Username:
B
I
U
S
"
url
img
#
code
sup
sub
font
size
color
smiley
embarassed
thumbsup
happy
Huh?
Angry
Roll Eyes
Undecided
Lips Sealed
Kiss
Cry
Grin
Wink
Tongue
Shocked
Cheesy
Smiley
Sad
1 page
--
--
List results:
Search options:
Use \ before commas in usernames
Game Page: Doesn't exist yet

Hammerfight (Any %) (Single Segment)

Decision: Accept

Congratulations to Kyle 'chairbender' Hipke!
Thread title:  
Run Information

Hammerfight (Any %) (Single Segment)

Verification Files

http://v.speeddemosarchive.com/hammerfight/

Please refer to the Verification Guidelines before posting. Verifications are due by Dec. 13, 2013.

Please post your opinions about the run and be certain to conclude your post with a verdict (Accept/Reject). This is not a contest where the majority wins - I will judge each verification on its content. Please keep your verification brief unless you have a good reason otherwise.

After 2 weeks I will read all of the verifications and move this thread to the main verification board and post my verdict.
Edit history:
Eternalspirit: 2013-12-01 03:35:59 pm
Eternalspirit: 2013-12-01 03:35:43 pm
A/V seems very good.

No cheating that I could see.

This run should probably go into an "any % with deaths" category, as the runner intentionally dies to save time. An any % should probably be a run with no intentional deaths or no deaths at all.

There is one unintentional death that costs about 1:02, and based on my viewing it would have been very difficult for the runner to prevent it, as it involved an item being greatly obscured by the stage and would have been difficult to notice quick enough to dodge it.

Otherwise, I must say I really love watching this type of game get run, mainly because it is so physics based, so while it has a lot of randomness and unpredictability, it also opens up a very high skill ceiling. The runner also deals with the randomness very adeptly, in my opinion.

There are mistakes here and there that cost a non-inconsequential amount of time (less than a minute but certainly in double digit seconds), but I would say with the things that the runner did good more than make up for that.

As far as timing, I believe this should be timed from first input to closing the last text box.

First input: ~14 seconds in (pressing play button after save name input)
Last input: ~36:41 in (last text box appear)

So my very rough estimate is 36:27. The runner got 36:22.20. Someone will need to make a decision on this and get the exact frames. Smiley

This is an easy accept from me.

I would also like to note that I think this game should be included in a future marathon. It's short, and is very entertaining for viewers (in my mind) and allows a runner to really show off a lot of skill, and this runner in particular appears to have pretty well mastered this game.
I know public verifications are meant to be short-ish, but there are no submission notes in the thread, and I feel that all runs on SDA should have a long explanation so that people can understand the game and the run. Additionally, I have played this game (in fact, I've completed it multiple times, and made some suggestions for routing it); I offered to private-verify it, but presumably I was the only one, because it's here now.

First, some general comments about the game. It's a physics-based game; you can't manipulate your weapon directly, but rather you move your vehicle and your weapon moves as dictated by gravity. Thus, you need to get it swinging first like a pendulum, then round and round, in order to get the strongest possible hits on your opponents. In order to make this faster and easier, the usual technique is to tell the game's options screen that you have a very slow mouse (thus allowing very fast movements); I use this technique, and although the options screen isn't shown in the recording, I'm pretty sure that it's being used for this run too.

Level naming for this game is confusing. Levels sometimes show their name (as two sets of Roman Numerals separated by a hyphen) before the level starts, but sometimes don't. I'm listing level numbers (converted from Roman Numerals to the normal numbering system) when they're present, and otherwise just being descriptive. Some of the levels appear to fit into the normal numbering system but don't have numbers displayed: for these levels, I give a descriptive name, and a number in brackets.

The best technique for defeating enemies quickly is to stunlock them. Colliding with obstacles will stun a vehicle (unless it's a very light contact), whether it's the player or an enemy who collides. Therefore, if you can hit an enemy into a wall or floor hard enough, you can just keep hitting them repeatedly until their vehicle breaks. The amount of damage needed to make an enemy surrender is less than the amount of damage needed to kill them; normally, all both work to clear a level, but the first is very slow because the game does not acknowledge the surrender for several seconds. (Allowing an opponent to escape offscreen in the process of surrendering, while it would be a noble thing to do in real life, is a reset in a speedrun.)

