Donkey Kong Forum
Donkey Kong Strategy => Advanced Donkey Kong Strategy => Topic started by: JCHarrist on February 02, 2013, 11:38:51 am
-
Ever since I started point pressing, I've tended to avoid the bottom hammer on level 4 because I wanted to minimize my exposure to the heat seeking wild barrels. Recently I've heard discussion from some top players that level 4 is actually safer WITH the bottom hammer if you know how to play it correctly.
What are some of the keys to playing the bottom hammer on level 4?
-
I will let others elaborate, but the idea behind taking the bottom hammer on L=4 is to primarily stall for some time until the game switches over to L=5 difficulty-style wild barrels, which are of course much easier deal with. Someone correct me if I am wrong, but I believe the difficulty switches 33 seconds into the screen, so usually by the time you use the bottom hammer the switch will have already occured. Me personally, I like to park it just to the right of the long ladder while waiting for the best time to grab the bottom hammer. Reason being, I find it much easier to manage L=4 type wild barrels thrown while staying in that area because they are pretty easy to dodge by moving to the left. That's usually my plan, but of course improvisation comes into play depending on what the first fireball decides to do. haha
-
Yeah, I agree with what Mark says. The difficulty switches when timer gets to 6000. Stay in the middle areas is safest to avoid the really wide swinging wild barrels. Just run away from them and be prepared to jump if the come near your feet. I try to avoid timing a hammer hit on them because of their angles, they can easily slip through.
-
I agree with Mark, however, after playing a bunch of the wild barrel hack and noticing how level 4 wild barrels relate to level 3 wild barrels, I'm thinking it might be best to wait near the right side of the 2nd girder and just steer the wild barrels off the screen. This way you also stay closest to the hammer. I haven't had time to thoroughly test this yet with a save state. Maybe Jeff Willms can make a hack for level 4 wild barrels! ;)
-
I thought the wild barrel hack was just all wild barrels of the level you're on, so when you get to level 4, you would have all level 4 wild barrels. So, the hack is level 3 wild barrels on every level?
-
I agree with Mark, however, after playing a bunch of the wild barrel hack and noticing how level 4 wild barrels relate to level 3 wild barrels, I'm thinking it might be best to wait near the right side of the 2nd girder and just steer the wild barrels off the screen. This way you also stay closest to the hammer. I haven't had time to thoroughly test this yet with a save state. Maybe Jeff Willms can make a hack for level 4 wild barrels! ;)
Level 3 and level 4 wild barrels are the same, just like level 1 and level 2 wild barrels are the same.
I thought the wild barrel hack was just all wild barrels of the level you're on, so when you get to level 4, you would have all level 4 wild barrels. So, the hack is level 3 wild barrels on every level?
No, that would make it too easy once you reach level 5, so I made them level 3 difficulty for all levels.
-
Ah ha, that explains it.
I save-stated my way to L5 and said, "wait a minute, they're still pinballers, no fair!"
Honestly, if a player can make it through L1 through 4, I think they deserve L5 barrels. It would also make a kill screen possible on this version!
-
I really like the idea of just all level 3 wild barrels. Then it becomes almost completely a matter of skill and method to get a good score on this hack.
Jeff, are you sure levels 1 and 2, and levels 3 and 4 wild barrels are, respectively, the same as one another? Did you find this in the code? 'Cause my experience tells me they are significantly different -especially after having played for hours with the level 3 wild barrels in your hack and then getting very surprised when I die due to (seemingly different) wild barrels on level 4 of normal DK.
-
I always thought there were subtle differences between L1-L2 and L3-L4 wild barrels, but I have no clue if there really are or not. L1 and L2 both kinda just fall down randomly...sometimes L2 do seem to bounce a little more toward you as they fall. L3 and L4 are both evil, but L4 seem to bounce even crazier than L3...faster/sharper angled bounces.
But yea, first time I played the wild barrel hack I knew right away they were either all L3 or L4 wild barrels...evil.
As for taking the bottom hammer on L4, aside from what Mark already said...it's also safer because if you rush up to girder 3/4/5 and a wild barrel gets tossed at you, you have less time to figure out how to dodge it/steer it out of your way. You don't want to be on the 4th girder when a L4 super sharp angled wild barrel gets tossed your way. Often these are unavoidable deaths...even if you jump away from it...it'll still hit you in the feet quite often.
