We talked some on discord, and I can see the argument for allowing a save state. I will say that I don't really like it from the standpoint of it being a bit of a slippery slope, making things easier to cheat/harder to verify, but I doubt there would be much of a problem on this small of a board.

The other issue is that there isn't much run left if you remove the grind.

If we are going to require a grind though we should probably require it to be in the highlight, which isn't currently in the rules.

Looks like it was, I probably timed it and then had an off by 1 error

I feel like getting the moonstones makes more sense to me than all of the items, but I don't have an objection to adding one or both categories if there is interest in running them.

I have moved the old runs and verified your new one.

That was the idea, yeah. just because that would make it harder to set up the RNG for a manipulated run.

I guess this leaves a question of if I should move over runs older runs that didn't have the game over, but were before people started doing manipulations. I guess I'm thinking I should move them and just make a note that older runs were grandfathered in

I have created a new category as discussed here, and included the previous game over requirement.

I was unaware of those details for Dragon Warrior and Final Fantasy manipulations. Interesting info

I thought about splitting the categories when people started submitting manipulated runs. At the time I was thinking the same way as NESCardinality. I think it would be possible to make a run where you aren't following a top level RNG route, but you are using knowledge of the RNG to your advantage (and not revealing it).

I think that is still true, but I think you could argue the mashing through the start, the decision making, and the time should give you a strong hint as to whether RNG was manipulated. I'm also not sure I'm convinced it is a prestigious enough category that someone would go through the effort, and therefore the honor system may be sufficient.

Especially since NESCardinality is part of this discussion I wanted to mention a few other examples of 'no RNG manip' categories. It seems like (to a lesser degree perhaps) Dragon Warrior and Final Fantasy have the same problem. Dr. Mario definitely has the same problem, and perhaps even Tetris. It seems like in these other cases the rules just state that manipulation isn't allowed and that seems to have been sufficient.

I'm inclined to make a new category, but I'd like to allow time for other opinions or rebuttals first.

XtremePrime (McDonald's series mod) messaged me (M.C. Kids mod) directly suggesting NES M.C. Kids should be a part of the McDonald's series. It is an official McDonald's title, so I agree, I think it should.

Fixed. I checked the numbers on the time when I verified, but not the units apparently. I'll try to verify your new one when I'm home later

Theoretically it is possible, I think the current TAS may be using controller inputs. If you used the continue code it also wouldn't be too bad.

Trying to 1CC with just the controller probably isn't realistic though. The movement is too slow for the crazy ball speed. Even manipulations would be much more difficult, a 1 frame tap still moves the paddle 3 pixels, so some pixels you can't even get to. With the vaus controller you have pixel perfect control

I think the stuff Bissmark posted seems correct. As for categories I think all stages and all stages (warpless) are appropriate. 16 stage also works, but I still like all stages. The warp doesn't actually skip any stage it just ends the one you're on, so you still visit every stage.

I guess we can agree to disagree on rounding.

As for the ending there is a better way to try it out than just testing a lot. I checked it out with the TAS tools and the results are interesting. If you are on the last pixel and deliver the last blow by hitting B and down on the same frame the game will blue screen and hard lock right after the TV host appears. Press down any later and the end cutscene plays as normal. The fade to black happens on the same frame whether you suicide or not. I'm guessing suicide being the way to go was just for fun not to save time. I think all of this supports last hit as the place to end the timer. Even if you are foolish enough to double KO, the cutscene does start, so I'd say that counts as completing the game.

I guess I don't recall if the 'last hit' was in the rules when I became a mod, but I'm guessing it was.Based on your last statement a way to shortcut either character control or the end screen transition tells me that last hit is probably the most appropriate thing. The concern would be if there was a way to deal the final blow and then die, thus managing not to finish the game after all. I'm assuming this can't be done, but I don't think I've checked, maybe you tried it?

As for the timing, I'm pretty sure I re-timed Murphagator's run along with the other top runs when a number of us came on board. Since there wasn't a need for ms I timed to seconds and I hate the convention of rounding down. To me if your run is really 14:12.786 a claim of '14:12' is misleading. There is definitely a period of 14:12 where you have already started and have not yet finished.

I guess if we find that you can fail after the final hit maybe it makes sense to change to screen transition, though I suppose it won't affect the times much.

This doesn't have anything to do with the timing, but I guess the thing I was thinking about the other day was the enemy AI. If you could figure out a way to keep them from ducking that would probably save quite a bit of time. May be worth doing some research.