It's also worth mentioning that levels are skippable in this game (and a good thing too, because it has so many autoscrollers and annoying long levels). I tested the rules for this recently: it seems that in addition to dying three times on the level, you need a total value of 10 in coins (where bronze coins are worth 1, silver coins are worth 2, gold coins are worth 8) to be offered the choice to skip. (Be careful: the counter in the bottom-left corner, although it also mostly counts coins, counts them as 1/2/10 or sometimes 1/2/20, not 1/2/8, and thus is unreliable; it also sometimes counts things other than coins.) Skipping a level drains all your money (including any gained on the level before you skip it), meaning that it's impossible to skip two levels in a row (and sometimes you have to play two levels in a row due to not gaining enough money; additionally, sometimes it's faster to play than skip, either because it gives you a better level to skip later on, or because skipping the level is time-consuming due to needing multiple slow deaths, or because it takes too long to gather 10 coins again).

This run also has a very minimalist weapon usage; this is almost a low% (it doesn't waste time specifically for the purpose of avoiding gaining weapons, but it gets very few naturally, and even misses some that almost any first-time player would naturally get via skipping levels). This is because the starting weapon is one of the better weapons in the game (because it's very good at stunning enemies, does lots of damage, and its main disadvantage is that it's hard to hit with, something that shouldn't bother a speedrunner); better weapons exist, but they generally need unacceptably large detours to collect.

Now, here are some level-by-level notes on how the run went:

Title screen: This is a clear save file, but a glitch in the game causes it to display some statistics from the previous save file in that slot. As far as I know, this glitch is entirely graphical and has no effect on gameplay.

Tutorial (1-1): This is actually not perfect: although the runner gets an excellent first hit on the target (taking a hit on the way but that doesn't matter in the tutorial), they then get the weapon stuck on the wrong side of the vehicle and don't get another good hit in for four seconds. Now, it's far from possible to do that instantaneously, but it is possible to do faster. Perhaps a second or so was wasted here compared to optimal movement.

Tutorial part 2 (1-2): In these sorts of levels (with a circle showing), your machine can't leave the circle (there's no penalty for trying, you're just forced back inbounds). This strategy is a pretty good one, as a result (you don't want to let the opposing machine away from a wall because then you can't stun them). The first two hits are close to perfect, but the runner needs three tries at the third hit, losing around two seconds.

Route choice screen: You get a choice of two routes here. I haven't timed any of the route choices to see which are faster. However, I'm pretty much entirely certain that the route that's full of slow autoscrollers is slower than the route that isn't full of slow autoscrollers, so I agree with the route choice here.

2-1: A pretty straightforward level. Six misses here, and two of the hits weren't as strong as they could be; this is still much better than I'd do in a typical playthrough of the level, but nonetheless, it's still something like five seconds from theoretical perfection, so there's time that could be made up here. One of the misses was either a disarm attempt, or else it missed by a really long way; these sorts of levels (against allies, where you're just trying to see who is stronger rather than fighting for blood) can be won instantly by disarming the opponent, but disarming is difficult in general and particularly difficult against early-game opponents (probably because their weapons are light and they lose in a few hits anyway); I've only managed to pull it off once and I have no idea how.

2-2: Likewise, these sorts of levels can be lost instantly via disarming yourself, and given that there's a key to drop your weapon (hold the right mouse button for a while), skipping the level is much faster than playing it out. (This level has a bunch of one-on-one fights, one at a time, meaning that although it isn't an autoscroller, it nonetheless takes a long time.) Note that the runner is using a different weapon from normal; the way that gaining weapons works in this game is that they're given to you at the start of certain levels, and if you're holding a weapon when you win specific levels, it goes into your inventory and you can thereafter use it whenever you want (if you've unlocked the Hall on this or a previous game). The hammer is thus forced into inventory for playing this level.

3-1: This is very impressive. There are only a few misses, and although I'm not sure, I think they don't waste any time because the runner has to wait for the machines to spawn anyway.

