> you should number the rooms in the maps or something
I actually have numbered them in this map.
http://bisqwit.iki.fi/kala/snap/cutman-numbered.png
(Sorry, no numbered maps for other stages.)
The numbering corresponds to the internal representation of the rooms in the game. Some numbers appear multiple times, but that's how the game works. Everything is compressed :)
> About cutman: In one of the vertical screens, with the armoured cannons (you should number the rooms in the maps or something)... when you jump to the left of a platform and in to the right instead of going around it, it looks like you have to wait for a while to land on the ground. Could you grab-release the ladder to get faster above the floor to make your next jump?
Unfortunately I couldn't understand what you are referring to.
But I think I have tried grab-releasing in all applicable places and chosen the fastest way.
> I don't understand why you did the rooms with the red blocks the way you did. In particular, in one of the laggy rooms you take damage and pause to go vertically trough one of them.
Because it's faster. :P
The first jump should be obvious (the apparent delay before jumping was because I had to move Rockman 1 pixel to the right), but why I paused in the second jump? It was to eliminate lag. The room lags a lot. Rockman's movement is purely vertical in that position, and pausing gives lag-free vertical movement at the cost of 2 frames.
Edit: Actually, I think I should try to do it in room "3" too!
Edit 2: I still probably should dare to try not using the magnet beam in the previous room. It could eliminate the lag in that next room.
> At the boss, could you switch weapons on the way down on the right side of the guts block? It looks like it takes a while before you can pick it up anyway, and it's only vertical motion.
That's a good idea! I'll do that.
(I recall this idea is not new, but somehow I had forgot it.)
> I loved how you scrolled to be placed just above the lethal spikes, and then went upwards to scroll the screen down.
Yeah, it looks funny. But it was invented by Morimoto, not me.
Thank you for the feedback!
It goes just before the "b1=RdMem(_PC)" line in x6502.c. For every frame, it prints what bonus the game would give if an enemy spontaneously died at that frame.
It works perfectly.
Analyzing the output, I see that the "temp" variable changes less than I thought. If there are no other objects on screen and Rockman stands still, the "temp" variable keeps a steady 0x7C value. When Rockman climbs, it keeps a steady 0x03 value.
However, when there is even one "flea" jumping around, the number changes once every few frames. It seems, for the most part, to always contain values that are evenly divisible by 4, but not really always.
If the value didn't change, I could predict the bonus drops for hundreds of frames ahead, but when it changes, I can predict nothing because it works in a bit similar way as the CRC calculation works: tiny ripples change everything.
But I keep trying to find ways to abuse it.
Curiously, sometimes there are situations where you wouldn't get a big weapon drop no matter how many frames you waited. The RandomSeed can enter a short loop of unhelpful values. The loop can only be broken by influencing the TempVar.