I'm open to discussion, but my instinct would be to say no. Long intros can be annoying, but I don't know that 20 seconds is really overwhelming.

I'm sure I spent some time looking at the memory and the game code when I made the tas, but even if there isn't a direct influence on RNG there are better reasons it shouldn't be allowed. Most notably there is no way for someone verifying your run to know exactly what state you're starting from and if you may have modified your save state to your advantage in some way.

Hopefully the intro doesn't completely deter you, I look forward to seeing some more runs.

I submitted that virtual R.O.B. run, but I prefer not to verify my own runs if I'm not the only one on the leaderboard. Can you verify it MrJanwar?

Based on the tree shape of the map and the image I linked below I'd argue that finishing the desert is about as good of an any% as you're going to get. You are correct that there is no way to loop back to area 1, but at that point you're forced to loop back to somewhere. It could also be argued that a run through the boat area (9) should be included. A score run would also be legitimate due to the arcade nature of the game.

http://s16.photobucket.com/user/TheCarpenter/media/ONPGSpyHunterPic.jpg.html

This game was picked to be a part of Big 20 #8 and due to the difficulty and confusion about the goal choice for the race I ended up doing a bunch of research on how the game works. As a result we ended up changing the goal for the race, but the research also has implications for the leaderboard categories.

The original goal for the race was "1 loop", but as it turns out the game branches (sometimes backwards) rather than loops. The goal was updated to cross 4 double bridges. This seems good for the race, but I don't think it makes sense for an any% category. As seen below the map is really sort of tree shaped, and ares 7 and 8 (desert areas) are really the end of the branches. I would argue that the end of the desert (either one) should be the end of any%.

I'm guessing the first question will be 'what about the snow?'. Snow palette and physics happen on top of existing areas (e.g. Area 3 after finishing desert) depending on multiple criteria, one of which seems to be the least significant bit of a frame counter (e.g. 50/50 chance of snow). The fact that the snow is random and also not a unique area probably mean it should not be included in any%.

I also think an 'all stages' category should exist. Players would be required to visit each of the 10 areas and the timing would end when they finish the 10th. I do think a snow area should be required for this category. Even though it is random, I think it should be considered as part of the full game experience. Also since the water area (area 9) loops to itself, but there is no good visual indication I think it should be sufficient to enter and exit the area once (the alternative would be to require a loop and a half).

Here is the mapping information I came up with after checking into the game with TAS tools and watching the pointer to the map while testing everything out.

Spy Hunter Map Area Num, Left path, Right path, length(segments), description Area 1, Area 2, Area 3, 73, green Area 2, Area 4, Area 5, 100, green. pink trees Area 3, Area 6, Area 5, 100, green. blue trees Area 4, Area 8, Area 5, 100, brown and green palms Area 5, Area 8, Area 7, 100, dark green and palms Area 6, Area 2, Area 7, 102, light green and palms Area 7, Area 2, Area 3, 132, desert Area 8, Area 6, Area 3, 132, desert Area 9, N/A, Area 10, ???, water Area 10, Area 6, Area 3, ???, desert after water

Thanks for your response. I generally like to be pretty liberal when it comes to categories and what constitutes a speedrun. When describing the controls the manual says the start button is used "After you have completed a phase and restacked the blocks, press the start button to add up the score and move onto the next phase.", and later step 5 of the direct mode description says to "Press the start button immediately after the restacking has been successfully completed, and the current score will be added up. Then challenge a new stack.". I'd still argue that the current category in no way does these things and probably shouldn't exist. All that said if you feel strongly about keeping it I'd suggest moving it to a 'Misc' category. With that in mind I will at least submit a run.

As for other categories I'd like to request a Virtual R.O.B. category (I did a run here ). The movements were made that would theoretically stack the blocks correctly (if R.O.B. operated perfectly and quickly). I would note that I used a Virtual R.O.B. GUI to track my movements (which I put on Github here https://github.com/pgrimsrud/VirtualROB). I would note that I'm not really stuck on the naming of the category. I think it would be a good idea to also add a R.O.B. category just in case, but I don't have the blocks or a run for that.

I did verify the runs you submitted, always nice to see a new runner. I do have a few suggestions though. First, and most importantly, please include the power on or reset at the beginning of your submissions. If you find it helpful you can start your splits at -1.11, the time it takes to load the first stage. There isn't much activity, but if you use discord we have a channel here https://discord.gg/6A9vh7Z. Last, and least importantly, I'd recommend fixing your aspect ratio. Not something I would reject a run for of course, I just personally dislike a stretched screen