Also, rushing to the top and grabbing the top hammer on L4 leaves you open to wild barrels being tossed on your head...or a sharp angled one that'll sometimes hit you in the foot after bouncing under your hammer.
-
Jeff, are you sure levels 1 and 2, and levels 3 and 4 wild barrels are, respectively, the same as one another? Did you find this in the code? 'Cause my experience tells me they are significantly different -especially after having played for hours with the level 3 wild barrels in your hack and then getting very surprised when I die due to (seemingly different) wild barrels on level 4 of normal DK.
I had looked at the wild barrel code before (although not very in depth) and had seen that difficulty 1 and 2 are handled by the same block of code, difficulty 3 and 4 are handled by the same block of and difficulty 5 is handled by its own block of code. I hadn't looked at the details or anything.
Just now I took another look through it and I think I have a more precise understanding of how wild barrels works. From what I can tell L1 and L2 wild barrels behave identically and are purely random. Both speed and direction of these barrels are random, although only a certain range of speeds is allowed.
L3 and L4 wild barrels also behave identically and use Mario's position to determine the direction in which to bounce. The speed is influenced by randomness. If the barrel is bouncing to the left then the speed will be chosen at random within the same range of allowed speeds as L1 and L2 wild barrels. If the barrel is bouncing to the right then the speed will also be chosen at random, however, the range of allowed speeds is such that the minimum allowed speed is the same as the maximum allowed speed of L1 and L2 wild barrels, and the maximum allowed speed is twice that amount (i.e., for L1 and L2 moving to the right, they can have speeds from 0 to x and for L3 and L4 wild barrels moving to the right they can have speeds from x to 2x). Mario's horizontal distance to the wild barrel does not appear to have any influence on the speed.
L5 wild barrels are purely deterministic with both speed and direction determined by Mario's position. The greater the horizontal distance between Mario and the barrel, the greater the speed will be in the direction towards Mario.
It's always possible that there's more to it than this, if some other random place in the code is having an influence on the wild barrels, however, these behaviours that I have described based on the code do match my personal observations of wild barrels. I have not noticed any difference between L1 and L2 wild barrels or between L3 and L4 wild barrels.
-
Interesting. Is there any chance you could make hacked roms for each of the other types of wild barrels? How difficult is that to do? Maybe if I can learn how to use the MAME debugger to see what files I need to remove/change I can do that myself. It seems doing this would be the easiest way to prove this to everyone once and for all, since, I'll admit, I'm still a bit unsure.
-
People draw a lot of conclusions about the DK code that I don't think they necessarily should.
Anyone who has worked with any sort of code, whether a programming language or even something as simple as HTML, knows that once you stick enough code together, parts are going to interact with other parts in unexpected ways to produce phenomena that are not intended or explicitly programmed, phenomena that not even the developers would be aware of.
A few years ago, a player who believed that he wasn't getting enough 800 smashes on the rivets stages might have been dismissed as imagining things.
And yet, lo and behold, the recent discovery that the program is bugged in a way to cause just that problem (http://www.classicarcadegaming.com/forums/index.php?topic=4164.0).
I think we might be unknowingly dealing with a lot of things like that.
For example, most people say that the rivet screens on Donkey Kong are identical to the rivet screens on Crazy Kong, and that only the sprites (fireball vs. firefox) are different.
As someone who has played a lot of CK, I'm pretty adamant that the rivets play differently on CK, and in a way that makes them generally harder. The fire simply doesn't behave the same way. I can't quantify it, I just know that I die on the rivets far more often in CK than I do in DK, and am constantly ending up in tougher spots.
The closest I have come to consciously articulating it is that the fire in CK seems to take longer strides, and to climb ladders more readily. But I don't know.
Robbie Lakeman agrees that CK rivets is not the same as DK rivets. He also can't put his finger on it, but concurs that *something* is different.
People who know the code, however, seem to be of the opinion that there's no difference, only the sprites.
While it may be true that there's no *explicit* programming to make the rivets different in CK, some other unknown, accidental wrinkle might be causing that screen to run differently, for reasons similar to the ones that lead to the 800 smashes being unfairly distributed.
