Posts for PlayPatrice

Posted: 2/10/2020 11:06 PM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
The first watering can means almost 60-80 frames per square per day plus movement and re-filling. Cash gold wise, it's fewer frames per GP to grab mushrooms. If you are trying to consolated into earliest DAY, then yes, spring farming is good. If you are going for fewest frames for startup capital, you harvest in the fall, and use the sprinkler / horse in spring / summer years 2 and 3. In a TAS you WOULD overplant spring and summer crops to match movement to the shipping bin, because you can force rain RNG to completely ignore the watering can / sprinker all together. The horse is the big time saver, no need to run to the bin.
Posted: 2/7/2020 12:16 AM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
Poision mushrooms only show up in the Fall. The only "crop" that grows in the fall is grass. So you are limited to eggs, milk, cave herbs or mushrooms for incoming cashflow.
Posted: 1/25/2020 9:05 AM
Post subject: THANK YOU
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
Thank you zin. That's going to make that a LOT easier.
Posted: 1/3/2020 8:25 AM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
*Victory!* Last year I had a hard drive failure that cost me all the work I had done previously on this particular TAS, but now a year later I've come back to it to continue banging away at it. Without understanding what causes the Fishing RNG to set to a particular interval (Time to Fish), I HAVE figured out how to consistently manipulate the final time to fish. The 60/20 rule: For every exact interval of 60 frames waiting on a "Pause" screen, such as dialogue or select menu, you will receive a reduction of 20 frames off your total time to fish to a minimum of 40-60 frames. 4 cash fish on day 1, is physically impossible - even for a Tass (I've been trying). 3 cash fish and the fishing power berry on day 1 IS possible, if you keep your time to fish around 50-60 frames per fish. You will have to bin the first fish and get a "Screen Transition Lunch Skip" - Bin the second fish, Sprite the Third Fish, and rush back to bin the 4th fish. This gives you 900 gold, and the potential for the maximum 4 power berries on day 1. This is important for options in stamina management and cashflow planning, opening up options that are not feasible to RTA runners. As it currently stands, the Hardwalls are really nasty on day 1, and I spent a grand total of 7,426 frames of RNG manipulation to pull this off. I'm certain there are faster "Seed Conditions" to cut this number down considerably (At least in half), but without understanding or being able to debug the code (lack of skill), It will not be found by me in the near future. A typical round trip up the mountain costs approximately 2.8k to 3.2K of frames to go up the mountain, starting from the house, and then to the shipping bin. But we are getting an extra 10 stamina starting on day 1 and maximum starting cash. If I can shave an additional 2-3k in frames off the 7k total, I believe the 4 fish start would be optimal. At the very least this is a good piece of kit for those that try to TAS "Earliest possible day to have X-Y-Z."
Posted: 4/11/2019 6:01 PM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
All right. Thank you for the reply. Thank you for the link. ^.^ I will see what I can do.
Posted: 4/11/2019 12:03 AM
Post subject: How Perfect does RNG have to be?
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
For the last few weeks I've been struggling with Harvest Moon's (SNES) Fishing RNG. The goal of the RNG is pretty simple: Get a grand total of 9 fish into the shipping bin by the end of Day 4, while collecting all the Power Berries. By doing things this way we cut out nearly 4,800-5,600 frames from the TAS by not having to make an extra trip into the mountians to collect either the Chicken/ Stump power berries during the night, or put off collecting the water sprite power berry until after we hit our Gold totals needed for cash crops. The RNG is important here as we have to manage it in a specific way to preserve our daylight timer in order to get all 9 fish to the bin, while still getting everything else done. These are the only scheduled fishing trips my routing requires. Once fishing is done, we never do it again. So long as we spend less than 4,800 frames directly manipulating the fishing RNG we are considered to have saved frames or come out frame neutral. The fastest result I've been able to achieve spends has roughly 620-680 frames that can be directly attributed to getting fishing RNG seeds to line up: 1) If my frame counts are correct there is another potential 600 frame reduction thats possible until ideal / optimal conditions. Theoretically, there are up to 800 frames that could be squeezed out of it with "God Mode" RNG manipulations. 2) My fastest "Fishing" solution of 9 fish and all four power berries uses a sub-optimal startup. There is a 16 frame minor pathing mistake, and two swag name entries, roughly 15-20 frames each over optimal. So roughly 35-50 frames could be shaved at startup. An optimal 1st day pathing causing some bad fishing seeds starting with the second fishing trip on Day 2. To get back to the good rhythm I know is there I would need to add in a 50-60 frame pause before the end of day 1 if I "Optimized" the visual appearance of Day 1, by correcting the 16 frame pathing issues and only using "A" as the names for jack and the dog, (As opposed to TasB and DwAC recpectively). In short I've allready spent nearly 100 hours trying to brute force faster fishing solutions. Of the 5 total solutions I've developed (All 3 days of fishing) the best solution I have is 1,400 frames faster, but has a 500 frame gap that could be trimmed out. Unfortunately I'm pretty sure at this point to get past that gap would mean exploring other entire fishing sequences. Meaning the entirity of all 9 fishing attempt would have to be re-explored, and each fishing attempt takes 30 minutes to an hour to optimize, and gets worse if we have to manipulate earlier attempts / timings to get a better seed. My question is - At what point can RNG intensive sections of a game be considered "Good Enough"? Do I need to use my "Optimized" day 1 solution as opposed to my Non-Optimized "Swag" start, even though the SWAG start gives me a better RNG sequence on day 2? Without help from someone in terms of RAM watch or RNG reverse engineering, spending another 100 hours trying to brute force a solution that /Might/ be faster than what I have doesn't sound like an effective use of my time considering I have another 1.5-2.0 hours of gameplay I need to do frame inputs on.