3-2: The starting cutscene for this level has a tendency to glitch out, leaving me staring at a sunset and no machines for something like a minute (I think the camera spawns in the wrong place). It doesn't happen to this runner, though (I think it might be something to do with my computer). The level itself is done really quickly; getting the "Strong Hit"s on the enemy beasts is not easy, and it's necessary to be able to kill them in one or two hits. I'd be lucky to pull that off with anywhere near that consistency. (The victory condition here is killing six beasts; killing the machines is unnecessary – I just tested – but the runner does it because they can get in the way, and they have to wait for the beasts to spawn anyway.)

4: The runner got really bad luck here. The aim of the level is to protect the caravan from the sophit, but sometimes, for some reason, they spawn on top of each other and there's nothing you can do about it, thus forcing a restart of the level and costing 11 seconds or thereabouts. The second time goes much better. The sophit has a tendency to run off the screen (which is why the runner doesn't chase it after the first hit; it wouldn't work), but the runner manages to keep it onscreen via baiting it with the allied caravan after the second hit (in order to get in more damage before it leaves), and resulting in a very fast level (I just played through it to test that the quick death on the next level is possible; it took me something like 5 minutes per attempt, and I didn't eventually succeed and had to skip it, because that level is a pain when not using a sword). To visualise how fast this completion of the level is: this is faster than skipping it is (the destruction of the caravan doesn't count as a death towards skipping). Oh, and as demonstrated here, you can totally collect coins after the level ends, including inside cutscenes and for about half a second after it fades to black; this is purely useful when speedrunning, but is a complete pain when trying to test how level skipping works.

5: The aim here is to do a certain amount of damage to your opponents (shown graphically at the top of the screen). That takes a while, so the runner smashes right into some conveniently placed spikes instead. For some reason, this destroys the machine even though it doesn't do nearly enough damage, allowing for a very fast level skip. (I've tested this: it is possible, and in fact very easy, but I don't know why it works.)

6-1: Autoscroller: the aim is to dodge until time runs out. Why is the previous level skipped, rather than the autoscroller? Because the autoscroller is only 50 seconds long, and you can't skip it faster than that (dying is hard on this level because you're chained in place and the enemy isn't that good a shot, reloading it afterwards takes time). Also, the previous level is pretty long in its own right.

One thing that interests me is that I don't know what happens if you kill the opponent on this stage, and if it can cut the autoscroller short, it needs investigation. In my tests, I got the opponent down to 20% health; although you have no weapon, you do have the ability to collide, both to do damage directly, and to knock the opponent into stage hazards. (The health bar for whichever opponent you attacked most recently is shown as a yellow bar at the top of the screen.)

6-2: This looks quite sloppy, with 24 misses (and considerably fewer hits needed than that; I've done it faster than this, although I can't do it faster than this reliably, which surprises me given how much better the runner is than me at the game). It's not as bad as it looks: the hammer that the runner's been forced to use attacks very quickly, and quite a few of them couldn't hit anyway due to the maximum movement speed of the machine, or the wooden spikes on the floor blocking the way (the major difficulty of the level). Still, this level can be done a lot faster; the first hit almost got the runner in position to stunlock the enemy, and a successful stunlock, hard though it is to do with the hammer, would have saved something like 15 seconds (maybe even 20). (Note that towards the end the enemy was trying to surrender, but finishing them is faster.) The runner gains the hammer permanently at the end of the stage, incidentally, making this not a low%; you can avoid gaining the hammer permanently via throwing the hammer for the last hit. It's possible to pull this off in realtime, and it even saves a few frames because you don't have to dismiss the prompt saying you gained the hammer, but given how little it saves it's far too risky for a single segment.

6-3: This is an autoscroller, although it takes some time for the visible timer to start, and some more time to win after it disappears. So skipping the level is much faster. The level skip is quite awkward, and uses a perhaps surprising technique: if you throw you weapon, you can take damage via rubbing along it, which is what the runner does here. (If you look carefully, you can see a steady stream of blue damage numbers coming from the sword that the runner was forced to use; the individual damage is low but it happens very frequently, possibly every frame, leading to a comparatively speedy death.)

