Serious Sam: The First Encounter

Edit history:

SDAVerification: 2020-06-02 01:42:02 pm

SDAVerification: 2020-06-02 01:41:32 pm

2020-05-16 01:06:56 am

Game Page: http://speeddemosarchive.com/SeriousSamFE.html

Serious Sam: The First Encounter (pc) (pc) [Any %] [Segmented] [Normal]

Decision: Cancelled

Reason: The runner is working on an improvement, submission cancelled

https://queue.speeddemosarchive.com/verificationfiles/2264/

This run will be available for a month. After that these link(s) will no longer work.

Thread title:

SDAVerification

2020-05-16 01:06:56 am

Run Information

Serious Sam: The First Encounter (pc) (pc) [Any %] [Segmented] [Normal]

Verification Files

http://v.speeddemosarchive.com/SeriousSamFE-20200514/

Please refer to the Verification Guidelines before posting.

Please post your opinions about the run and be certain to conclude your post with a verdict (Accept/Reject). If you wish to remain anonymous, you can also send a pm with your reply to 'sdaverification' (please state clearly in that case which run you have verified). This is not a contest where the majority wins - Each verification will be judged on its content.

LotBlind

2020-05-25 04:01:50 pm

Right, so that old run should never have been accepted at all. There were known improvements even back then, as I understand. Also if it only segmented 10 times, that was less than once per level... but this is still kind of the main thing with this run as well: we like it when segmentation usage is taken a bit further than this. What's the point of making a segmented run just a little bit better than ILs? Since there isn't an inherent limit on how often you can segment, we don't consider a less segmented run better than a more segmented run, we just want to see how far a game can be pushed (in real-time). Think about it from the viewers' perspective: the number of segments is just a number, but the more little optimizations, the more impressive the run looks in a very tangible way.

In this run there are definitely moments that look like they could have been even faster. Rocket jumps that don't always quite land exactly where they optimally should have, it would probably have been possible to manipulate the charging bull-thing more efficiently to get the boost at 3:38 even sooner... here's a few observations or ideas.

4:33 - Can this RJ be done in such a way as to vault the ridge completely? Looks like you get slowed down when you land a little below the edge.
4:58 - Really cool that you can actually pick up items from OOB. That isn't often the case.
Dunes - Based on your comments, more segments hear could have enabled getting those less damaging Werebull boosts in case that allows skipping a pick-up (or just doing more of them), and of course making them boost you in exactly the right direction each time.
9:39 - Here a slightly smaller boost could have gotten you over the gate, couldn't it? Wouldn't this have saved a very small amount of time spent picking up one of the two +25 potions afterwards? Possibly applies to some other boosts (vertical ones) throughout the run.
10:10 - Obviously not getting here is possible, right? Actually, might it be possible to combine a RJ with the bull butting you to gain horizontal speed when escaping the trench? Of course any boost that sends you airborne is best done down a hill since it increases the air time, barring too much fall damage, but this game doesn't look to have that.
15:48 - Again, with another segment here, you could probably save at least half a second by using the clock to time when the last guys will spawn and prefiring the shot. I'm assuming there's no randomness to the timer. Is it impossible in theory to kill them faster than the about 7 seconds that it takes you to fire the cannonball and get to the door, if you waited at the door instead and just shot them one by one?
23:10 - When you go pick up the super health here (it gives you 100 hp, right?), looks like you were already close to 200 hp. Can any of that extra health be used to jump around the arena usefully during the fighting? If you collect the armor after the health? In any case, if you're using a projectile weapon to get the last kills to finish off a wave (or a boss of course), you should be as close as possible to the last enemy that you shoot when you fire at them.
25:20 - I think there might be opportunities to use the cannonballs in a slightly more versatile way. They bounce off walls, after all. It looks like, unless they deal a lot less damage after bouncing, you could have fired the first of the two cannonballs that kill the last two bulls ahead of time so it kills it on the rebound, leaving you to just prefire the other bull down directly. A clear time saver AFAICS. Was there no opportunities to make use of the large radius when it blows up? Could one such cannonball kill both bulls?
27:00 - When you do the RJ after the one that you combine with the launch pad, does that effectively save more time than just doing a random RJ without another kind of boost prior? If so, I'm wondering if there wasn't somewhere else where you could have done that right off a bull boost or something like that. Is a bull + rocket boost better than bull boost then rocket boost, btw?
28:20 - Whoa, this double kill can't be on purpose can it? O_o
29:10 - Is it technically a waste of time to have rockets blow up on approaching enemies instead of flying over to the boss? I feel there has got to be more optimizations in this boss fight with more segments.

