12.5, probably 11.xx later): The edge hit is inspired by the MBU paths for Pitfalls and Winding Road.

Upward Slants (~21): Many years ago I tried to find a Super Speed path similar to that in the old King of the Mountain path. While I didn't succeed back then, I did succeed in finding a path now.

U-Tube (~50, maybe even 1 minute or more): Thanks to Matan for the path. I added in some tweaks. The path is much better than the normal path that takes 50 seconds or so.

Wave Jumping (10.2x): It is actually possible to get in the second block from the corner. It was tricky when I first tried it but it became much easier once I got the hang of it. I think the launch is slightly slower than the third block but it is offset by the earlier launch. (I have 7.8x using the third block.)

I hope you enjoy!

Songs: Nightwish - She is My Sin, Nightwish - Rest Calm, Nightwish - Scaretale.

Video of epic proportions. Should have saved it for the last day of the year because that is just epic awesome. It's great seeing so many old customs levels being destroyed, and the edge hits are insane one by one. Please continue, you're amazing.