While segmenting a lot does in a way make the run “cheaper”, I think it’s really silly to not do certain routes because it requires an extra segment. The point is to go fast and not restrict yourself to an arbitrary amount of segments. I don’t think using timescale would be really worth it in old engine since it can often lead to jittery mouse movement and it wouldn’t even save that much time.
What do you mean by “within” the game? The game allows you to save whenever you wish so why would you not use that if it allows you to go faster?
Well I understand that, and honestly, before I joined the HL2DQ team I was setting out to do a TAS of HL2 and many of these tricks were going to be employed. I enjoy TAS runs; it’s fun to see a lot of impossible tricks and 100% efficiency. But I have a feeling this run isn’t going to label itself as such. The segmenting policy is a lot easier if you cringe every time you make one (that isn’t a level change). If I see a sweet turnstile jump and then I find it took a segment to do it, the entire effect of “omg this guy is amazing!” is completely lost. The more segments are created, the more accessible the run is for anyone to complete without much practice.
As far as save/loading goes, these aren’t save points, so it just feels very cheap. I was expecting that man-hack video to be some sort of crazy clip, but instead all i see is Save… Load… what? To me, when you load a game, you have left it. If you end up in the same situation you saved it in, great, game continued, but if you are advanced from where you were when you were still playing, you are not loading the right game. That’s what I mean when I say it feels like you are going “outside” the game.
EDIT: To Jukspa: Actually, the point of speedrunning is in fact to limit yourself to a certain number of segments. This is why in SDA they adopted the 0.5 seconds lost for every segment rule, which yes is in fact arbitrary and in my opinion not strict enough, but the reason this was adopted is this: segments = BAD. The more segments you have, the closer you are to having a TAS. The difference between a TAS and a speedrun is skill. I was a little ashamed to end up posting our run with 21 additional segments. We made a few sacrifices for the sake of completing the run and a few awesome shortcuts. But even then, the segments were limited, and avoided as often as possible. d1_canals_12 took forever to complete and LightningX submitted a 2 segment run of it insisting we use it for the sake of time. I refused and ran that map for a very long time and ended up beating his 2 segment time with just 1. 1 segment per level is CLEAN, because the game forces you to autosave and break your motion at those points anyway. 2 segments is DIRTY, more than 2 is filthy, and so on and so forth.
SDA recently dropped the segmentation penalty if you aren’t aware. If you’re looking to display skill and more segments = bad then why not do a single segment run at that point? The routes are pretty similar(if you don’t do any save glitching) the only difference is that a segmented run can obviously be much more optimized.
The whole one segment per level thing seems pretty silly. Say Valve made the maps twice as big, so that two maps were combined, would it be “dirty” to use two segments per map in that case then? Or if they were more split up, would it be okay to do one segment per map then, or should you do two maps in one segment?
A single segment would certainly be more impressive, if they retained a good number of feasible tricks. Running a game level by level goes back to QdQ, and therein lies the challenge. Obviously I’m okay with a little variance if the trick or run dictate it. Our run had 21 additional segments, as I said. In spite of that, most segments were still loaded with difficult tricks, not divided up trick for trick, and we took care to avoid them in the interest of skill. I only mention this because at some point, the idea of introducing skill into a run seems to be very heavily neglected. That’s great if it looks cool, but can you pull it off often enough that you can also complete the rest of the run with it?
The more great things you do with less segments, the more impressive the run is. Let this facet of a run slip, and your run becomes relegated to a mental exercise. Which is still cool: I like TAS’s too. If that’s what you collectively want, the fastest possible time that someone could run the game if they could play the game 5 seconds at a time, then just call it what it is. It’s about striking a balance between skill and efficiency, and I only bring this up because there doesn’t seem to be anyone on the team advocating the balance of skill at all.
Sidenote: If SDA has removed their segmenting policy, whats to stop runs with 20,000 segments from being submitted?
We’re just looking to get the game finished as quickly as we can using bugs and exploits to our advantage, not impressing people by cramming a bunch of tricks and glitches into one map in one go, I was more impressed with dwahmov because of the tricks and strats they pulled off in each segment, I was in awe the whole time, not because it was done amongst other tricks in the same segment, but because they were amazing to see.
These segs get really freakin difficult at times! You need plenty of practice to get anywhere in these runs. We’re talking about milliseconds now, we even get rid of good segments for ones that are 0.015 seconds faster.
