48ths vs percentages for HP

[Stratos]

may as well make a thread about this.

So on cartridge HP is displayed as a fraction of 48 rounded up, whereas on PS (and every other sim) it's displayed as a percent. I'm not sure how percent is rounded—rounded normally, except at 1% and 99%?

Benefits to percents:

• More information (roughly 2x as precise as 48ths allows for slightly more accurate knowledge of rolls to KO)
• People are used to working with percents—it may take longer to think whether an attack that did 18/48 of damage is a 3HKO, than if it did 38%.

Benefits to 48ths:

• No more "8% sand." With the sole exception of 4x resisted stealth rock, all passive damages in the game are exact multiples of 1/48. So for example if a Pokemon is at 6/48, SR will ALWAYS kill, at 7/48 it NEVER will. edit: i forgot life orb. The only other thing in the game which is actually made less accurate by changing to 48ths is life orb, which will occasionally kill when HP is displayed as 5/48 but not usually. I'm not sure if life orb always kills when HP is displayed as 9% but never when it's displayed as 10%, or if it sometimes kills when it's at 10%.
  
• Cart accuracy is always nice.

The way I see it, the only benefits to percents are basically that people are comfortable with the status quo.

Sure I've seen the arguments about information and while they're not really wrong, the difference in information is really incredibly minor. The only time I can see it coming into play is if a Pokemon is at some wack-ass HP, say, 19/48, and you're using an attack that has a min roll of 39.0%. Since both 38% and 39% display as 19/48 you can't be sure whether the attack will guaranteed KO. But then again, that same shit happens with percents, it'll just happen slightly more often now. The important rolls—guaranteed 2hkoes and 3hkoes—are exact multiples of 1/48 so they won't be affected by the loss of information (and hell, since guaranteed 3hkoes arent an exact multiple of 1/100, theyre actually helped), and since damage rolls vary by 15%, honestly, the 1/100 information you will sometimes lose out on is well within the typical error of randomness anyways.

As for what people are more used to working with—true, we're all comfortable with percents, but it wouldn't take long to adjust to 48ths. There's a reason GF picked that number, it's one of the most convenient numbers for fractions, since it cleanly divides into 2, 3, 4, 6, 8, 12, 16, and 24.

On the other hand, the benefits to switching to 48ths are real and permanent.

I know 48ths failed once before, but imo that was because it was done without any consulting the community and was poorly and unintuitively implemented. I took screencaps of a few games to show the current implementation of the 48ths method and my proposed implementation:

DOU game with percents:

VGC game with 48ths:

My proposal:

to clarify since someone asked about this in PM: when it shows how many 48ths of HP you lost, it would simply be showing the difference between the new 48ths value and the old one, not converting the HP lost into a 48ths value. It would give you absolutely no information you don't get in-game.

thoughts?

 

[Nix_Hex]

At least you're telling us before unlike cathy x_x

I don't have much to say but cart accuracy is of utmost importance. BUT I don't see why we can't show % to the n'th digit since that doesn't show anymore info than thr player gains by counting pixels and dividing by 48 with a calculator.

In favor.

Edit: to clarify, show BOTH.

 

[Zarel]

I think a lot of casual users will have trouble with e.g. is 7/48 enough to survive a turn of burn damage? etc etc

Having both might work. Something like "13% (6/48)"

 

[Stratos]

I think a lot of casual users will have trouble with e.g. is 7/48 enough to survive a turn of burn damage? etc etc

Having both might work. Something like "13% (6/48)"

Not 100% sure what you're proposing here. If it's that the exact HP be displayed as both /100 and /48, then you run into the same issue as decimal percents, which is being able to figure out exact spreads. Granted it's a lot more complicated math than with decimal percents, but it's doable. I remember that being an issue the last time this was discussed, though i'm too lazy to find the post i really want so just have this one

If you're saying just take whatever x/48 is and display it as a percent (rounded up?), sure, sounds like a fine idea.

 

[Zarel]

If you're saying just take whatever x/48 is and display it as a percent (rounded up?), sure, sounds like a fine idea.

Yeah, that's what I meant.