Obviously arena fights can always be done even faster even by segmenting before the final group of a wave has spawned.

I'm assuming this game doesn't have the kind of "zig-zagging" that I've seen in one of the other SS games.

Does your momentum stop if you jump while running and release movement keys? Because if it doesn't, another minor optimization would be anytime you need to turn around to shoot at something, you first jump, then turn mid-air, then resume moving to stop there being those frames when you're not moving at full speed. Doesn't cost a thing, in case this can be done that way.

So yeah, I like this run for what it is but a low segment count often means a run runs afoul of the elemental "can't easily be improved" principle. I do like the beelining which I've seen done much more sloppily, and it's clear from your comments many alternatives have been tested throughout the run. Do you have any interest in redoing the run with let's say 3-4 times more segments (at least)?

Right, so that old run should never have been accepted at all. There were known improvements even back then, as I understand. Also if it only segmented 10 times, that was less than once per level... but this is still kind of the main thing with this run as well: we like it when segmentation usage is taken a bit further than this. What's the point of making a segmented run just a little bit better than ILs? Since there isn't an inherent limit on how often you can segment, we don't consider a less segmented run better than a more segmented run, we just want to see how far a game can be pushed (in real-time). Think about it from the viewers' perspective: the number of segments is just a number, but the more little optimizations, the more impressive the run looks in a very tangible way.

In this run there are definitely moments that look like they could have been even faster. Rocket jumps that don't always quite land exactly where they optimally should have, it would probably have been possible to manipulate the charging bull-thing more efficiently to get the boost at 3:38 even sooner... here's a few observations or ideas.

4:33 - Can this RJ be done in such a way as to vault the ridge completely? Looks like you get slowed down when you land a little below the edge.
4:58 - Really cool that you can actually pick up items from OOB. That isn't often the case.
Dunes - Based on your comments, more segments hear could have enabled getting those less damaging Werebull boosts in case that allows skipping a pick-up (or just doing more of them), and of course making them boost you in exactly the right direction each time.
9:39 - Here a slightly smaller boost could have gotten you over the gate, couldn't it? Wouldn't this have saved a very small amount of time spent picking up one of the two +25 potions afterwards? Possibly applies to some other boosts (vertical ones) throughout the run.
10:10 - Obviously not getting here is possible, right? Actually, might it be possible to combine a RJ with the bull butting you to gain horizontal speed when escaping the trench? Of course any boost that sends you airborne is best done down a hill since it increases the air time, barring too much fall damage, but this game doesn't look to have that.
15:48 - Again, with another segment here, you could probably save at least half a second by using the clock to time when the last guys will spawn and prefiring the shot. I'm assuming there's no randomness to the timer. Is it impossible in theory to kill them faster than the about 7 seconds that it takes you to fire the cannonball and get to the door, if you waited at the door instead and just shot them one by one?
23:10 - When you go pick up the super health here (it gives you 100 hp, right?), looks like you were already close to 200 hp. Can any of that extra health be used to jump around the arena usefully during the fighting? If you collect the armor after the health? In any case, if you're using a projectile weapon to get the last kills to finish off a wave (or a boss of course), you should be as close as possible to the last enemy that you shoot when you fire at them.
25:20 - I think there might be opportunities to use the cannonballs in a slightly more versatile way. They bounce off walls, after all. It looks like, unless they deal a lot less damage after bouncing, you could have fired the first of the two cannonballs that kill the last two bulls ahead of time so it kills it on the rebound, leaving you to just prefire the other bull down directly. A clear time saver AFAICS. Was there no opportunities to make use of the large radius when it blows up? Could one such cannonball kill both bulls?
27:00 - When you do the RJ after the one that you combine with the launch pad, does that effectively save more time than just doing a random RJ without another kind of boost prior? If so, I'm wondering if there wasn't somewhere else where you could have done that right off a bull boost or something like that. Is a bull + rocket boost better than bull boost then rocket boost, btw?
28:20 - Whoa, this double kill can't be on purpose can it? O_o
29:10 - Is it technically a waste of time to have rockets blow up on approaching enemies instead of flying over to the boss? I feel there has got to be more optimizations in this boss fight with more segments.