Then call it a TAS. When many people watch a speedrun, they do not simply marvel at the techniques being used, but also at the idea that “someone actually did that”. This is obviously diluted to some degree by level saves, even more so by AHK in our run, and now almost completely by multiplying the number of segments by 10. There were maps in our run that very few who attempted it could replicate each trick in one go, let alone come within a few seconds of the final time. With so many segments here, I imagine you would HAVE to decide a run by fractions of a second just to differentiate the skill level of anyone who has more than a year’s experience on an FPS.
As an aside, HL2DQ was originally started without AHK. Many maps, for example in Coast, were very similar to what you see in the finished product, but obviously a little more raw. Completing 95% or more of your bhops was a must if you wanted to compete with the top level times. Tricks like the mine jump, sloped flying, and most of the collision jumps you see were already being completed with just the mousewheel. Players like suga and LightningX were hailed by many for their skill at the game. It wasn’t until Lonerville revealed AHK that many became disenfranchised with spending so much additional time perfecting hops when it became increasingly clear that not everyone was on the same playing field. It’s the same situation sports would be in if all steroids were always 100% scientifically undetectable. We just opted to be forthright to our viewers about the decision to “legalize” AHK and increased the difficulty level of each run in response to it.
I mention this because it seems to have contributed to this growing idea that being good at a game doesn’t matter in a speedrun, just the creativity of the end result. Why bother trying to actually become a better player when you can just recreate the moves of an impossibly good player by creating 500 segments. In spite of the diluted skill level that AHK presented, as a very competitive gamer for 15 years and someone who spent over two years on this game, its just disappointing to me. Sorry if that involves being a downer, but someone had to say it.
Thing is, the direction segmented speedrunning was going in is viewer entertainment as opposed to skill showcase, and to be honest DWaHMoV seemed a lot more impressive to me compared to HL2DQ. A lot of segments also means extremely high standards for those segments, you pretty much have to get everything perfect before the segment is called done. This is why a couple of runs are actually not progressing at all for a long time nowadays, because there is some (short) segment that no one can get at a sufficient level of perfection. As BitRain said, we push those segments really hard, and often tick by tick.
And I wouldn’t call it a proper TAS if it doesn’t involve perfect strafing and stuff.
What are the circumstances with TASes in Source? From what I’ve heard a properly optimized TAS of Portal is possible because of the lack of RNG (though I could be wrong, I don’t have much knowledge of the games nor this engine), but surely HL2 would suffer the same problem as HL1 has for such a long time with RNG?
If you think about it, one of the main reasons we use such high amounts of segments in our projects is because we want to show off every single timesaver that is possible in the game which is essentially what a TAS achieves in most other games, however to me it makes sense to do this currently due to what seems to be a farfetched concept of completing a ‘proper’ TAS. The other thing is that even if TASes were fully feasible and optimal in HL2 from the start, would they be preferred over seeing a human-performed run that was just as optimized trick/strat-wise? A TAS might actually be less entertaining due to the robotic movement and control and that’s surely something you have to consider when deciding what the goal of a segmented run in this case is.
I’ve concluded that (even if the line is blurred between these two definitions) in my mind there are really two types of segmented run. One that is highly scripted and segmented to show off everything possible in the game taking entertainment as a priority over raw skill (essentially what we’re doing at the moment, creating the closest thing to a TAS without actually being a TAS), and the other being a QdQ style segment-per-map skill-based run which is not really being done here (though I would imagine that’s part of the reason why the original HL2DQ was named that).
In GoldSRC (I hope YaLTeR doesn’t mind me talking about this), the concept of creating an optimal full-game TAS doesn’t seem so distant now that the RNG has been researched and looked into, so they could very well be a thing very soon. This has brought me to have a re-think about segmented runs; if I ever am going to start a new one off myself, what do I want the goal to be? After all, what’s the point in a TAS if a segmented run with unlimited segments and scripts is feasible and probably going to be more entertaining? It’s made me consider that if we are to have 3 main categories (single sitting, segmented and TAS) that need to all have clear distinctions, then I might start making segmented-per-map runs so that there are ‘TAS strats’ available (in GoldSrc’s case it would be things like jumpbugs and ducktapping which require a very high amount of precision), with the focus being more on skill and leaving the TASes to focus more on entertainment value (though of course there is entertainment found in seeing more skill-based stuff, but that might have been killed off with the already high amount of unlimited segmented projects that have already been released).
To summarize, in my mind the heavily segmented/scripted runs should continue being as they are until A. an optimal TAS is viable and B. it is decided that TASes are actually even worth it as replacements over the unlimited segment runs.