 

[Nails]

percentages allow for way more precise play. seeing that my icy wind dealt 4.2-8.3% to the opposing kyogre is moderately frustrating but i put up with it because it's better practice for vgc events. i do not want that to be applied to smogon metagames however, i really like the precision allowed with percentages. edit: if someone wants me to go into greater detail lmk, this is a 4 am post and it's kind of messy so w/e

also @ your points in the op, life orb will always kill a mon displayed at 10%. and due to rounding/greater general precision, i think percentages are still more accurate than 48ths even when dealing with numbers that are fractions of 48, eg a 323 hp mon switching into 1 layer of spikes and 283 / 323 = 0.8761 being displayed as either 88% with the current system or 87.5%-89.6% with 48ths. the second time it switches in, seeing that it took 10.4%-14.6% from spikes is also somewhat frustrating.

please tell me if i'm missing something or messing up how pokemon displays hps. i don't think i am though. the current system is very user friendly, does not actually break cart mechanics (players could agree to report their hp as a percentage to the opponent every time their hp changes. it would be tedious but yeah, it's just a rule that this hypothetical scenario enforces, we're still playing pokemon), and it has been used for years. 100ths are the perfect balance of information to provide.

sorry for the rambling post

 

[Stratos]

percentages allow for way more precise play. seeing that my icy wind dealt 4.2-8.3% to the opposing kyogre is moderately frustrating but i put up with it because it's better practice for vgc events. i do not want that to be applied to smogon metagames however, i really like the precision allowed with percentages. edit: if someone wants me to go into greater detail lmk, this is a 4 am post and it's kind of messy so w/e

Yes, please do. The reason it displays the percentage ranges like that is because it could technically be going from the very top of the last 48th to the very bottom of the next 48th. But this isn't really a one-to-one comparison with percentages. If I attack a 59% pokemon and I drop it to 52%, the damage window SHOULD read that it dealt 6.1%-7.9% since anything in that range could bring a mon from 59% to 52%.

Back to my proposal, it wouldn't say that the icy wind dealt 4.2%-8.3%, it would simply say that it dealt 3/48.

So once again, there are only 2 percentages per 48th, which is barely any additional precision.

also @ your points in the op, life orb will always kill a mon displayed at 10%. and due to rounding/greater general precision, i think percentages are still more accurate than 48ths even when dealing with numbers that are fractions of 48, eg a 323 hp mon switching into 1 layer of spikes and 283 / 323 = 0.8761 being displayed as either 88% with the current system or 87.5%-89.6% with 48ths. the second time it switches in, seeing that it took 10.4%-14.6% from spikes is also somewhat frustrating.

That's because the system we use now for 48ths is very silly. You know exactly how much damage Spikes dealt—it dealt exactly 6/48 rounded down of the opponent's HP. If the sim displays that as 10.4-14.6%, that's because it's showing the entire range of possible damage losses which could result in that change of 48ths. Again, compare to my above example. If the damage display methods were one-to-one, it would show that the pokemon has 87.1%-88.0% HP after switching into one layer of spikes with the percentage method.

please tell me if i'm missing something or messing up how pokemon displays hps. i don't think i am though. the current system is very user friendly, does not actually break cart mechanics (players could agree to report their hp as a percentage to the opponent every time their hp changes. it would be tedious but yeah, it's just a rule that this hypothetical scenario enforces, we're still playing pokemon), and it has been used for years. 100ths are the perfect balance of information to provide.

sorry for the rambling post

The only problems you've really brought up are the issues with the current implementation of 48ths which is rather poorly done. Please look at my proposal at the bottom of the OP and see if it's still really that repugnant to you.

 

[Nails]

The only problems you've really brought up are the issues with the current implementation of 48ths which is rather poorly done. Please look at my proposal at the bottom of the OP and see if it's still really that repugnant to you.

I mean. My main issue with 48ths are that they are less accurate than 100ths. The only difference between your proposal and the current implementation is that yours provides less information than the current system. My points that I made were 1) that I very strongly prefer the current set up with 100ths and that it's basically a matter of preference since 2) it is not a cart mechanics question in any way.