Posted: 4/9/2019 8:02 AM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
These are the known timers that can cycle when fishing. Fishing 1 I think is a simply "Play Animation" state that only cycles when jack performs certain actions. 255 is him sitting a the docks, and there are count down timers that coincide with each of the fishing animations. Sometimes fishing 2 cycles, sometimes it doesn't. But I have seen no correlation between it and the frame you can pull fish up or the time to fish on the next session. Fishing 1: 7E02AA Fishing 2: 7E08F5
Posted: 4/9/2019 7:58 AM
(Edited: 4/9/2019 8:05 AM)
Post subject: Ram Search / Watch - How do I find Permutative RNG?
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
So I'm working on a TAS of Harvest Moon (SNES). For those that don't know, it has a fishing mini game with some very difficult RNG. In a nut shell, with proper RNG values I can prevent an "Extra" Trip up the mountains to handle item collection, netting a frame savings of 4,800-5,600 frames. I've been trying for nearly two weeks to brute force the propper event with some results. I started with some 2,300 frames in fishing RNG manipulation, and have steadily been working it down. But now I'm running into a brick wall that I need help with. Namely in working with, and understanding RAM watch to find the values or events that cause fishing RNG to permutate out of controll. So far my best outcome is I spend approximately 650 frames in dialouge screens to achieve the results I need. But under the right conditions it could be very easy to cut down to less than 100. Another 600 frames could be potentially shaved off the mini-game itself with "perfect" fishing. Changing the time to snag a fish from an average of 150 frames, down to a minimum of approximately 25-35 frames. Some of the Animation Timers for the fishing event have been found for the game (which I've posted below), but I can't seem to work or understand the Ram Search function to find what triggers the time required to get a "Snag" on the fishing line. Here is what I know (Or rather, very seriously suspect): * A Fishing "Hardwall" script may exist. If you try to fish too quickly, or show up on the mountain within a certain period during the daylight timer - you can get hardwalled. Creating a situtation where the game will not allow you pull up a fish under any circumstances untill a minimum total number of frames has passed on the system clock. This can be bypassed and returned to the normal RNG by waiting inside a building or inside a convo screen. * Fishing seeds for the NEXT trip up the mountain seem to be determined by something before going up the mountain. Once on the mountain, the seed seems to stabilize. * Even if frame counts remain the same, both when you put your rod in the water and when you pull up your fish can change the fishing seed the next time you start fishing. * The "Minimum" time to fish can be manipulated to some extent by changing the fram at wich you touch the docks. * There is an "On Hit" frame perfect chance to shorten the current minimum time to fish that can be determined after sitting at the docks, and depends on when you drop your rod. Usually it shows up in the first 10-15 frames of sitting on the docks if this short cycle is available (Sometimes it is not). Considering all the variables involved, it is incredibly time consuming to brute force my way through these RNG permutations. If I change anything earlier in the run, it throws off all the RNG after that point (many times even if it's frame neutral, but not allways). Changing the frame I sit at the docks, sometimes changes both the minimum time to fish and for certain the "Short Cycle" frame. Changing when a fish is caught changes the fishing seed for the next fishing cycle. So changing a frame, and then searching 10-20 "On Hits" to get the necessary maximum frames to get 4 fish in one day (or rather 3 fish + power berries), can take 1-2 hours per fishing session. Meaning a single "Day" of fishing after a change somewhere earlier in the run, can take up to 8 hours to find an acceptable result, which also causes desynch issues on other parts of the run and forces you to reroll any fishing that comes after that point. So changing anything on day 2, forces me to completely re-do the 6 fishing events on day 3 and 4 (which can be another 4-16 hours of brute forcing). I need to pull up 9 total fish, and I know It's possible to shave another 600 frames off this sucker and theoretically, move a power berry from Day 4 to Day 2 (which changes my current routing optimizations). In Ram Watch I've tried searching frames against a known value (# of frames required to snag a fish), but either I don't know what I'm looking for, or that's not the right variable I'm looking for. How do I narrow down and find the values and trigger variables generated throughout this Fishing RNG process?
Posted: 4/6/2019 2:58 AM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
Frame Costs for Crops: After doing some extensive spreadsheet monkeying, here is what I came up with: Numbers are Gold per frame spent Fish 0.06G No Horse At an average of 6 squares of movement per crop (up to 12 squares away from shipping bin). With Rain RNG, we do not have to water crops: Up to 85 pieces of day. Turnip - 0.53G Potato - 0.70G Tomato - 0.88G Corn - 1.06 By Dropping the distance down to an average of 5 squares per crop, we go up to 92 pieces. 3 squares (Double Row), you can nail up to 108. At a distance of 9 squares (18 away), you could get 70. Actual results may vary. But basically, you could very easily TAS your way to 10, 3x3 crops in dense formation, to be harvested in a single day. With Horse Optinal Horse Farming (Assuming 1.5 squares of movement for Horse whistling/positioning). Up to a theoretical maximum of 125 per day. Which is 20 bags of 3+3 crops in parallel per day. Turnip: 0.77G Potato: 1.03G Tomato: 1.29G Corn: 1.55G For a Ranch Master 999, We need at least 450,000G to make all in game purchases and cover the farm in grass, or 350,000G if we just cover everything in spring crops by the end of Spring 3. So - this number isn't completely accurate, but gives a good benchmark of what needs to happen. This is total frames required harvesting each at the assumed rates: If I have done my frame math correctly - we are looking at somewhere between 1.03 Hours of Optimal Summer Corn, or 2.6 Hours of Horseless Turnip farming to hit those numbers - just in crop harvesting. So, between that, field prep, and everything else that needs to be done in the game - we are looking at a movie that will be around 2.5-4 hours long. Optimally, we /might/ be able to get a sub 2h, but it takes almost 20-30 minutes to get past the first week of gameplay, not to mention affection management for girls and cattle... Must... find... tech....