EDIT: As for the saveload stuff that really seems to be a completely arbitrary and opinionated subject; if saving and loading to pull off awesome tricks/glitches (in this case pretty revolutionary ones with the OoB stuff) then why not allow it? That’s the one place where I’d say that even in a segment-per-map run it should have an exception in the ruling. It’s in the best interest of speedrunning (finding glitches/tricks etc) so I don’t see why saveload stuff shouldn’t be allowed.
If you want a segmented run with the least possible number of segments, don’t you end up with something sitting awkwardly between a single-sitting run and a perfect run? What is the point of making that kind of run? Such a run simultaneously fails to show off the full speed potential of a game and shortchanges through segmentation the “display of skill” which Max]I[muS-X has argued for. Why not go all one or the other?
It seemed a lot more impressive to you because they had 4 additional segments to do it. And “extremely high standards” to fractions of a second means that the segment is repeatable enough to hold it to that standard. Honestly, if the segment is 10 seconds or less, it’s possible to do with your eyes closed in enough attempts. At that point you could literally memorize the time and duration of each button press. Good luck doing anything of the sort in a minute or more. To your point; TAS also measures it’s improvement tick by tick, takes thousands of hours to create, and yet has zero gaming skill involved.
I’m not saying there was not skill on display in the last movie; the tricks were much more difficult given the fact that they had to move backwards at such a high rate of speed. What I’m saying is that it becomes diluted, and I feel that giving in to so many segments in the last run has created a sense that unlimited segments is no big deal. I certainly don’t think a segment for a turnstile jump is warranted. I have no idea if this was done or not, but something tells me that it, and other jumps like it, are being segmented heavily to account for 150 segments in only 5 maps.
It just bothers me to see the runs even being compared when one is being done on Easy with hundreds of additional segments and exploiting the save load feature (a feature we literally tried to avoid using…). Why would you even compare the times at this point? But yes, HL2DQ was modeled after the original QdQ and became bastardized with AHK and some additional segments, but we at least TRIED to keep some of the original sanctity intact. I actually don’t mind the concept that PJC describes in his post (despite the fact that I will continue to think that the save/load clip is not a true in-game glitch), I just hate the comparison to what we were trying to achieve.
P.S. Yes, save load was cheap in OoB as well and made it less entertaining. If you have to load the game again to complete a trick, then it wasn’t doable while you were playing the actual game and it doesn’t count. This is obviously an opinion but I don’t think I’m alone in this.
P.P.S. How are you going to handle the timing of this? Saving and loading each take a relative amount of time to the speed of your PC. Are you going to make up an arbitrary value of the save load process, or just cut it out altogether? They both seem a little disingenuous.
The point is that the game forces you to save at this point. The game is already divided in this way by the creators. The demo files are split up automatically, and the load times in between are dependent on the speed of your PC.
The other point is that there is a balance as you say between accomplishing more difficult tricks due to shorter playing periods, having periods so short that any skill required to accomplish it is completely negated by sheer repetition, and having periods so long that no especially impressive trick is possible. Anything less than 10 seconds is repeatable to the second by anyone with a few hours on their hands. Sure, you can then try and measure it to fractions of a second, but who notices fractions of a second, and more to the point, who cares? The divisions automatically appointed by the size of a traditional level creates a nice, uniform benchmark that already comes with the game.
Well what about the games that don’t have any transitions whatsoever? What about the most recent GoldSource engine that keeps the recording through loads? What about maps that start with 20 minute cutscenes and contain extremely hard tricks in the end? This is arbitrary in my opinion.
Non-linear games don’t get traditional level-based linear game treatment. QdQ, HL2dQ, are level based, linear games. They can easily be divided as such and the individual levels make great unbiased, pre-made benchmarks for speed that can be compared for times. It can be considered a little arbitrary, I’m not disputing that, and exceptions can be made. But there is a difference between making a few exceptions for 20 minute cut-scenes (as we did on most of our two segment splits with the long dialogue periods), and simply making a new split for every difficult trick, and especially for adding unnecessary difficult tricks all throughout the map that barely create a timesave and creating arbitrary splits for each one. Do you see the difference?
The latter has a place, and it’s called a TAS. Now, if you pulled off all these new moves in a single segment or so (barring the big, completely random game breakers), that’s just called a godly speedrun.
It just bothers me to see the runs even being compared when one is being done on Easy with hundreds of additional segments and exploiting the save load feature (a feature we literally tried to avoid using...). Why would you even compare the times at this point?
I kinda feel like this is the main point. Yes HL2DQ and HL2DQer are very different in what they are and what they are trying to accomplish. However also recognize and respect that this is the way the community has decided they want to make segmented runs.