Yes, please do. The reason it displays the percentage ranges like that is because it could technically be going from the very top of the last 48th to the very bottom of the next 48th. But this isn't really a one-to-one comparison with percentages. If I attack a 59% pokemon and I drop it to 52%, the damage window SHOULD read that it dealt 6.1%-7.9% since anything in that range could bring a mon from 59% to 52%.

Back to my proposal, it wouldn't say that the icy wind dealt 4.2%-8.3%, it would simply say that it dealt 3/48.

So once again, there are only 2 percentages per 48th, which is barely any additional precision.

You misunderstood why I was upset with 4.2%-8.3%, I don't care about how the battle log is phrased I just don't like there be so much ambiguity about how much damage my attacks dealt. The information in your or zarel's implementation is the exact same. I literally don't care which is used for the sake of this argument, I dislike both of them when applied to anything but vgc.

 

[Stratos]

I mean. My main issue with 48ths are that they are less accurate than 100ths. The only difference between your proposal and the current implementation is that yours provides less information than the current system. My points that I made were 1) that I very strongly prefer the current set up with 100ths and that it's basically a matter of preference since 2) it is not a cart mechanics question in any way.

sure it is. whether or not you think the "report your HP" thing is reasonable, it's absolutely against cart mechanics to not display HP in 48ths. But since reporting both is even worse, we prefer to break cart mechanics. In the end what it means is that reporting 100ths de facto breaks cart mechanics since it means you can't report 48ths.

You misunderstood why I was upset with 4.2%-8.3%, I don't care about how the battle log is phrased I just don't like there be so much ambiguity about how much damage my attacks dealt. The information in your or zarel's implementation is the exact same. I literally don't care which is used for the sake of this argument, I dislike both of them when applied to anything but vgc.

Fair enough. How about reporting HP to 200ths then? Any FE pokemon that would ever run HP investment (read: not dugtrio) has at least 200 HP min, so it shouldn't give away any spreads. Logically, reporting to 200ths is strictly an upgrade over reporting to 100ths.

Realistically, the consequences of doubling the precision of our HP reporting would be next-to-none, and the same goes for halving it. I don't think your Icy Wind doing 4.2-8.3% is that much different than it doing 5.0-6.9%. But the logical place to plant your flag would be on the value that the actual Pokemon games give us, which is 48ths.

 

[Zarel]

I mean. My main issue with 48ths are that they are less accurate than 100ths. The only difference between your proposal and the current implementation is that yours provides less information than the current system. My points that I made were 1) that I very strongly prefer the current set up with 100ths and that it's basically a matter of preference since 2) it is not a cart mechanics question in any way.

1. The accuracy question shouldn't matter. As I posted in the other thread, /100s has always been about convenience. Accuracy is never a reason to break cart mechanics.

2. 48ths actually are more accurate, for determining things like "do you survive rocks?" etc.

Fair enough. How about reporting HP to 200ths then? Any FE pokemon that would ever run HP investment (read: not dugtrio) has at least 200 HP min, so it shouldn't give away any spreads. Logically, reporting to 200ths is strictly an upgrade over reporting to 100ths.

Realistically, the consequences of doubling the precision of our HP reporting would be next-to-none, and the same goes for halving it.

Doubling/halving precision is actually kind of huge. It's not enough to look at a value in isolation and have a good chance of narrowing down the EVs significantly, but it's enough to make a variety of inferences. A bot can narrow down an HP EV in a few turns, and double precision halves the number of turns it takes, and vice versa.

Again, I'm willing to make this trade-off for convenience (re: humans prefer to think in percentages), but not for other reasons.

 

[Alfalfa]

How would calculating damage in 48ths be compatible be compatible with fixed damage attacks, like Seismic Toss and Night Shade? Wouldn't setting damage percentages in 48ths conflict with the actual damage output of these attacks?

 

[Texas Cloverleaf]

Again, I'm willing to make this trade-off for convenience (re: humans prefer to think in percentages), but not for other reasons.

I support this message

 

[Nix_Hex]

How would calculating damage in 48ths be compatible be compatible with fixed damage attacks, like Seismic Toss and Night Shade? Wouldn't setting damage percentages in 48ths conflict with the actual damage output of these attacks?