6-4: Another autoscroller, but of course you can't skip two levels in a row. This one's interesting in that the goal is to survive as long as possible, with money gained over time. You count as passing the level if you survive 30 seconds; the runner manages to suicides into an enemy weapon at 0:32 (you don't need to lose all your health to lose, just enough that the announcer will stop the fight). Incidentally, my PB on this level beats the runner by two seconds :-)

6-5: It's hard to count the misses here, but it's around 15: not bad for the hammer, which you're forced to use again for this level, but doing better would be entirely possible. If you ditch the hammer in 6-2 in order to save a few frames, I think you have to ditch it again in 6-5 in order to avoid gaining it permanently and losing the few frames you gained, making the low% route even more unwise to try in a single segment.

Route choice screen: What option will a speedrunner take: a boss fight (against a sophit, similar to level 4) that gives them an item that restores their HP when they deal damage? Or the ability to leave this story arc early and go right on to world 7? Somehow, I'm not surprised that the runner chose not to take a detour in order to pick up an item that's no use on a speedrun anyway (even if it is one of the best items in the game for casual play).

The Hall: You have a few options here when speedrunning the game. You could change weapons to the hammer the runner "wasted" a couple of frames unlocking earlier, but the mace is simply just better, so that would be counterproductive in addition to wasting time. You could also buy items, but skipping 6-3 means that you don't have enough money to buy anything good (in fact, you'd have to also avoid skipping 5 if you wanted to be able to afford items that save a noticeable amount of time, such as Siege Bombs, and that would cost much more time than the item would gain). Additionally, the shop inventory is randomized, meaning that a single-segment run shouldn't bother checking to see what's on sale, because the time it takes to check isn't worth the almost certain disappointment. Thus, the best option is to go straight to the exit.

Incidentally, you would normally have the sword equipped by default at this point, but the runner avoided permanently gaining the sword (via skipping 6-3 in this run, but there are other methods). I haven't tested to make sure that the game defaults to the mace (which I should, technically speaking, as a verifier), but I can believe that it does, and I'm not going to replay the entire game again in order to make sure that something that I think is likely is actually correct.

There are more mandatory Hall screens in the game (as well as some optional ones that no sane runner would go out of their way to visit), but they're all dealt with the same way for the same reasons.

7-1: Stunlock says hi. Don't blink, or you'll miss it. (The mace is just so much better at stunlocking than the hammer is, guys!)

7-2: This is all about suiciding quickly: killing opponents just makes more spawn, after something of a delay. The enemy weapon selection is (as far as I know) based on the RNG; some enemies kill faster than others, but the difference is quite small and this is entirely out of the runner's control anyway.

7-3: Much the same as 7-1, just with more enemies. The runner gets a strange alternate game mode here (achieved via a particularly strong hit on an enemy, and some other conditions I don't know); this causes an effect where everything lights up. Pros: you do more damage. Cons: the game enters a sort of mild bullet time effect, so everything happens slower. Overall, it mostly balances out in this case.

The runner gets an Arquebus (that they have no reason to use, although they carry it around with them all game) for free here. I think that's just a property of the level; it happens to me too, but I don't know the precise trigger. The Arquebus actually affects the machine's physics slightly, and can be dropped without losing time, but it's only a very minor change, small enough that I don't know whether it helps or hinders.

7-2 and 7-3 are a fertile ground for considering possible route changes. First, 7-2 and 7-3 are pretty similar levels: 7-3 has more enemies, but 7-2 has more text (and its enemies are a little harder to kill). I'm thus unsure which level is better to skip. Secondly, the first enemy in 7-3 seems likely (guaranteed?) to spawn with an Iron Mace (the intended upgrade from the Stone Mace, the weapon that the runner has been using all game so far), and it costs no time to change weapon here (beyond a few frames from the message that shows when permanently gaining a weapon). The Iron Mace does more damage, meaning faster kills. However, it's slower to swing, which is a potentially large disadvantage (especially on 10-6). I'm not sure which weapon is better for the route, but this is the first weapon in the game that's at least worth considering.

7-4: Textbook stunlock. I think the Iron Mace would do slightly better here (it would establish the stunlock faster, and kill in fewer hits); the Stone Mace does pretty well, though, and apart from maybe a couple of seconds lost to a couple of missed swings, this is impressive.