Posted: 4/6/2019 2:04 AM
(Edited: 4/6/2019 7:20 AM)
Post subject: Crop Planning with frame Counts
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
Crop Planning using frame counts: The Clock starts at Hour 6:00:00 and is HH/MM/FF (Hours, Minutes Frames). The Shipper "Pauses" The clock at 17:00:00 as the shipping deadline for the day. At 12:00:00 you are forced to wait through a 98 frame "Lunch" Animation. It's 60 frames per minute, 15 frames per hour. 11 Hours to the shipping bin (Shop Close), 12 hours until daylight pallet swaps stop (and it's nighttime) Any time you "Leave" The Path map screen (and only this screen) - time is incremented by 1 Hour. (A loss of 900 daylight frames. So going to the mountain (or town), and coming back, automatically gives you a loss of 1,800 frames in travel overhead in addition to whatever time you spend in those maps. The Clock "Pauses" when starting a screen transition, and returns when you gain control of Jack. Occasionally there are 16-32 "Dead" Frames after entering a screen. This frequently happens in the late afternoon, or when you are "Fishing" to quickly and trigger the Fishing RNG hardwall (It's the game stalling you). When NPC's hit a "Scripted" movement tick, you will experience 1-2 dead frames of no input (but time keeps ticking). This often happens in the late afternoon when the woodcutters are removed from the map. NPC movement can still happen when dialogue screens are open. Meaning if you change your convo timings, you can still desynch by a frame or two if an NPC moved somewhere via a script on the map. When the ambient daytime pallet swaps occur, you can loose 1-2 frames every 10-40 frames. On average you will loose somewhere between 10-20 frames to these frame transitions. After Hour 17:00:00 the pallet changes happen rapidly, and you can easily cost you 7-10 periods of "Dead" frames of 1 to 2 frames a pop as the colors change. This can cause re-synching issues if you changed anything earlier in the run. Other than scripted events, the daylight timer is only "Paused" when you are inside or have a conversation window open. == What does all this mean? == This gives you 9,800 frames of "Daylight" time to get things into the shipping bin. A single "There and Back" trip to either mountain or town costs you 1,800 daylight frames in overhead. You can Use nearby conversations to manipulate RNG without loosing daylight time. * Time to pick up a vegetable is 24 Frames * Time to throw a Vegetable into the Bin is 34 Frames. * A single Square of "Running" Animation is 8 frames of holding down the B button. So for maximum harvesting, you can assume 65 frames per piece + Movement. Giving you a maximum theoretical - zero movement harvest of 150 pieces per day (9,750). Realistically, you will need 10-20 frames of movement just to get to the starting area of your crops (or grab the horse). Without the horse, every square you spend going to the crop is a square you have to walk back. If we allow for an "Average" of 5.4 squares of movement to collect and throw, we can get up to 90 vegetables during Spring (10 plots). This calculation takes in the idea that vegetables will be from 0 to 2 squares away from throwing distance into the shipping bin. (Giving you an average movement distance of 4 squares per piece). Considering we have to go a little further out, we start being able to harvest 1 or 2 plots an additional set away. Much less if you have to start pushing further out than a 3x3 grid directly adjacent to the shipping bin. So summer "Row" farming with a horse, maximum, at 138 frames per 2 pieces is 140 vegetables, or 11.6 plots of 3+3 rows.
Posted: 4/5/2019 1:27 PM
Post subject: Desynch Caused by "Dead" Frames and where to find them.
