First GSC Research Thread

obi

obi

formerly david stone
is a Site Content Manager Alumnusis a Programmer Alumnusis a Senior Staff Member Alumnusis a Smogon Discord Contributor Alumnusis a Researcher Alumnusis a Top Contributor Alumnusis a Battle Simulator Moderator Alumnus
Hello, and welcome to the first GSC research thread! We need your help testing this stuff out, so don't be intimidated. Many of the things still in this thread will take about a minute to test. Even the stuff that requires a lot of testing can still be accomplished by a couple of guys testing while they wait for their pizza to be delivered! A few people submitting trials of 15 or so hits at a time can really add up. If you are unsure of how to test something, don't hesitate to ask. It's a lot better to ask how to test something than to spend an hour setting up an elaborate test only to find out that your results are inconclusive.

However, that doesn't mean you just post anything here. Questions about breeding chains and the like belong in another thread. This thread is reserved for questions about game mechanics.

Items
Moves
 
This information can only be tested in-game. Netbattle or any other simulator are known to have false information (and they run on the information we have as a community, anyway).
 
Obi, I'd like to help out, but I fall under your category of "not sure how to test". So, if you want me to double-check on Quick Claw/Flail/Reversal, could you let me know how I should go about doing it?
 
Another note on Flail/Reversal: there is no "damage range" when using Flail or Reversal. When you take everything into account, Flail/Reversal will deal exactly the average damage in the range. However, I am not sure whether, in the event of a .5 average, the number is rounded up or down.
 
GreenPikachu said:
Obi, I'd like to help out, but I fall under your category of "not sure how to test". So, if you want me to double-check on Quick Claw/Flail/Reversal, could you let me know how I should go about doing it?
Click to expand...

Quick Claw would be the easiest to test. Basically, just get a slower Pokemon to use a move against a faster Pokemon when the slower Pokemon has a Quick Claw. Count how many turns you use and how often the slower Pokemon moves first. I actually plan on testing this one myself later, but we'll need a lot of data to find out the actual %.

Another thing we could use people to do is to just go through the GSC descriptions. If there is anything you're interested in learning more about or wonder whether the data is correct, post here. Virtually nothing in there has been confirmed by actual GSC testing. A lot of it is DP / ADV testing or just "Yeah I remember it works like this in GSC", which is good, but not always accurate. A lot of older info is "Belly Drum Salamence". One site makes an error and everyone copies it.
 
The description of Psywave is incorrect; in GSC (as well as RBY) it can do anywhere from 1 to (1.5 x current level) damage. I'm pretty sure it's not bound to multiples of 0.1 x current level for damage either, but I'd need more precise numbers to check for that. This applies for RBY as well, for that matter.
 
Im testing the Quick Claw one. I didnt see the post about the do not use netbattle, so I had to get rid of my data and start over =( anyway I plan on pulling out my gold version and use quick claw snorlax on lance's dragonite. i will use a set of snore, sleep talk, rain dance, amnesia probably and just carry a lot of full restores and pp enhancers. Is that okay or should I do something else as well? I think that should give me like a few hundered attacks before my items and pp runs out and i start to struggle.
 
yeah that would work, although you obviously can't count the turns you use an item as anything (just pretend like nothing happened that turn).
 
Here is a good question i think needs a serious study do Hi Critcal hit moves really have 25% CH rate in both GSC and ADV or is it something else? it took a while to find out Hi CH moves only had a 12.5% CH rate in DP so i think this sould be doubled checked in both GSC and ADV
 
I'll do some CH samples for GSC.

Hi-Crit
Used Sandslash with Slash on wild Pokemon

Took 100 samples (five rounds of Slash PP) every time.
Format:
#. 100 samples, x CHs, (out of these 100%, average over this and all previous%)

1. 100 samples, 20 CHs (20%, 20%)
2. 100 samples, 30 CHs (30%, 25%)
3. 100 samples, 21 CHs (21%, 23.667%)

Quick Claw
Used Stantler with Quick Claw against Morty's Gengar. Sandslash Sand-Attacked it six times so that it could barely hit me, and I just used Leer and Tackle and checked whenever he did or did not get Quick Claw. I'm 100% Gengar's Speed is higher. It also seems to activate much more often when I use Leer, but I guess that is just random.