I would not be surprised if the same sort of thing were true of the wild barrels!
-
Either way, I think it would just be cool to have hacks of each type of wild barrel to practice. If some end up being the same, then we will only need fewer of them.
However, I realize that, if Chris is right, and that the thing that makes the wild barrels different is not explicit to the code, Jeff's very method of modifying the explicit portions of the code may still affect the actual outcome to make the barrels act the same in the hacks -though they might actually act differently in the normal rom. So even my idea of making roms specific to each type might not be 100% foolproof, but I would still assume it is likely a very good test.
-
In the CK vs DK rivet stage portion of this topic, I have put in a lot of hours for both games and my rivet stage deaths may seem like they happen more often in CK than DK because the rivet stage is really the only tough board in CK. But after observing the fireballs in both games I feel like that there is increased speed in the fireballs in CK. They seem to spend less time thinking about going up or down a ladder and just go as in DK they can hang around in an area for a while before they do. They also seem to go down the ladders at the same speeds for both games, but going up ladders is a different story. There are certain points where you get into a race to a top rivet with a fireball in DK and you can beat them to that rivet is a bit of a head start. In CK I almost always lose that race and have to hope it immediately goes back down. I don't know much about code for these games, but this is just what I have observed from the amount of time I have put into most games.
-
There's actually nothing programmed into the hack that distinguishes the wild barrels as being L3 difficulty as opposed to L4. Basically what I did to make them all behave like L3/L4 is I made the wild barrel code always branch to the L3/L4 difficulty code block (the same code block for L3 and L4) instead of choosing the code block to branch to based on the internal difficulty. I left the actual internal difficulty unchanged.
So with that said, if you want to be guaranteed to see true L3 wild barrels as they would appear in DK then I suggest you go to L3 of the hack and look at how the wild barrels behave there. To see L4 difficulty wild barrels go to L4 of the hack. If it turns out that the behaviour of them is different on the two stages then it is likely that the wild barrels you see on L1 in the hack behave differently from both normal L3 and L4 wild barrels.
You should keep in mind, though, that I made some other changes to the hack besides making all wild barrels L3/L4 difficulty that could theoretically change wild barrel behaviour causing the hack wild barrels to behave differently from actual wild barrels in DK. Specifically, the other changes are that I modified wild barrels to be capable of reusing the sprite slot of previous wild barrels without becoming bombs (so that you don't start getting only bombs) and I also modified the 2nd and 3rd barrels, which are normally required to be normal barrels, to be wild barrels like all the others. But assuming you don't believe either of these changes has an impact on wild barrel behaviour then I think you can be fairly certain that the wild barrels that you observe on L3 and L4 of the hack are the same as actual L3 and L4 wild barrels of DK respectively.
-
Someone correct me if I am wrong, but I believe the difficulty switches 33 seconds into the screen, so usually by the time you use the bottom hammer the switch will have already occured.
I think this is correct. It's important to note that this is 33 "real time" seconds.
The difficulty switches when timer gets to 6000.
This is much less accurate. The reason is that the switch is based on "real time" seconds. The rate at which the Timer ticks down on a Barrel screen can be highly variable based on how many wild barrels are released and especially based on how many objects are smashed, which pauses gameplay (including the Timer), but does NOT pause real time. Because of this, the switch can actually occur within a pretty wide range of Timer values (say, between 5600 and 6500 perhaps) although it's still good enough to give you an idea that you should be looking out for the switch. Remember, the switch affects all objects on the screen simultaneously, including objects that are already "in play". So, a wild barrel which was behaving like a Level 4 wild barrel during it's first 3 bounces or so off of the upper girders can suddenly convert itself to a Level 5 wild barrel on the last bounce of of the girder directly over your head and instead of bouncing wildly out of the way, it can just drop straight onto your head instead. Watch out for this transition.
L3 and L4 wild barrels also behave identically and use Mario's position to determine the direction in which to bounce. The speed is influenced by randomness.
Interesting analysis Jeff regarding how Level 3 wild barrels are working. However, I wonder if you can take another look through that portion of the code when you get a chance. You are using the term "speed" in your description of what is variable through randomness. My suspicion is that this is not accurate.