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
So, in re-optimizing my opening sequence, I found that many non-RNG related tasked kept desynching. I would go and fix those areas only to find other spots show up completely desynched. Considering the rather "Passive" nature of this game, I was extremely frustrated until I started understanding where it is coming from. The game engine systematically freezes or ignores inputs on a periodic basis. I assume it is cycling or checking it programing stack, or executing a very specific part of its code. Here are the predictable points I've found: * Every time the Daylight clock cycles by 1 hour after 10 am, you will loose 1 frame of input. * During the atmospheric lighting pallet swaps - you will loose 2 frames for each time the color shifts. This is VERY pronounced in the late evenings just before it becomes completely dark. Be prepared to go frame hunting if you make a sequence change that messes with anything after the shipper comes by the house. It can get pretty bad. Once it's completely dark, these "Dead" frames disappear. * Lunch and Exaustion animation - while predictable - will be required to be accounted for when re-sequencing or optimizing anything earlier in the run. They are rather lengthy, but can be handled by "inset" cut and paste. Just remember to remove the old "Gaps" to prevent frame bloat and additional desynching. * Any time an NPC has a "Scripted" movement. This is different from random movement. For example - there is a script for both the Dog and the outdoor Wood Cutter, that sometimes, if they are near your expected path (or they were placed there by the game), once you get within 2 squares of them, they WILL move to intercept you. When this script executes, you receive dead frames just before they move. This scripted movement may also explain the sometimes random dead frames I got when I left my dog outside (He was being script moved off screen). I put the dog inside the chicken coop - and my night time dead frames disappeared. This may also be why a 12 cow barn is so laggy. That's 12 cows constantly rolling for script movement. And why you get so many dead frames late game if you leave your livestock outside. * You are doing to much on Day 2. This one is real. If you go to the Shed to grab tools and thus start the clock before rushing for the mountain for your starting fish - the game seems to "Know" what you are trying to do. Just after screen transitions it will insert 16-60 dead frames at each transition in order to slow you down and prevent you from getting 3 fish and an additional power berry(s). If you just rush for 3 fish, you get your 3 fish and this "Dead Buffer" isn't added after screen transitions. It took me a while to figure out why the pathing sequences would all be horribly out of whack on Day 2 until I realized what the problem was. If I went to the shed first, or grabbed a power berry - these 16-60 frame buffer windows (In addition to horrible fishing RNG Hardwalls) that would show up and eat away at my daylight frames. Thus preventing me from getting an "Ultimate" first working day. It threw off my "Frame Math" that I had planned out using day 3 and 4.
Posted: 4/5/2019 12:59 PM
Post subject: How To Manipulate "On Hit" RNG
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
Other than fishing - RNG in this game seems to be biased on some random number generator that is tied to the Master system clock and is fairly easy to manipulate. * Once you have entered a map - the RNG seed is "Set". Changing the time you enter a map will change the seed. So any alterations to shave frames off an earlier segment will desync everything RNG related after that point. * to "re-roll" a on hit event (Power Berries from chicken, stump, soil or gold bags / gold coins / moles), is a simple matter of changing the exact frame you press the correct input (usually Y). Previous controller input, or excessive controller inputs on that frame does not change the result. * You can 're-roll' a result by simply activating the appropriate input on the next frame. Then sequentially go down the line until you get the desired result. * I found out (accidentaly) that hitting a different plot of soil on the exact same frame generated the same result. So instead of a bag of gold from the soil above me, I got it from the square to my left. * You can "Explore" The RNG results without loosing pathing frames! For example: When trying to get a Tree Stump Power Berry you can simply set up your pathing so that you strike the stump at maximum pixel distance from the stump. If after the final hit, you don't get a power berry, rather than insert a "Blank" frame, insert a movement frame, but ensure you input happens on the next frame. You will take one additional step, and still hit the stump without loosing any frames to dead space. Using this technique, I nearly walked all the way around the tree stump from righ to left, then from bottom to top looking for my power berry on the way to the Fisherman. I got it on the 3rd to last frame of available movement. But the point is - I got it, without having to "Wait" the 15 frames I would have otherwise had to do had I used blank frames. Things that are definitely "On Hit" RNG wise - * Weather and some Special Events, Dog Bark RNG (On Hit is when you open the Journal, if the journal dialogues are all equal) * Coin Drops From Soil * Gold Bags From Soil * Power Berries from Soil, Chicken, and Stumps * Moles From Soil * Romantic Conversation Questions from girls (If it's the right time / place / season for such questions) == More Notes on Fishing RNG == Fishing RNG works VERY differently than this. Fishing RNG seeds seem to be baised on when you enter the mountain (Daylight time) compared to either whether or not you have allready dropped off a fish - or did something else in the day. I find the less I do - the overall shorter my RNG times to fish are. When I start trying to cram in an extra wood chop on the first trip - or route in any of the power berries - the Fishing RNG tries to punish me by setting up hardwalls or extremely long times to fish - that are difficult to get passed even when waiting inside frozen time conversations with the fisherman.
Posted: 4/5/2019 12:39 PM
Post subject: Routing Power Berries with 3x Cash Fish
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
So I had to go back and redo my only 3 days of fishing for route optimization. Since I knew what I was doing it went a LOT faster than the first pass. And my original frame estimates / maths were wrong. As of right now I do not see a way to get all three Power Berries and Three Fish for the Shipping bin without making four trips. But you CAN split the Power Berries between two days. Grab the chicken and stump on day 3, and then the Fishing Power Berry on Day 4. But long story short: On Day 2, the Fishing RNG is very brutal if you do anything other than go straight up the mountain. When combined with the overhead of coming in and out of the woodcutters hut - hitting both chicken and stump - or the Water Sprite Power Berries - It's far more frame efficient to just blitz the first day of fishing and grab the chicken and stump on day 3. So, with frame optimized movement, each trip up the mountain for just fish will cost you somewhere between 2000-2500 frames of daylight (not including the 1,800 frames lost due to the hour incrementing from screen transitions when exiting the path). == Chicken and Stump == The Chicken + Stump has 576 frames of total daylight frames routing overhead (Along with a non included 62 frames of frozen time dialogue to mash through - Pushing it up to 630 total frames spent on route divergence). Basically, when the math is done, you have 570 frames from the time you sit at the docks to the time you snag a fish. This 570 frames needs to be divided amongst all three fish, and it gives you a zero frame margin of error. This leaves with approximately 190 frames per fish (From sitting to dock), and have 0 frames left over. For safety, you should probably aim for no more than an average of 150 frames per fish. == 4 Fish - 1 Day == You /CAN/ get the fishing power berry and 3 cash fish in 3 trips up the mountain. After catching a fish and throwing it in the pond - you can go right back to the docks and cast your line again if you don't talk to the fisherman. The Water Sprite represents 691 frames of daylight overhead (with 120 frames of frozen dialouge). On the day I pulled this off I spent 529 Frames on fishing with 157 frames remaining at the end of the day. This leaves you with 686 frames for 4x fish. Or roughly 171 frames per fish. Again, for safety I would not recomend going over 150 frames. == The Fring Theory theoretics == IF - and I do mean IF we could somehow manipulate or predict fishing RNG (using paused time in dialouges as I've been doing) - you could manage all 3 power berries in one day. On the 4x fish day, I had 686 frames available for fishing. The daylight frames overhead for Stump and Chicken is 570. Leaving you with 116 frames for 4 fish. Or 29 frames per fish. The two lowest fishing cycles I've stumbled on was a 16 frame fish (which I never found again) and a 24 frame fish. With most fish being either 100 or closer to 250-350+++. So that means on Day 3, over the course of exactly 3 trips, you could pull all three power berries AND three cash fish, without the need for a 4th extra trip or splitting your power berries amongst two or more days.