1. 54 samples, 12 triggers (22%, or 56.32/256)
2. 90 samples, 20 triggers (22.2%, or 73.14/256)
 
I couldn't find the answer to this and couldnt find it on site, but how does metronome determine what move you use? It can not be completely random as there some moves that are used more then others. I tested metronome with my togepi for some time and the majority of my moves were the following:

15/80 times: pursuit
12/80 times: safeguard
10/80 times: crunch

Sorry if this is already known but if not I was planning on doing a few tests of metronome on different pokemon and whatnot to see if their is sometimes a common pattern, perhaps linking to hidden power or ivs.

*Im also working on the quick claw tests. I will have them completed and posted in 2 days most likely.
 
No testing from me right now (maybe later), but:
- For AJC's question there about high critical moves, I think the way it was understood to work was that the critical hit levels ramped up as follows:
6.25% (normal)
12.5% (+1)
25% (+2)
37.5% (+3)
50% (+4)
and that you get +1 level for any of Scope Lens, high critical hit move, Focus Energy, and you get +2 levels from Stick on Farfetch'd or Lucky Punch on Chansey. So GSC Slash had a 12.5% rate in GSC the same as (you indicate) it has in DP. I have not generated test results to support it, so I'm fine if you find it insufficient, but I know this was the conclusion reached by both ROM hackers and testers back in the days when Scope Lens+Focus Energy+Slash Ursaring was considered viable in G/S.

With regards to the Hidden Power formula, all I know of it is that 13/13/15/15 yields Bug 70.

I'm hoping someone will ask a question I can test in a couple of minutes rather than taking hundreds of trials, so that I can help with real test data and not just spout the accepted knowledge from the good/bad old days without spending even more time doing tedious stuff in pogheymanz than I already do.
 
You could just use Slash on something once until you run out of PP.

If you really want to be helpful and efficient, you could use Slash on a slow Pokemon with a Quick Claw and test both effects simultaneously! You don't need hundreds of trials personally, you can just contribute small amounts which will help us in the aggregate.
 
Hidden Power formula from UPC

http://www.upokecenter.com/games/gs/guides/hiddenpower.html

BaseDamage = int((A*5+B)/2)+31
* A starts at 0.
* If the Attack DV is 8 or more, A=A+8.
* If the Defense DV is 8 or more, A=A+4.
* If the Speed DV is 8 or more, A=A+2.
* If the Special DV is 8 or more, A=A+1.
* B = min(Special DV,3)

Type = (A%4)*4+(B%4)
* A = Attack DV
* B = Defense DV
* 0: Fighting
* 1: Flying
* 2: Poison
* 3: Ground
* 4: Rock
* 5: Bug
* 6: Ghost
* 7: Steel
* 8: Fire
* 9: Water
* 10: Grass
* 11: Electric
* 12: Psychic
* 13: Ice
* 14: Dragon
* 15: Dark
 
Quoting from the best RBY and GSC guides ever:

Quick Claw: At the beginning of each turn, probability 60 to increase your Speed by 1024 until end of turn.

Brightpowder: All opponents' attacks which could cause damage or a change in status or stat modifiers of your Pokemon have their accuracy reduced by 20.

Focus Band: If an opponent's attack would deal damage to reduce your Pokemon's HP to 0 or less, probability 30 to make that attack reduce its HP to 1 instead.

King's Rock: Whenever you use a damage-dealing attack that can't change status or stat modifiers, probability 30 to make the opponent flinch.
Click to expand...

'Probability x' in GSC means x/256. So Brightpowder has probability 20/256 = 7.8125% of activating, while Focus Band and King's Rock have probability 30/256 = 11.71875% of activating.

About Hidden Power:

Now the DVs can give the HP specifics. Start with a power of 32.5. If the Attack DV is 8 or more, add 20 to the power. If the Defense DV is at least 8, add 10. Repeat for Speed (+5) and Special (+2.5). Then, if the Special DV is even, subtract 0.5 from the power, and if half the Special DV (rounded down) is even, subtract 1. Now disregard any 0.5 that may be left in the power.

Congratulations, you now have the strength of your HP. If it's outside of the 31-70 range, you've done something wrong and need to redo the calculations.