I recognize it, I’m acknowledging it, and I feel at this point I’ve made the point I’m trying to make. All the subsequent posts have been responses and clarifications, and I’m sure there will be a couple more after a few other people see the posts, but you’re right, the points are there. I could have just left it alone, but I couldn’t help but reintroduce a concept that I feel like has been lost since I left the scene. I just came here because I heard that the original HL2 was being re-run and I was excited to see what the new players would bring to the table in terms of skill and creativity, and I guess since it wasn’t quite what I was expecting. Hats off to you though, I no doubt will watch the finished product. I’m sure there will be some glitches I see there that will still blow me out of the water, like the teleporter glitch which I still don’t get how it works!
Wait a sec. I think you’ve missed something here and it turned into a segmenting discussion. I don’t mind that but I want to explain some things first about this run.
Nobody explained Maximus what this “saveloading” trick is. The crazy amount of segments in maps is because of a new trick that we’ve found. Actually, we don’t really count them as “real” segments.
Basically, when you get stuck in something or in OoB, you can save the game and then reload it. Every time you do that, the game pushes you A BIT upwards (downwards in OrangeBox engine) on each save reload until you go back inbounds/not stuck anymore. The huge amount of segments are actually saveloads since you need to do many of saves in order to finally reach your destination, as the pushes aren’t very big in distance. This trick is mainly what lets us shave 30 minutes off the previous runs as we can pretty much get stuck wherever we want (using props or Triggerdelay, a lot depends on the map), go Out of Bounds, walk on top of the map to skip some areas and then easily get back inbounds to trigger the mapchange or whatever. If for some reason we cannot perform this trick and have to stay inbounds, the amount of segments we do in those areas is pretty natural. It’s just the saveloading trick that makes it look like this run is TAS-like.
https://www.youtube.com/watch?v=94PLqo7lSKI best examples of the trick
The part where he goes up to OoB is those many little segments or “save reloads”, it’s part of the trick and I don’t think it should be considered as real segments.
From what I understand is that you want a segmented run to actually not be a segmented run, but a somewhat optimized SS run, which is, in my eyes, not the point of a segmented speedrun.
For me the only purpose that a segmented run serves is to complete the game in the fastest time possible. If you now leave out hard tricks because they need more segments I am not that impressed to be honest. When I go and watch a segmented run I expect top notch optimization and tricks that are impossible when doing them in one sitting. I want to see the game being pushed to its limits and just have fun seeing it getting destroyed, whether that be in 1 segment per map or 20.
If I want to see a good speedrun done in the least amount of segments I would watch Maxxuss when he playes OE or Gocnak or Jukspa playing on the current engine. I could now argue that these runs are even more impressive than HL2DQ because they are actually done in one sitting with the least amount of segments and are therefore not so “dirty and filthy” as a segmented run where you can try the maps over and over again.
Am I also right with the assumption that you dislike any RTA speedruns since you are allowed to save / load and die in those? If not, what if they are reloading many times to do a hard trick? Is it a TAS at this point? If you get shot by a guard because you only have 1 HP and you have to reload, does it become a TAS?
Of course not, nor do segmented speedruns that use a lot of segments in order to achieve a better time. The game is not played by a computer and it does it take advantage of external programs that allow you to perform better.
You are right in one point though where I have to agree. The more segments you use the less skill is required. Sure there are still tricks that are insanely hard and require the player to be very good at the game to even have a chance of getting it but the rest of the game can be done by someone with less experience by simply repeating short segments over and over again. But this does not lower the quality of the run because, in my eyes, a segmented speedrun is not about showing skill but showing extreme tricks and exploits that completely break the game in a way that is impossible to recreate when you try it yourself in one sitting.
As I already said in the beginning it sounds like you want a segmented run to be a really good SS run which does not match with the current mindset regarding segmented runs, at least not with mine.
exec, true but max]I[mus was making an argument for 1 seg per map which is a long way from where this run, even on maps without saveclips, is on segmenting. So the discussion turned to that philosophical difference.
You seem to only be interested in consistency rather than speed which are two different things, it would seem to me that you would enjoy watching/running single segment runs a lot more. Maybe i haven’t been around long enough but measuring in fractions of a second is all that matters to me, look at the difference in challenge mode from portal 1 where it only measures each separate second and almost everyone is tied for wr, and then portal 2 challenge mode with at least 100 runners and almost no tied times. There is no motivation to improve when you don’t care about saving a quarter of a second or less, and doing this on every map adds up to several whole seconds, last time i checked every speedrunner i know cares a great deal about fractions of a second, for good reason.