Obviously arena fights can always be done even faster even by segmenting before the final group of a wave has spawned.

I'm assuming this game doesn't have the kind of "zig-zagging" that I've seen in one of the other SS games.

Does your momentum stop if you jump while running and release movement keys? Because if it doesn't, another minor optimization would be anytime you need to turn around to shoot at something, you first jump, then turn mid-air, then resume moving to stop there being those frames when you're not moving at full speed. Doesn't cost a thing, in case this can be done that way.

So yeah, I like this run for what it is but a low segment count often means a run runs afoul of the elemental "can't easily be improved" principle. I do like the beelining which I've seen done much more sloppily, and it's clear from your comments many alternatives have been tested throughout the run. Do you have any interest in redoing the run with let's say 3-4 times more segments (at least)?

Blacksecret

2020-05-26 12:46:09 pm

Looking back at the run doing the levels single segment did make some parts of the longer levels slower than they could have been since initially this run was meant to show me what an optimised full game run would look like. So to answer your last question yes I'm interested especially because figured out some optimisations for the last level that save quite a bit of time. The only issue with making a run with many segments is that sometimes loading a save causes the camera to face straight down for some reason but I think I have a way to hopefully avoid that.

3:38 - I would have to mess around with that and extra segments would definately be handy for that
4:33 - It's something I didn't really pay attention to but I think yes
4:58 - You might think it's obvious that saving more health and armor would allow you to skip the 25 armor in the next level and add a boost somewhere but there is definately a tradeoff of a weaker rocketboost off of the bull and regular rocketboosts while you have armor left save almost no time at all so it's questionable if this part could be improved or not although segmenting more would let me improve the boosts from the bulls at the very least.
9:39 - picking up the medkit as soon as possible is the top priority here since walking that far triggers the enemy spawns that last for a bit over 2 minutes which you can't speed up and you have enough time to grab the extra +25 and I think I even got hit by a bull but it didnt really lose me any time there since I still managed to kill the last kamikaze just after it spawned. I agree that it probably didn't look very optimal but it only-- made it a bit harder for me to kill the last enemy on time
10:10 - same as previous it's possible but it really isn't necessary + I'm not sure if you can do that without losing over 50 hp
15:48 - Prefiring would save some minimal time but waiting at the door would take longer because there's atleast 20 enemies and they have too much hp to kill them fast enough one by one my only other idea would be prefiring very early and hoping most of them would die from the returning cannonball maybe saving some time that way.
23:10 - You can get the same effect by prefiring efficiently in which case segmenting will help a lot so there is no real reason to jump around the arena
25:20 - Sadly that won't work because the cannonballs don''t do enough damage to kill a bull after bouncing off a wall but there is a slight delay of the door opening after killing the last bull so maybe there is an angle that allows killing them both in one shot while reaching the door on time.
27:00 - Bull + rocket is definately better than bull boost then rocket upon landing. the reason why the boost after the launchpad saves more time that a regular boost is that you already have momentum which gets added to the rocketboost but you cant really carry over the speed you get from a regular bull boost this way since you have slowed down too much by the time you land. I also believe it would be inneficient to do rocket + bull boost and then a rocket upon landing since you would be too far from any bull after that.
28:20 - Yeah that was just a cool coincidence.
29:10 - The Boss heals back to almost full if you do too much damage to him and I could have done maybe 2 more cannon hits on him before the laser hit to trigger his healing so I don't think the rockets wasted time especially since they save me from taking damage. I did lose a little bit of time because I started to use the cannon on the boss at the very end a little bit too late.

There are also some new optimisations that let you do more boosts and skip some health in the last level.

Yes the zigzagging is a HD rerelese thing and honestly it's an absolute pain.

There is a second or two of time during which you have air control after leaving the ground and releasing the movement keys during that time stops your momentum so you could only turn to fire at things during rocketjumps etc. but not with regular jumps.

Thanks for the great feedback there's definately a few more areas in the run that would benefit from more segmentation and hopefully I will have time to do this soon since I'm pretty busy with work although many parts should take less time with more segments.