Route choice screen: I have no idea which of these routes is faster (I can't even remember which levels are on which routes), so I won't really comment on the runner's routing choice here. I will say, though, that most of the levels on this side can be done very quickly with the strategies in this run, meaning that I would be surprised if the other route were faster.

8-1: Spikes are handy when suiciding. Not much else to say.

8-2: This is pretty fast, helped out by the slow-time effect: the first enemy is effectively killed in just a few seconds (knocked down to surrender range). The runner wastes a few seconds trying to finish off the first enemy from there, though (which is still faster than waiting for the surrender), partly because the enemy weapon gets in the way. Then they get the slow-time effect again and use it to mop up the enemies that spawn in response to the first enemy's defeat.

8-3: Spikes are handy when suiciding, autoscroller edition. It's pretty much completely obvious that this is the fastest strategy for this level.

8-4: This wasn't quite as good as some levels, due to allowing one of the enemies off the screen (about two seconds), and due to needing some time to get the first enemy into position to attack them (maybe about another two seconds slower than it could have been). It's still pretty fast, though. The second enemy's weapon gets in the way, but in this case it doesn't waste much time due to the stylish backup strategy of killing them with their own weapon.

8-5: This is how mace fights against large machines go. Maces are overpowered enough as it is: it's kind-of devastating to see them against an enemy type that's weak to them. This is probably another level that can be safely placed in the "faster to play than skip even if you don't care about the levels either side of it" category.

Incidentally, that isn't an audio glitch: that noise is what explosions in Hammerfight sound like (presumably, they're so loud that the player's ears start ringing).

I don't know the trigger for the Assault Arquebus, either.

9-1: OK, so the weapon you're given for free on this level (the Steel Whip) is the best weapon in the game that doesn't have an unreasonable time cost to obtain: it's mace-type (which is good to start with), and better than the Stone Mace in basically every way (which is even better). The question is, is it worth completing this level the intended way to gain it permanently, rather than just skipping it (which is quite easy; the runner optimizes it here via throwing it downwards, which makes it go offscreen faster, triggering the level loss condition sooner)? A perfect solution to this level (in which you throw the weapon upwards at a diagonal to break lots of targets at once) is quite fast, but also very hard to pull off. Meanwhile, your new weapon would only be useful in levels 10-1, 10-3, 10-4, 10-6, 11-1, 11-2. It's something to think about for future runs, at least.

9-2: This level is mostly an autoscroller, but playing is is non-negotiable, because 9-3 absolutely has to be skipped. You can't speed it up (except by when you get the hit on the last target), but you can slow it down, via missing targets. This run misses two targets, and also gets the slow-time effect three times (I think you have to hit the targets less strongly than the runner is to avoid it), so it's far from perfect, but the time losses aren't that large in an absolute sense, with the mistakes costing less than a second each. The explosion ring is because some of the targets (marked with flames) are booby-trapped to explode when destroyed; you're meant to avoid them, but I don't see why anyone would, because they don't do enough to kill you anyway.

9-3: IMO this is the worst level in the game (even counting the one near the end which, according to the forums, caused more people to ragequit the game than all the other levels put together). A ball is delivered, really slowly; you have to hit it high up into the air and get points for how high you hit it, watching it all the way. Then you wait for the next ball, again and again and again. You can make things marginally faster via not hitting it as high, but then you'd stand a pretty high chance of not getting enough points and having to replay the level.

So what can you do instead? It's not at all easy to commit suicide on this level, but almost all terrain is destructible in Hammerfight given enough force. So the runner makes some flat space in which to rub against their hammer until they die. The first death was pretty sloppy (compare the area preparation to the second, which looks better but which is only a second faster because it has trouble rubbing against the hammer); the third doesn't prepare the flat space properly, and demonstrates why it's necessary, but it nonetheless goes well enough that by the time it finally goes wrong, the runner manages to recover with a spectacular backup strat.

To gain an idea of how much time can be saved here, I timed the three suicides (approximately, there's not really much point in using frame advance when estimating): they took 11, 10, and 18 seconds respectively. This leads me to believe that this level can be improved by at least 10 seconds with better execution.

10-1: Stunlocking's great, isn't it? This example loses at least a second through missed shots and shots that are weaker than necesssary, but I don't think it loses that much more than a second; it's still pretty good.