How do you figure? Damage is damage. The "did x to y% damage" report reflects upon 48ths depleted, not damage rolls. Since Night Shade and SToss do, in fact, do damage, the same principle applies.

 

[Antar]

Always in favor of cart accuracy. Does that mean we have different divisors based on the generation, though?

 

[Zarel]

Always in favor of cart accuracy. Does that mean we have different divisors based on the generation, though?

It's 48 pixels in every generation from 1 to 6.

 

[Tobes]

how this is going to go:

-Nobody posts in opposition of this because a) Hardly anyone actually wants this change, to such a degree that no one thinks it'll actually get implemented and b) Who actually reads Policy Review?
-Silence is taken as consent and change gets implemented to Showdown!
-Players quickly realize what happened and raise a shitstorm. People take up arms and have heated arguments with the vocal minority favoring removal of percentages. Much ill will is spread.
-Change is reversed with or without some shoddy compromise, status quo is restored. Everything's back to normal, except everyone's just a little bit more angry and annoyed than before.

Can we just save ourselves the trouble and not bother with this? I'm sure the opposite of Statros's proposal would work fine: Percentages as the default, but you could get 48ths if you mouseover the numbers.

 

[Zarel]

Can we just save ourselves the trouble and not bother with this? I'm sure the opposite of Statros's proposal would work fine: Percentages as the default, but you could get 48ths if you mouseover the numbers.

Err, what are you proposing here? Real percentages and 48ths displayed? That's too much information and can't be done. Estimated percentages from 48ths? That's exactly what this thread is proposing.

 

[blarajan]

how this is going to go:

-Nobody posts in opposition of this because a) Hardly anyone actually wants this change, to such a degree that no one thinks it'll actually get implemented and b) Who actually reads Policy Review?
-Silence is taken as consent and change gets implemented to Showdown!
-Players quickly realize what happened and raise a shitstorm. People take up arms and have heated arguments with the vocal minority favoring removal of percentages. Much ill will is spread.
-Change is reversed with or without some shoddy compromise, status quo is restored. Everything's back to normal, except everyone's just a little bit more angry and annoyed than before.

Can we just save ourselves the trouble and not bother with this? I'm sure the opposite of Statros's proposal would work fine: Percentages as the default, but you could get 48ths if you mouseover the numbers.

haha yeah #edgy

this is one of the stupidest posts i've read in a very long time. don't attempt to stifle discussion or proposals because you can't be assed to be involved with the conversation. the proposal makes sense. we are actively violating game mechanics by not presenting information in the 48ths, which has had impact in games (regarding hazard damage). while we are able to theoretically justify offering information in 100ths, we are not representing the game fully by failing to present 48ths. if we offer both, we can easily ascertain spread information through the combination of information. this is a conversation worth having. don't fucking come in here with your too cool for school shit and hand wave all over it for likes. like honestly what was that even "no one reads policy proposal so let's stop proposing policy1!!"

 

[Tobes]

Zarel: The former.

haha yeah #edgy

this is one of the stupidest posts i've read in a very long time. don't attempt to stifle discussion or proposals because you can't be assed to be involved with the conversation. the proposal makes sense. we are actively violating game mechanics by not presenting information in the 48ths, which has had impact in games (regarding hazard damage). while we are able to theoretically justify offering information in 100ths, we are not representing the game fully by failing to present 48ths. if we offer both, we can easily ascertain spread information through the combination of information. this is a conversation worth having. don't fucking come in here with your too cool for school shit and hand wave all over it for likes. like honestly what was that even "no one reads policy proposal so let's stop proposing policy1!!"

Point taken. While I maintain my belief that this is not a change the majority of people wants, that doesn't justify my being dismissive of the idea or my facetiousness towards Policy Review as a whole, and I apologize for that.

 

[Reymedy]

There's a lot of things that are available to us when we play on a simulator as opposed to when we play on the game cartridge, so I'm really not buying the "yea cartridge" argument.
And actually, there's not much than that to support this change.