Posted: 4/3/2019 11:25 PM
Post subject: 3 Fish RNG + Power Berry - Early Starting routing options.
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
Here are the "Practical" Rules I've figured out for Fishing RNG that allows us to net 3x Gold fish and the Fishing Power Berry on the same day, With the -possibility- of grabbing the power berries for Fishing, Chicken, and Stump on Day 2 (First "playable Day), without needing to make a separate specific daytime trip for the Fishing berry, or a night time trip for the stump and chicken. Or basically, 3 Berries, 3 Fish, Only 3 Trips, all in 1 Day - Either at a net reduction in total overall frames, or at least breaking even (Which helps with Calender Management). Unfortunately, brute forcing the correct RNG timing is considerably time consuming, but if I can get a little help finding the correct ram-watch values we can nail down this mechanic very easily. 3 Gold Fish and the Power Berry is very doable - and I've got a routing solution that accomplishes this on Day 4. But for optimizaton purposes I want to move this up to Day 2. Either for a 3x Gold Fish - Power Berry + Nighttime Trip (We route in 250 wood.) or we handle all the berries in just three trips. I would prefer just three trips because I can route in the wood with a single stamina refil on Day 5 for Crop Calender managment (Need to water at least half my crops). But - all that shinanigans aside - Here are the practical "Rules" To RNG Fishing - A full "Round Trip" from the Shipping Bin-Fish-Shipping Bin costs us somewhere between 2,100-3,000 frames. A decent 200 frame "Fishing" cycle will land us somewhere around 2250F while my fastest fishing trip was 2087F. Doing an average of a 350 frame fishing cycle will net you 3x fish and plenty of time to do other things (Either grab the fishing sprite, or both the Chicken and Stump Berries). And there is still a little bit of wiggle room left. The water "Sprite" sequence, from the time you snag your fish - eat the berry - and drop your line agan - is 820 +/- Frames. Giving us 1,280 frames of RNG "Overhead" in order to break even compared to the "Average" Three fish 200frame cycle. A break even scenario would be used for calendar management, but god mode optimized we are talking about a 1,800- frames in total savings that can be used to do a few other cool things during those trips - or just save time. More Importantly - if we can pull off god mode RNG manips - It would open up a lot of routing options - and more importantly - save the TAS'ers considerable time compared to brute forcing the RNG sequences. It takes 2-4 Hours to find a viable 4 fish day - and I haven't yet fount the Godmode 4 fish and 3 power berries option that is out there. But if my Frame Math is right - it's possible - But just barely. * Breakdown of Fishing Mechaics From a practical viewpoint - there are three separate fishing variables that contribute to how long it takes from [Drop Line] to [Line Wiggle]. Definitions (These are my own to help me keep things straight in your head). [DFF] Daylight Fishing Frames - This is the number of frames from the time you SIT at the dock (not drop your line) - to the time you can snag and pull in a fish. This is our metric of success and what we need to reduce in order to pull of some fun RNG optimized routing. [Sit] Sit - This is the frame you activate the sitting apixel art on the docks. [DL] Drop Line - This is the frame you hit the "Y" Button to throw your rod in the water. [Snag] - This is the "Snag" Frame where you actually pull up a fish. [SC] Short Cycle - Drop line has a frame perfect window that can reduce your overall time to fish (A "Short Cylce" if you will). It is usually found somewhere in the first 8-15 frames from [Sit]. There can be a single frame that, when compared to the other 50 frames "Surrounding" it, can cause a 50-200 frame reduction in time until Snag. Sometimes it's just one frame, sometimes its every 5th or 8th frame. Either this is an addition +Frames variable (Which is ignored on a short cycle), or its a number that reduces overall fishing frames if you hit it. Not sure which. [STF] - Static Time to Fish - This is a minimum number of frames needed. This varible exists. If you move your drop line forward 20 frames, your snag line will also move forward 20 frames if you don't hit a short cycle. This variable rotates slowly (Every 100-250 frames). Sometimes it changes a lot, sometimes it changes just a little bit. It usually ranges from like 0-300 frames - But I've seen it as high as 700. This variable is set static when you sit down, and usually takes a LONG time before sitting, or waiting before you drop you line to move it around. When you drop your line in the water. I imagine this variable should be one of the easier ones to find in RAM Watch. [HW] - Hard Wall: This Variable is the bane of my existance. It's a hard time set to the system clock of the earliers frame you are allowed to get a Snag. I don't know what causes it, or sets it in place, but if you find that you have a very LOOONG fishing cycle, and no mater what you do in a 800-12000 frame window, you allways get a snag in the same 20 frames - you found a hardwall. We can get around Hardwalls though in a practical sense. It is set to the overall system clock - and NOT the daylight time. So if you find yourself