Forums  /  Paperboy series  /  Paperboy (NES)  /  Split categories

With the flash cart and emulators ability to manipulate the house pattern, I feel the categories should be split. One being "Original Cart" the other being "Flash Cart / Emulator". What are everyone's thoughts?

NerdyNesterNerdyNester likes this. 

If something has an advatage of the other and no way to reproduce it on all types. Then a split does seem like a good choice. Especially to runners that run OG cart and console. All about fairness to me.

Peanutfan22Peanutfan22 likes this. 

House order manipulation is not caused by a flash cart or emulator, it is caused by starting with a fixed, predictable outcome and then setting up your deaths. If done the same each time you can start with a certain house order after game over.

Everdrive, Original Cart, and Bizhawk give the same results upon reset on a Front Loader. This key to establishing fixed predictable outcome after a reset is the Front Loader, as the Top Loader handles its resets differently and completely randomizes outcomes upon reset, which is why runners with the Front Loader or using Bizhawk emulator is greatly preferred. There are no advantages between Everdrive, Original Cart, or Bizhawk, and all three can easily reproduce the same outcomes.

Read the following article and let me know if you have any questions about how this works. https:/​/​www.​speedrun.​com/​paperboy_nes/​thread/​ahf4e

When you reset in Bizhawk you will always get the same outcome, where as on Front Loader, you will get this same outcome most of the time. This occurs because of PPU/CPU alignment. The NMI will sometimes occur at a slightly different time so you see two different outcomes. But most of the time you can get the same outcome at Bizhawk. As for FCEUX and other emulators that do not emulate the PPU start up at the same time as Bizhawk will yield a different house order but is predictable upon every reset. Bizhawk actually correlates to Front Loader outcomes and so is the perfect tool for learning the game to play on Front Loader. To learn more about the PPU/CPU effect read the following article: https:/​/​www.​speedrun.​com/​Who_Framed_Roger_Rabbit_NES/​thread/​e1ues

The reason this occurs is because of nmi/vblank, the popular emulators have never bothered to emulate it in the same way that most front-loaders seem to work. This is a CPU/PPU alignment issue.

In a real NES the CPU/PPU are two separate chips that run at different speeds and communicate with each other, such that they both have to boot up, the cpu might end up booting when the ppu is a few cycles ahead or behind. This often doesn't matter, unless the game executes code that will change state when the NES is finished rendering a frame. Whenever the NMI occurs any code the CPU was running will be stopped and the interrupt code will run. In a game like Donkey Kong, they execute their RNG code in a loop until the NMI. This causes the RNG to differ based on the number of operations that occur on a given frame. If the CPU/PPU startup is always aligned as it is in emulator, the NMI behavior becomes more predictable so you can do things like get a more consistent RNG manipulation at startup.


However, since we can very easily align Bizhawk and Front Loader outcomes there is no substantial reason for separating Emulator from Original Hardware. And since Original Cart and Everdrive yield the same results on a Front Loader because the Papreboy rom fully initializes the RAM and therefore there are no RNG-related uninitialized memory addresses, the outcomes are always identical.

Once again let me know if you have any questions on how this works with Paperboy.


@Peanutfan22Peanutfan22 If you are interested in how to manipulate house order on Front Loader I am more than willing to show you how you can test outcomes, and use emulator to test contingencies, so that you can try to find a faster route.

First, If you plan to use Original Hardware then use the Front Loader only.

Second, You can use the Original Cart or a Flash Cart, it doesn't matter.

Third, You will use Bizhawk to test outcomes and to explore the game.

Fourth, using the initial default house order in Bizhawk you will begin to create a spreadsheet. This will require many, many hours of data collection. You will watch memory address values, as shown in my article, to ensure a different outcome, document how and where you died to get that outcome, and then test whether the house order has a dog house at the end. If not then move on to the next one.

Fifth, once you find a good starting house order then you will play it through testing all contingencies, meaning that you will utilize all combinations of "orb manipulation" which is the breaking of the small orb which will change outcomes. Also keeping in mind that there is usually only a 7 frame window to break the orb and that whether you break the orb during the first few frames will yield a different result than breaking the orb during the last few frames. While this will not always be the case, it very well can be something you will need to pay attention to while exploring contingencies.

Sixth, once you have found a good route then you can do basic lag reduction strats by playing out each day in preparation for runs. Make notes for each day if you need, always remembering that you must manip the dog to be at the dog house and run into him at the optimal moment.

This will require a lot of data collection, trial and error, and save stating all possible contingencies with all orb manips for all days. As soon as you reach a dead end then you are back to death manips to find the next house order with a dog house at the end.


My apologies. I was mistaken and unaware this was doable on OG hardware with OG cart. Thank you for the clarification. Since all work the same there is no reason to separate categories. I withdraw my proposal. 🥜💚

Chambers_NChambers_N and threecreepiothreecreepio like this. 

No worries, happy to be of help in any way.