This is also the first point at which you can use the free gem that you obtain earlier in the game. I'm not sure what determines what size gem you get, nor what size the socket in the Stone Mace is. It would definitely save time (probably tens of seconds over the course of the game) to use the gem if it fits, though; orange gems do bonus damage, meaning fewer mace swings are needed to complete the game. (In fact, if the size of gem is consistent, this may be what ends up swaying the optimal choice of weapon. Or it might not; you wouldn't get to use it for long.)

10-2: The runner apparently lost count of how many coins they collected in 10-1: six silver is enough to be able to skip a level. The execution here isn't bad, but given that there are five enemies the level takes 41 seconds, and although I'm not sure (I can't easily time it because the netbook on which I'm writing this is not fast enough to run Hammerfight at full speed), I believe that skipping it is faster than that.

Shop: Throwing weapons could conceivably save time; even though they don't do very much damage compared to any of the maces that the runner could have at this point, it's possible to throw them at a spawnpoint to hit enemies as soon as they spawn. Actually pulling this off, though, is very hard to do and saves one swing of the mace at best; it'd be statistically insignificant compared to the number of missed shots. And of course, you wouldn't have any money if you skipped 10-2, like an ideal route would. So I don't blame the runner for not trying.

10-3: This would have been faster if the enemy had been knocked into the bottom-right corner earlier (where in addition to damage from your weapon, you get damage from the wooden stakes and from the wheels on the ground as well). The runner may want to consider attacking from a different direction in order to save a couple of seconds.

10-4: I think the first enemy was killed by the floor hazards, which is why it was credited as "defeated" rather than "killed". This is a very fast completion of this level.

10-5: Suicide skip, again. The second and third skips are slower than they would have been with perfect luck, because the AI and RNG didn't cooperate. Not much you can do about that in a single-segment, though.

10-6: This level is quite hard to use the standard strategy on: the enemies have green gems (allowing them to turn temporarily intangible in order to dodge weapons), meaning that it's hard to get them into position to stunlock. The runner's strategy here is to simply perservere until they screw up; I don't know a better strategy, but it feels like there might be one (especially because the runner actually got hit amongst all that, losing 4 seconds and ending up in severe danger of dying unintentionally and having to start the level, although they pulled through with a few hitpoints left). The enemies also use projectiles; the runner lost a second due to impaling themself on one after it had landed, so picking those up might be a better option in future runs (it doesn't really lose time, just hold the left mouse button while you're busy stunlocking an enemy against the ground).

11-1: Very impressive. This sort of thing is why all speedruns of this game should use mace-type weapons.

11-2: The level itself is done pretty quickly, and I don't have much to recommend. However, there was time to pick up that orange-gemmed steel whip during the victory fadeout; it breaks my heart to see someone leaving one of those on the ground, because it's the best speedrun-obtainable weapon, and the game gave an opportunity to grab one at no time cost bar the couple of frames for the resulting text box. How much time would doing that save? Sadly, probably more than a second or so, there are hardly any levels left to use it.

11-3: The enemy actually escapes the stunlock here, but the runner re-establishes it again soon after, meaning only around 1 second was actually lost.

12: Huh, I can go back to counting missed shots again (four, plus another four or so that were too gentle to make much difference), with the level maybe five or so seconds away from where it could be. It's been a long time since an enemy had enough armour that they survived stunlocking for long enough to force the runner to actually have to aim.

The Emperor: If you frame-advance to actually read the text, you'll discover that this game actually had a MacGuffin hunt, but all three MacGuffins were found by NPCs off-camera. I like the way that this game deals with boring sidequests by making them happen entirely in the background, with the player unaware of them.

Finally, comes the last and possibly most interesting choice in the route. One of the options allows you to keep your weapons; the other causes you to temporarily lose them (for some reason, you can just re-equip them a couple of levels later via entering the Hall via the main menu, but the game never tells you that this is possible and it's too slow to bother with in a speedrun anyway). Each has its own set of levels. Of course, given that the runner has used the Stone Mace all game anyway, I'm not surprised by the choice, even if I have no idea which set of levels is faster (again, I'm not clear on which levels belong to which set).