Looking back at the run doing the levels single segment did make some parts of the longer levels slower than they could have been since initially this run was meant to show me what an optimised full game run would look like. So to answer your last question yes I'm interested especially because figured out some optimisations for the last level that save quite a bit of time. The only issue with making a run with many segments is that sometimes loading a save causes the camera to face straight down for some reason but I think I have a way to hopefully avoid that.

3:38 - I would have to mess around with that and extra segments would definately be handy for that
4:33 - It's something I didn't really pay attention to but I think yes
4:58 - You might think it's obvious that saving more health and armor would allow you to skip the 25 armor in the next level and add a boost somewhere but there is definately a tradeoff of a weaker rocketboost off of the bull and regular rocketboosts while you have armor left save almost no time at all so it's questionable if this part could be improved or not although segmenting more would let me improve the boosts from the bulls at the very least.
9:39 - picking up the medkit as soon as possible is the top priority here since walking that far triggers the enemy spawns that last for a bit over 2 minutes which you can't speed up and you have enough time to grab the extra +25 and I think I even got hit by a bull but it didnt really lose me any time there since I still managed to kill the last kamikaze just after it spawned. I agree that it probably didn't look very optimal but it only-- made it a bit harder for me to kill the last enemy on time
10:10 - same as previous it's possible but it really isn't necessary + I'm not sure if you can do that without losing over 50 hp 
15:48 - Prefiring would save some minimal time but waiting at the door would take longer because there's atleast 20 enemies and they have too much hp to kill them fast enough one by one my only other idea would be prefiring very early and hoping most of them would die from the returning cannonball maybe saving some time that way.
23:10 - You can get the same effect by prefiring efficiently in which case segmenting will help a lot so there is no real reason to jump around the arena
25:20 - Sadly that won't work because the cannonballs don''t do enough damage to kill a bull after bouncing off a wall but there is a slight delay of the door opening after killing the last bull so maybe there is an angle that allows killing them both in one shot while reaching the door on time.
27:00 - Bull + rocket is definately better than bull boost then rocket upon landing. the reason why the boost after the launchpad saves more time that a regular boost is that you already have momentum which gets added to the rocketboost but you cant really carry over the speed you get from a regular bull boost this way since you have slowed down too much by the time you land. I also believe it would be inneficient to do rocket + bull boost and then a rocket upon landing since you would be too far from any bull after that.
28:20 - Yeah that was just a cool coincidence.
29:10 - The Boss heals back to almost full if you do too much damage to him and I could have done maybe 2 more cannon hits on him before the laser hit to trigger his healing so I don't think the rockets wasted time especially since they save me from taking damage. I did lose a little bit of time because I started to use the cannon on the boss at the very end a little bit too late.

There are also some new optimisations that let you do more boosts and skip some health in the last level.

Yes the zigzagging is a HD rerelese thing and honestly it's an absolute pain.

There is a second or two of time during which you have air control after leaving the ground and releasing the movement keys during that time stops your momentum so you could only turn to fire at things during rocketjumps etc. but not with regular jumps.

Thanks for the great feedback there's definately a few more areas in the run that would benefit from more segmentation and hopefully I will have time to do this soon since I'm pretty busy with work although many parts should take less time with more segments.

LotBlind

2020-05-27 01:26:09 pm

The camera turning to face down when reloading: what kind of problem do you think this might be? It can't be helped but I don't think it's a major minus for the viewing experience.

Yeah, I really think going totally ham on segmented runs is what gives them their appeal. Glad you find it the same way. Some runners don't want to go out of their way for very minor optimizations but I think those add the spice into it. Furthermore, the more optimized a run is, the greater the chances the next runner will do a different category or game instead of just an improvement on an old run, which means more runs overall!

9:39 - I didn't expect all my feedback to be on the mark. Since you want to get to the health pickup in the arena fast, I guess if it's possible to get more horizontal speed from the boost, that would still make getting there faster, unless I've misunderstood this.

15:48 - Actually, I had the same idea afterwards. You should definitely try whether a richocheing cannonball, or even multiple, could be used here. One could be rolling, one could be about to explode, and the third one could the one you fire directly at them. Maybe it's not worth spending that many cannonballs?

23:10 - You're right, prefiring might take care of that. I'm trying to imagine how you could prefire e.g. both of the scorpions that appear on top of the pillars by the gate so you kill them at the same time (in case those were some of the ones that had to die fast). At the least, with cannonballs, couldn't you send one flying in a high arc and fire the next one directly at the other one?