I believe that what is variable through randomness is more closely related to "angle". If I were to guess, I would bet that the vertical speed of a wild barrel is fixed. Meaning, there is probably some code that is reused often that handles the physics of gravity -- I'm betting that Vo is always the same (initial vertical velocity after a bounce -- y-direction) and the acceleration due to gravity is applied to the object in a close approximation of what would be expected by gravity (notice that bombs, which do not bounce and therefore do not reset their Vo, accelerate noticably so that their vertical velocity near the bottom of the screen is much faster than near the top). What I'm not clear about is whether the actual vertical distance between bounces is variable (due to the ramping of the girders) -- it appears that they are -- and so it might be true that not all wild barrels reach the bottom of the screen in the same amount of time.
Anyways, if the "angle" of a bounce becomes sharper, and the vertical distance is always traversed in the same amount of time regardless of whether it's a shallow path or a sharp path, the result is that a sharper bounce will yield a faster object. I actually believe that this was an oversight by the programmers, but they left it in after testing because it was just another way to kill the player faster in a manner that is only slightly unfair.
-
After about 2 hours of running as many tests as I could think of using the Wild Barrel Hack and the normal DK rom, I am ready to say I agree 100% with Jeff's conclusions. Level 1 and 2 wild barrels are the exact same, and only have 1/2 the potential mobility of level 3 and 4 wild barrels. Likewise, levels 3 and 4 wild barrels are the exact same as well.
I realized, after running the tests, that the supposed 'difference' I thought I had seen between level 3 and 4 wild barrels was simply me having experienced a significant amount of very 'sharp-to-the-right' wild barrels (due to the innate random element in the wild barrels' angle producing that specific sharp angle) and me thinking that, therefore, level 4 wild barrels had even greater mobility than the ones on level 3. However, even after having almost been fooled by this very same thing in my tests (I had several very sharp level 4 wild barrels right away during my tests), I eventually realized my error. I did a handful of tests by parking Jumpman in certain unsafe and safe locations on both the normal rom and the hacked rom and all the results came up consistent with what Jeff stated. There were no statistical differences I noticed that would merit saying level 2 or level 4 wild barrels are their own 'type'. I'm certain now there are only 3 types of wild barrels: levels 1 and 2, levels 3 and 4, and levels 5+.
Thanks for setting me straight on this issue everyone! :) Now to go re-edit a certain portion of the DK manual... ;)
-
Interesting analysis Jeff regarding how Level 3 wild barrels are working. However, I wonder if you can take another look through that portion of the code when you get a chance. You are using the term "speed" in your description of what is variable through randomness. My suspicion is that this is not accurate.
When I refer to "speed" I am referring to horizontal speed, and when I say direction I'm referring to horizontal direction (either left or right). I agree that vertical speed is likely determined by fixed gravity alone and is always deterministic. The code actually has no angle calculations for wild barrels, it just adjusts the horizontal speed of wild barrels. When it sets the horizontal speed to greater values it will result in a steeper angle, as well as a faster overall barrel speed.
-
Maybe he has just started doing this more frequently or maybe he has been doing this for quite a while and I just now noticed it, but I noticed the other night how Dean grabs the bottom hammer. Instead of kinda dancing around and moving between the left and right sides of the bottom hammer as the fireball gets closer to the ladder below it (as people often do, when they are trying to prepare for when the fireball will climb), Dean simply positions himself on the left side of the hammer and waits (unless the fireball is not near the ladder, in that case, he is busy grouping barrels and jumping them as need be -which often means he is further away from the hammer). Once the fireball starts to climb the ladder, he scoots over just a bit, so that Jumpman is kinda half-way under the hammer, and then jumps to grab it. Because of how the hammer hitbox works, he is able to grab the hammer without being perfectly below it. This gives him the necessary room to be able to land in time to smash the fireball.
Again, I don't know if it is just me noticing this technique for the first time, but it seems so simply and efficient that I can't believe I never thought of it. Here I was, doing all this silly running around under the hammer and trying to find the best position to be in in order to grab the hammer when the fireball climbs but still be able to jump barrels. Silly me! :P Thanks Dean!