The Abyss (13-1): The runner may have lost their weapons, but there's a hammer lying on the ground. They then immediately use it to kill an enemy with a Stone Mace and swap weapons. At least this runner is consistent in their item choices. The rest of the level is done pretty quickly; there are a couple of mistakes but they don't lose time because you have to wait for the enemies to spawn anyway.

I don't recommend trying to follow the plot of the game past this point. It doesn't really make sense.

13-2: If I remember correctly, this level is really, really long. No wonder the runner just skips it via suiciding on their mace and against enemies. Skipping this level is more luck than skill, really: it depends on which enemies spawn and whether they actually get attacks to connect.

13-3: You can only really damage that machine on the core, so most of the apparently missed shots are just keeping the mace swinging while waiting for it to present a target. (The runner actually gets hit once, presumably due to impatience; that didn't really waste time, but looks bad.) It's convenient that a short level with enough coins generated to buy a level skip comes between two long ones.

13-4: Another skipped level. The enemies do more damage than rubbing against the stone mace does, so it might have made more sense to let them damage the runner to get the third death, than spend a while flailing around trying to drop the mace and inadvertently dodging them in the process.

14: The mistakes earlier in the level (failing to kill the enemies as fast as possible: compare with how fast they spawn in the second attempt) cause a moderate time loss; it's unclear how bad this is. However, the unintentional death definitely causes a very large time loss (61 seconds, to be exact). The death seems to come completely out of nowhere, but after carefully reviewing the recording, it turns out to be due to a spiked hammer (not sure what its actual name is) that came loose and is hidden behind the wooden obstruction across the centre of the screen. The second attempt goes much better, and the runner (having played the game before) attacks the centre of the level during the victory screen in order to progress the plot. (Most first-time players will spend a long time wondering why nothing is happening before they start destroying the terrain.)

14-2: Autoscroller. The runner wisely skips the level using a method mostly based on damage from enemies damage (although they rub against the mace a small amount, that hardly matters; discarding the mace is useful mostly to prevent it hitting enemies by mistake). By the way, IIRC this is the level that causes most people to ragequit (I think it's this one; it might be 13-4). Especially if they're using metal weapons.

14-3: Sadly, two long autoscrollers in a row means that they can't both be skipped. You don't gain much from attacking here, so the runner actually discards their weapon altogether for better manoeuvrability, and just focuses on dodging. It's not like it'll be needed for the rest of the game. (The text on the timer is always bugged like that, by the way; this is definitely not a Hammerball match.)

15/Epilogue: Although you can generally skip levels in this game, this level is different: it's the final boss, but you can skip it simply via rejecting the option that the game gives you to fight it, another level of WTF on top of the confusedness of the plot of the end of the game. The runner does so (the boss doesn't take that long, but fighting it is definitely slower than not fighting it), and gains the good ending; I'm not entirely sure what triggers one ending or the other, but this is perhaps the biggest surprise of all.

I'm not sure what the timing rules are for this game, but the last time player input can affect the game (rejecting the fight on the final boss) happens at 36:31, and the first time player input can affect the game (gaining the ability to move the machine at the start of the tutorial) happens at 00:20, so I'd guess a time of 36:11. (This run should be marked as "with deaths", by the way; deathless is much slower due to the inability to skip levels, and thus a completely different category. And a rather boring, autoscroller-filled one at that.)

As for accept/reject, this is tough. The run clearly shows a very high level of ability at the game; it's a lot better than anything I could manage, for instance. However, there are several levels that are clearly improvable, in addition to a costly unintentional death. I'd definitely reject this run if it were segmented. For a single-segment run, though, mistakes are going to be inevitable; nobody can be expected to get all the levels perfect, and for every level with ugly misses, there's a level that's done so quickly I had to replay it several times to work out what happened. As for the route, I wouldn't reject this run based on the route even if it were segmented; it's unclear whether the route is the best one or not, but there's nothing obviously wrong with it (even if it is almost a low%), and improving on it will be difficult.