For type, take your Attack DV, and "mod-4" it. In other words, divide by 4 and look only at the remainder. Multiply that number by 4. Next, mod-4 the Defense DV and add to the previous result. You'll get a number from 0 to 15, and then it's a simple chart to find the type:

15 = Dark
14 = Dragon
13 = Ice
12 = Psychic
11 = Electric
10 = Grass
9 = Water
8 = Fire
7 = Steel
6 = Ghost
5 = Bug
4 = Rock
3 = Ground
2 = Poison
1 = Flying
0 = Fighting

Finally, note that Normal is the only type missing, yet it's always what HP's type is shown as for the game interface. This is because Hidden Power only takes its calculated type during your own turn and only on turns when it's used; it's Normal at all other times. All this means is that HP can always be Countered, and never Mirror Coated, regardless of type; and that Conversion and Conversion2 see it as a Normal attack.
Click to expand...
 
According to same site:

Flail: Power 20. Power +20 if user has less than 68.75% of its maximum HP. Power +40 if user has less than 35.42% of its maximum HP. Power +20 if user has less than 20.83% of its maximum HP. Power +50 if user has less than 10.42% of its maximum HP. Power +50 if user has less than 4.17% of its maximum HP. Uses no Critical Hit Domains.
Click to expand...

Reversal has the same description.

Psywave: Reduces opponent's HP by a random number, chosen from a field with lower bound 1 and upper bound (UserLevel * 1.5).
Click to expand...

About Critical Hits: in GSC, high critical hit moves have a 25% chance of scoring a CH, not 12.5%. Here's what the site says:

The new formula is best expressed by what I call "Critical Hit Domains". There are five of these in all. Each domain is a set of integers:

Domain 1: 0 to 15
Domain 2: 16 to 31
Domain 3: 32 to 63
Domain 4: 64 to 84
Domain 5: 85 to 127

Each move starts out with one active domain (#1; each activation takes the lowest-numbered available domain). Aeroblast, Cross Chop, Crabhammer, Karate Chop, Razor Leaf, Razor Wind, and Slash start with 3. Scope Lens activates an additional domain, as does Focus Energy. Lucky Punch or Stick on the appropriate Pokemon (Chansey/Farfetch'd) activates two domains. Anything that would somehow activate more than 5 domains is ignored.

Then, a 1-byte (0-255) random number is chosen, similar to a lottery. If that number falls into an active Critical Hit domain, up goes the x2 bonus.
Click to expand...
 
Well I had to redo the entire game and I still am not finished yet. However I did get the Quick Claw and put it on Togepi and then tested against the much faster gastlys and haunters in mortys gym. Quick Claw activated .21 times. Or 12/55 times whatever way you wanted. Actually only .02 off from what the standard number was. My results were pretty simple:

Code: 
1. No
2. No
3. No
4. No
5. No
6. Yes
7. No
8. No
9. No
10. No
11. No
12. No
13. No
14. No
15. Yes
16. Yes
17. No
18. No
19. Yes
20. No
21. No
22. No
23. Yes
24. No
25. No
26. No
27. No
28. Yes
29. No
30. Yes
31. No
32. No
33. No
34. No
35. No
36. No
37. No
38. No
39. No
40. No
41. Yes
42. No
43. No
44. Yes
45. No
46. No
47. No
48. No
49. Yes
50. No
51. Yes
52. No
53. Yes
54. No
55. NO
 
It seems like X-Act provided you with all the answers, but I feel like I shouldn't let my testing go to waste, even though it isn't much.

I tested Quick Claw's probability to activate with a sample of 220:

49/220 = 22.27%

It is very close to the 60/256 or 23 something %, so with adding up everyone elses trials, and X-Act's finding, I think it'd be safe to say that Quick Claw does activate 60/256, like you thought.

-Also, what is the probability for being imobilized by love? In one of my matches against the computer in PS2(Although it's probably a fluke, or similar to the haxy battle tower in DP), my Lv.55 Snorlax was immobilized by love against a Lv.50 Female Flareon 6 consecutive times, and a total of 8/10 times against the same Flareon. I thought that being immobilized by love had a 50% chance each time. Although this probably isn't much, and it was just luck, maybe Attraction has a higher percentage?
 
From same site:

Attract: Normal, Power 0, Accuracy 256, PP 16. If the two players' Pokemon are of opposite gender, then until a Pokemon switches or is KO'd, the opponent has probability 128 to lose each of its turns.
Click to expand...

Probability 128 means 128/256, or 50%.
 
You must log in or register to reply here.
Back
Top