25:20 - I see. What about firing the first cannonball in a high arc here actually? Or what if a cannonball explodes in-between them and you also fire a rocket between them. Would that combine into enough damage?

27:00 - Good to know you've considered this.

29:10 - I see. That sounds like something that you might want to mention in your comments.

Getting all this stuff out of the way will make the next verification easy. I have no doubt you'll be able to produce a more than sufficiently high-quality run just by adding those segments and including all the improvements you already know about for the next verification whenever you've found the time. Sounds like it won't be submitted right away, so this submission will probably be cancelled completely.

The camera turning to face down when reloading: what kind of problem do you think this might be? It can't be helped but I don't think it's a major minus for the viewing experience.

Yeah, I really think going totally ham on segmented runs is what gives them their appeal. Glad you find it the same way. Some runners don't want to go out of their way for very minor optimizations but I think those add the spice into it. Furthermore, the more optimized a run is, the greater the chances the next runner will do a different category or game instead of just an improvement on an old run, which means more runs overall!

9:39 - I didn't expect all my feedback to be on the mark. Since you want to get to the health pickup in the arena fast, I guess if it's possible to get more horizontal speed from the boost, that would still make getting there faster, unless I've misunderstood this.

15:48 - Actually, I had the same idea afterwards. You should definitely try whether a richocheing cannonball, or even multiple, could be used here. One could be rolling, one could be about to explode, and the third one could the one you fire directly at them. Maybe it's not worth spending that many cannonballs?

23:10 - You're right, prefiring might take care of that. I'm trying to imagine how you could prefire e.g. both of the scorpions that appear on top of the pillars by the gate so you kill them at the same time (in case those were some of the ones that had to die fast). At the least, with cannonballs, couldn't you send one flying in a high arc and fire the next one directly at the other one?

25:20 - I see. What about firing the first cannonball in a high arc here actually? Or what if a cannonball explodes in-between them and you also fire a rocket between them. Would that combine into enough damage?

27:00 - Good to know you've considered this.

29:10 - I see. That sounds like something that you might want to mention in your comments.

Getting all this stuff out of the way will make the next verification easy. I have no doubt you'll be able to produce a more than sufficiently high-quality run just by adding those segments and including all the improvements you already know about for the next verification whenever you've found the time. Sounds like it won't be submitted right away, so this submission will probably be cancelled completely.

Blacksecret

2020-05-28 02:20:31 pm

9:39 - The more damage you do to yourself the stronger the boost will be and you don't really keep the speed during the whole flight so from what I understand firing under you late into a jump will end up being faster than turning around to attempt a more horizontal boost.
15:48 - The cannonball explosions are really small and weak but I think 2 rolling ones might work and you have enough to do that.
23:10 - I will be very suprised if a falling cannonball does enough damage to kill one of the scorpions because I'm pretty sure fully charged cannon shots explode in the air before they fall back down
25:20 - I think I can drop enough grenades to kill one bull and cannon the other one.

And about the camera I just thought it would look bad but hopefully I can avoid the problem.

LotBlind

2020-05-29 08:03:31 am

9:39 - Right, because of the air resistance effect. I see.

23:10 - Another idea I had was what happens if you shoot a cannonball so it clips the edge of one of those pillars? Depending on the physics, would this cause it to ricochet at a near-vertical angle so it falls back down on the scorpion before it's exploded?

Is there somewhere where you can shoot two cannonballs early and complement one of them with some kind of hitscan fire to enable a fast double kill that way? I'm just trying to be thorough with my suggestions so as many tricks can be packed into the next submission as possible. By the way, is there anywhere where you pick up health without you needing quite that much to be at full health? If so, maybe a small rocket boost before then?

Blacksecret

2020-05-29 01:13:05 pm

23:10 - I will test what you can do with the cannon when I get to that point

I don't think I ever grab too much health like that and i will have to check if the cannonballs will even do enough damage to be useful that way.

I'm 3 levels in so far and I managed to save 4 seconds on the third one which was quite a surprise to me so far so good.

LotBlind

2020-05-31 12:06:22 pm

Obviously a rocket -> hitscan combo could be better. Anyway, I'll wait for the next submission now before returning to this.

SDAVerification

2020-06-02 01:41:33 pm

Decision posted.