So I guess it comes down to, do I think the runner could improve this run? I think that with enough practice, they could, and maybe there'll be another submission in a few months, or a year. However, I don't think that it'll be easy to improve, or that submitting this run now is a case of being sloppy. As such, I'll tentatively suggest accepting this run, although I could certainly understand if it were rejected.
There were some notes attached but, uh.. yeah.. if the runner missed anything, you certainly made up for that. Smiley
Edit history:
LotBlind: 2013-12-04 10:26:56 am
FINISHED

14 (the spinning morning star level) - The death looks pretty random, I didn't even get what caused it before reading the comments. And it's true the runner almost chain kills the robots on the second try. Also it's the 3rd to last level.

Seems this game has several endings. Does it change the game a lot to choose a different path?

Accept
Not a walrus
One question for those of you who have watched it: If you had to guess, how much time to the intentional deaths save?
5-10? With the understanding that this is a complete guess..
What.. a weird.. game..

Anyway A/V is excellent, no cheating, blabla
No idea how much time deaths save, because I never played it. But I'm guessing it's substantial.

Accept
Edit history:
chairbender: 2013-12-03 09:53:42 pm
TwitchTV: chairbendr
Hi everyone, thanks for all of the feedback and thanks for spending your time to look at the run! Thanks especially to ais523 for missing my run notes and writing up better notes than even I did!

To answer the deaths time savings estimate: As part of planning this run, I played through every single branch of this story and played every single level to figure out the fastest possible completion time doing a suicide vs an actual completion of the stage. I estimate that a deathless run would be 5 minutes slower (seriously, it's almost exactly 5 minutes slower according to my slightly-out-of-date measurements). It also wouldn't exhibit some of the interesting suicide methods you can employ Wink

If you want to try to make sense of my spreadsheets and figure out the times for different route options (the individual times for each stages are not the most recent, but they're close), you can view them here. Each measurement was generally taken from when you click after the starting screen for a stage (the one that shows the stage number) to when the victory screen fades to black or you left click to accept the level skip.

By the way, there is a knowledge base page for Hammerfight (I've written most of it). It's here - https://kb.speeddemosarchive.com/Hammerfight.
Yeah, this game is... interesting...  Looks kinda fun though.

A/V good, no cheating detected.

The death in 14 is really the only thing that makes me hesitate, but it seems to me that getting the rest of the run to this level is no small task, so I'll give this an accept.
I actually played through the levels that chairbender skipped, for fun, a while ago. It takes me a lot more than 5 minutes :-)

Seriously, though, many of the skipped levels can be completed quite quickly, although not as fast as skipping them. (Including the one everyone hates, and the one I hate; the one everyone hates took me a huge number of tries, but a successful completion doesn't take long. I did use a Siege Bomb on it, though, but I think it missed, so that shouldn't affect the timing.) Btw, I discovered (now that I suck less at 9-3) that it finishes early if you hit the record score early, which means that skipping it doesn't save as much time as I thought (although it did save time).

The really big reason that level skips save time, though, is that many of the skipped levels are autoscrollers, which can't be sped up no matter what you do. The branch choice here is mostly designed to minimize on autoscrollers (they tend to come in groups).

A different ending would not noticeably change the run. With the route choices in this run, you can get the bad ending via choosing to fight the final boss. (That said, I don't know how long it takes to beat with the stone mace; I've never tried that. If its projectiles behave the same way as they do when you're using a machine flail, you'd want to buy a Rushid's Vial at some earlier point in order to avoid the need to wait for its first phase to end, because attacking early means you take damage, but given that they're purple, and purple in this game means electricity, there's a chance that they react to metal specifically.) Incidentally, remember when I said a machine flail 2HKOs everything? That works on the final boss too, once you get into position to attack.

Btw, I found a much faster strategy (with the stone mace, as used in this run) for 10-6: feint at the opponent to make them use their gems, then quickly swing it back round and hit them before the gem recharges. I think it might only work for me because my computer is so slow, though; someone with a faster computer will need to test to see whether it's possible with the game running at full speed or not.

I've also been testing the following item choices: Iron Mace; Steel Whip + orange gem; Machine Flail + purple gem. I don't think any saves a noticeable amount of time in this category, although I'd recommend upgrading along the sequence for a no-level-skips category (the Machine Flail 2HKOs basically anything, the purple gem hardly helps but the fastest one to obtain comes pre-gemmed).
Decision posted.