hardwalled, you can simply hit the select button or stay in the fisherman conversation so that you will sit at the docks and drop your line in closer to hardwall. If minimum fishing frames is greater than the hardwall - you have to wait out your minimum frames. If it is shorter than the hardwall - you have to wait for the hardwall to show up. I've never seen a Hardwall on the first fishing trip of the day. But they constantly show up on trips 2 and 3. The Hardwall seems to get worse the faster you are at fishing. It's like the game coders wanted it to be VERY difficult to get 3x fish in a day - and even then, you don't get anything else done. Between the "Lottery" aspect of fishing, and the coding implimentation of this hardwall - I don't think you were supposed to get 3 fish to the bin in a single day - much less 4 fish (With the Power Berry). Without finding the RAM addresses for these variables, I have found a very consistent (But time consuming) method of manipulating the fishing RNG. 1) Make note of your [Sit] frame. 2) [Drop Line] on the next frame 3) Place a [Snag Line] somewhere between 200 and 250 frames (300 frames if you are just trying to 3x fish it and nothing else). 4) run the movie forward and check for jiggles. If your rod returns with no fish, go back and move the [Drop Line] forward a frame without changing your frame count. You are looking for the short cycle (Do this even if you get a wiggle in the time frame, you can get 50 frame cycles). Repeat this for 20 frames or so and keep a result you like. 5) If you Don't get a hit or result, go back in the movie in leave the fisherman conversation open another 10-20 frames and repeat the process. You are trying to find the point that your Static time to fish changes. When you find it has changed off small incriments, it usually takes 100 frames to trigger the next change in the static Variable. Convo's pause the daytime timer, but the other fishing variables still continue to cyle at their slow pace. This allows you to save frames overall, by adding 20-40 or even 100 frames to the convo can net you 200-500 frame reduction in overall fishing time. And more importantly, preserves your time to get down the mountain to drop off your fish, allowing you to easily handle a 3x Fish. If you sacrifice a few overall frames it's not too difficult to route in Both Mountain Powerrries or the Fishing Powerberry into a single day for Calender Management. 6) If you have found you have been Hard Walled, you have to leave the fisherman conversation open long enough to get past the hard wall count. Hard walls are usually identified by 500+ frames of waiting for a fish to show up, and no matter how hard to push frames to try and manipulate them, the fish keep showing up near the same frame counts. The onyl way to brute force this without altering your seqeunce BEFORE comming up the mountain, is to leave the convo open long enough to get past the hard wall, meaning you could be waiting 500-1200 frames (8-20 Seconds) just for the chance to re-roll your static time to fish. Without RAM-Watch (Which I have no clue how to set up) - this is considerably labor intensive. But the payoffs in early game setup are potentially huge. Being able to Nail a 50 frame fishing cycle when you want would really help with early routing opportunities. For now, I'm going to try and brute force the fishing power berry on day 2, and manage the stump and chicken on day 3 (or in the late evening of day 2). It just takes a while.
Posted: 4/3/2019 10:34 AM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
Okay, that answer / explanation helps out a lot for what my run needs to focus on from here forward. I'll continue working on the run with only the Ranch Master and All Days category in mind, and worry about the other stuff / definitions after that hurdle is passed. Thank you.
Posted: 4/3/2019 10:07 AM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
RTA Best ending involves the following: - This is the exact copy/paste from both the Speedrun.com and Admirils guide. Married 2 children 900+ happiness 10 Power Berries eaten, 10,000 G own 1+ cow own 1+ chicken, hugged dog 100+ times Best Ending (For both Married and Unmarried), involves quite a few more things. (Again, see Admirals guide). And there is an even larger list of things to accomplish in a ranch master 999 game. Most notably in either of the above two "Best endings". The ranch master score does not go above 999 and points are capped in each category. Are we saying "Best Ending" is determined by the top tier of the avaible cutscenes? We play all available cutscenes, or only specific ones (Like the RTA'ers use?). But you said that only "Best Ending" or Ranch Master are the only two allowed categories for TAS'ing. Ranch master is well defined by an in game score. Whats the criteria for "Best Ending?" Are we defaulting to the RTA'ers? Or one of the other "Best Endings" also described in Admirils guide? * Edit * Thank you for responding. Your video was old enough I didn't know if you were still around on the site or not. Sorry if I come across as argumentative or what not, it's just I find the lack of solid definitions frustrating when trying to do planned routing. Which is not your fault, it's just a side effect of this game not having a large speed running or TAS community.
Posted: 4/3/2019 8:23 AM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
Techokami wrote:
Still watching this, but if you want to watch both the main and the Tails POV encodes side-by-side, you can use this: https://viewsync.net/watch?v=YvtI4gFMwlo&t=0&v=vU6J2e--hhg&t=0
Honestly, if the authors are willing to remaster a 4:9 video, with both side by side - it would DRASTICLY help in watchability and clarity in this amazing piece of work. Just watching one of the videos is a jarring, jumpy, spaztic nonsense that I had to turn off after the first 5 minutes - it gave me a serious headache. When watched side by side like this poster did above - your eyes (And more importantly your brain) can piece together what's happening much easier and naturalistically. And if the screen starts jumping around to much that it's discomforting, you naturally glance at the other (typically more stable) screen, where something significant is probably happening. I was actually able to complete the entire video with the side by side playthrough without any issues. But yeah, wow, this is definately a lot of crazy awesome tech and work. Congrats!~ * Edit *
Evil_3D wrote:
For people that missed s3k thread, Selicre maked an Atlas encode of CNZ 1+2. He is testing the dynamic camera the keep both characters on screen as possible, still have some camera bugs and is far form perfection, but the work he did is amazing.
The idea movie format would be to have Two Split screens and a widescreen on one wide display video encode. Looking something like this: |...................................| |..[--Sonic ]..[ Tails--]..| |..[Window 1]..[Window 2]..| |...................................| |..[------Screen 3-------]..| |..[-----Distance View----- ]..| |...................................|
Posted: 4/3/2019 6:08 AM
Post subject: Harvest Moon - Category Concept Check - Opinions wanted.
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
So I've started a TAS for Harvest Moon (SNES). I've also done a bit of research, checked some of the previous submissions and Forum Posts regarding this game. Notably These Two: http://tasvideos.org/5825S.html - "#5825: ViGadeomes's SNES Harvest Moon "Marriage" in 29:31.77 (Rejected)" http://tasvideos.org/3516S.html - "#3516: samurai goroh's SNES Harvest Moon "any%" in 02:18.78 (Rejected)" Bot of these were rejected for following the RTA game categories over on Speedrun.com. (https://www.speedrun.com/hm1) Namely, the use of a password and specific ending game cutscene achievements as a benchmark of achievement. So I need a viability concept check for my work in progress (allready 135k frames in) before I go much further so I can make appropriate corrections. After doing some reading I understand I'm putting together a "vault" style movie here: If I were to simplify the categories is qualifies for: Category: 100% Married, Glychless, All Days, Ranch Master. I'm making a disctinction between 100% runs and Ranch Master (Max Score 999). 100% runs (For Married or Unmarried) revolve around aquireing all the collectables and unlocking the top tier ending cenematics for each category of your ranch. The differences between a Married run and an Unmarried Run involve the deadlines of when to buy both house upgrades (as well as the deadline to get married by) and farming affection for the other girls. In a Married run, you only have to worry about one girl, but you have to schedule in your house and your marriage over the course of the game. In an Unmarried game you have to reach +/- 500 affection with all 5 girls (a considerable time sink), but the deadline requirements on the houses are lifted. Technicly, every Ranch Master (Max Score 999) is a Married Run. But it technically isn't required to be a 100% run. A 100% Married run does not require you complete all affection scores for the other girls. You are only required to reach max affection with 1 cow rather than all 12. The final crop requirements are much lower / less quirky, only needing 200+ rather than Farming in exact Lots of 500 +/- pieces. Plus in ranch master you can "Technically" miss one or two collectables (The Milker specifically) and still achieve a final score of 999. Note that these three (Ranch Master, 100% Married, 100% UnMarried) is different from the "Best Ending" category the RTA runners use. The "Best Ending" Category in this case is actually the "Happy Family" ending as described in The Admiril's guide over on gamespot. (https://gamefaqs.gamespot.com/snes/562623-harvest-moon/faqs/52336). RTA runners use this, as it gives a basic ending. The run to achieve the necessary benchmarks and then use a password to trigger the end cutscene. Their other categories also revolve around this password. This cuts down a game that can takes somewher from 3-8 hours to complete, and cuts it down into a manageable 2 minute, 30-45 minute, or 2 hour run. RTA Categories: Any%: [2 minutes] - Complete Day 1, Use password. - Involves the micro pathing of day 1 to achive the lowest possibel time. Marriage: [40 minutes] - Earn 6,000 gold, Upgrade your house, get date and marry as quick as possible. Activate Password. All Days: [43 minutes] - Activate the daily journal 330 times. (Very boring video). Best Ending***: [2 Hours +] - Do the minimum activities required achive the "Happy Ending" as described in Admiral's game guide*. Activate Password. I bring all this up - because as I said the last two videos were rejected baised primarily on their concept - trying to directly translate the RTA game standards over to TAS. So I'm trying to put together this TAS to meet the concept requirements implied by previous rejection notes. The hope, is that if I can get it finished, it might spark interest and expand category concepts for more TAS'ers to attempt this game (and allow for a template to canibalize for other TAS'ers to use in attempting other categories). Harvest Moon (SNES), has a lot going for it in the optimization department and requires multiple layers of routing, Cash, Stamina and time management, and it's a farily simple game to manage frame inputs on (some exceptions aside, gosh darn fishing RNG took a while to deconstruct). So, If I were to finish this tas - here is specifically what I am aiming to do in as few possible frames: Category: 100% Married, Glychless, All Days, Ranch Master. * All Days (No Password) * Glychless (No Major Glychyes?) - No Power Berry Duping, Maricle Potion, or Overflow shenanigans. * Married with 2 Children * Aquire Every Collectable Item * Play the "Top Tier" End Game Cinematics - For Each Category * Hit the Maximum Ranch Master Score of 999 * Trigger Every In Game Cutscene at least once (Girl Events, Beanstalk, Etc.) * Minimize all cash earning activities (Only 3 days of fishing needed, minimum crops harvested in appropriate lots - Rain RNG manipulation, so minimum watering needed, etc.) * Minimize the number of frames needed, and maximize stamina efficiency, good routing, calendar management choices - etc. Am I on the right track here for a viable TAS submission? I'm NOT asking "If I make a video like this - will it be approved?" So I'll post it one last time: Category: 100% Married, Glychless, All Days, Ranch Master. What I am asking - if I finish this video, is it reasonable to assume that this can be a viable concept? Would a video meeting these criteria recieve appropriate consideration baised on the TAS's merrit, rather than being rejected based on it's concept? Does it make it to the "Starting Line?" Or do I need to adjust the overall goals here? I would say I'm somewhere between 1/4 to 1/3 of the way throught this and I'm pretty excited with where it's going. I just want to make sure I can make any appropriate conceptual adjustments needed to make it to the starting line. Thanks for reading. ^.^ - Pat
Posted: 4/1/2019 5:52 PM
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
feos wrote:
TAStudio menu -> Config -> Set autosave interval -> Enter 0. Otherwise, what would really help me fix this is if you record a video of how you get this crash from fresh unzip of 2.3.1 release. Maybe reducing the Capacity will help to trigger it sooner so you don't have to record too long a video.
When you say record a video, What program do you recommend to record the screen capture. I am assuming you don't mean to send a copy of my TAS. What other computer spec information will you need and where do you want me to send it? Next up: I set the memory values to max, with the save states on the longest interval possible as well as the Autosave interval to 0 (Thank you for letting me know where that was). I am MUCH more stable. The constant saving seemed to consistently stall out my system. I'm also relatively stable with the interval set to 120. Setting the save states to the largest memory available along with the longest interval between save states did most of the stability fixes. The screen crash only happens now if I try to unpause the game from the Bizhawk window (Rather when seeking or autosaving) and is avoided by closing and re-opening Bizhawk before doing Large Sequence check. (I close TAS studio before playing manually, as TAS studio seems to prevent my my controller inputs from being accepted) Bizhawk threw my first "Error" screen yesterday when re-opening TAS studio after such a play check. I appologise I forgot to copy/paste it. Next time it happens I will paste it here. * Edit == Found the Exception Window == Goofed around with my editor, Closed Tas Studio Window (So we are just Bizhawk Now), Opened my Ram watch, tried to Re-open TAS studios. Ram watch is only looking at one memory address that is relatively static and only rotates when you trigger an in game day. I'm not an expert, but I have Bizhawk installed on my D drive, as my C drive is my Operating system SSD. Would the split between assembly on my C drive and the executable on the D drive make any difference? Debug Note See the end of this message for details on invoking just-in-time (JIT) debugging instead of this dialog box. ************** Exception Text ************** System.NullReferenceException: Object reference not set to an instance of an object. at BizHawk.Emulation.Cores.Nintendo.SNES.LibsnesApi.Enter() at BizHawk.Emulation.Common.MemoryDomainIntPtrMonitor.PeekByte(Int64 addr) at BizHawk.Client.Common.Watch.GetByte(Boolean bypassFreeze) at BizHawk.Client.Common.ByteWatch.get_ValueString() at BizHawk.Client.EmuHawk.RamWatch.WatchListView_QueryItemText(Int32 index, Int32 column, String& text) at BizHawk.Client.EmuHawk.VirtualListView.OnDispInfoNotice(Message& m, Boolean useAnsi) at BizHawk.Client.EmuHawk.VirtualListView.WndProc(Message& m) at System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)
Posted: 3/31/2019 7:24 PM
Post subject: BizHawk-2.3.1 - Tas Studio Keeps Crashing - Autosaves?
Quote
PlayPatrice
He/Him
Experienced Forum User
Joined: 3/31/2019
Posts: 20
BizHawk keeps crashing for me, and I think it has something to do with the autosave features of TAS Studios. I'm new to TAS'ing and I'm fairly computer literate, but coding illiterate. I'm currently working on a frame by frame input run of Harvest Moon (SNES, Eng, NTSC), and now that the frame count is getting up there, I'm seeing things like 181 save states and the whole thing crases any time I try and do an unpause playthrough to check larger stretches of the video. I went into the Memory settings and doubled a few of the values and that seemed to fix my issue for a few hours, then it started crashing again. The savestates then jumped to 361, and the crashes started again. Most of my crashes "Feel" like they revolve around autosave or savestate feature. With one exception I've never lost any of my work to one of these blackout crashes. It doesn't throw an error, just says it's not responding until it asks me to close the program. I'm running windows 10, with BizHawk running in windows 8 compatibility mode (Which prevented a lot of crashing early on). Does anyone have any advice on how to set up or disable the autosave feature to prevent crashes like this (Or prevent massive file bloat?). A finished 100% run of Harvest Moon is going to be a TAS that is at least 45 minutes to two hours of game time. And I'm only 15 minutes in while running into consistant crashes so far. The memory usage in my Task manager only hits like, 27-37% and I've got over 16gb of ram in my system. If anyone could give their advice or suggest how to fiddle with the Programs settings to make it more stable, that would be much appreciated. Edit: I'm using the BSNES core if that makes any difference.