The bottomline being

humans prefer to think in percentages

And it's a pretty important point, I'm not sure that I want to start thinking tommorow with your 48's at all. In fact, I know that I don't want to because % are way more convenient.
Let's assume 48's are slightly better, would it still justify to change everybody's habbit all of a sudden ? I don't think so.
My point being, you need a way stronger reason to back-up a change like that relating to users' experience. You're going for sure to worsen it in the short-term (humans do not prefer to change what they're used to), to "potentially" improve it much later. That sounds like everything but a good move to me.

Besides, who would like to make every single turn less accurate (by that I mean, less convenient to access accurate information), just to make that one single turn that happens every 30 games, more convenient (the one where your switch your something into SR and wonder why it just lost 8%).

With that said, I like the idea of having access to both information with the mouse-over thingy, I mean why not. That'd be actually a good feature and a relatively satisfying compromise between cartridge and simulator's optimization, but only if % are what is displayed by default, and 48's the bonus information that you can access via whatever is the most convenient for Zarel to implement.

 

[Aberforth]

Why not is pretty clearly outlined already, cause it would mean you can calculate opponent's EV spreads really easily.

Also 48s are not 'slightly better' as a counting system, base 12 counting is a lot better than base 10 counting systems.

Aside from that, still dont particularly care. % are convenient cause of everyone using a decimal system (despite, yes, being the worse counting system of the two) and it works fine as a status quo that I dont feel a particular pressing need to change it.

 

[OrdA]

Hi,

I believe the convenience argument in favour of retaining 100ths is weak.

The server can just send the data in 48ths and the client will let you toggle between displaying 48ths and displaying %s, calculated by (100 * x) / 48. If displayed with enough digits, you can always tell if a mon is going to live stealth rock damage, even from the displayed % value.
The default display should be %s, because this is more intuitive for new players.

Edit: The real question is whether we want the ability to always tell if a mon is going to live sr, toxic, etc.

100ths -> 48ths:
- easier to read out 16th-based damage (rocks, sandstorm, toxic, etc)
- cartridge-like behaviour
- half the accuracy

This can be taken apart into two independent decisions, the second being about accuracy vs. cartridge behavior.

100ths -> 96ths:
- easier to read out 16th-based damage (rocks, sandstorm, toxic, etc)
- similar accuracy

96ths -> 48ths:
- cartridge-like behaviour
- half the accuracy

 

[DragonWhale]

In regards to HP accuracy I think the percentage mod is the better option for Smogon formats because it's more consistent with the official formats (Battle Spot and VGC) in that regard. Smogon formats have around double the max HP so personally it makes sense to double the HP accuracy.

Honestly the passive damage argument is mostly wrong. You're not guaranteed to lose 6/48 when you switch in to rocks or take 3/48 to sandstorm. A full health zard with odd max HP will lose 23/48 due to all pixels rounding up (except 47-> 48). You can argue that 5/48 is guaranteed to die to rocks but then again so is 12%. edit @ below: o tru

The player who needs to determine whether or not they survive passive damage has access to their current hp/ max hp, which they should always use regardless of HP accuracy. I've probably played on cart more than I have on PS and I've never counted pixels to determine my next move lol. It's always through current/max hp.

The only good argument for 48ths is staying true to the cart imo.

 

[Halcyon.]

If this gets implemented then we had better change sleep clause to reflect carts as well. I wanna see Randomuser69 lose instantly the moment he tries to put two of my mons to sleep.

Trying to simulate cart mechanics is fine, but don't act like it's ok to follow them sometimes and not others. We don't have mandatory timers (stall would be less common if we did), we don't have accurate sleep mechanics, and we don't show things in 48ths. I don't see why the percentages are an issue if we continue to change things for our own benefit anyway.

But like I said, change what you want. Just make sure that we consider all the other modifications we have made that have actual effects on games.

 

[Stratos]

In regards to HP accuracy I think the percentage mod is the better option for Smogon formats because it's more consistent with the official formats (Battle Spot and VGC) in that regard. Smogon formats have around double the max HP so personally it makes sense to double the HP accuracy.

disagree. except for dragon rage and sonic boom, everything is twice as big at level 100 as it is at fifty, so it'd be more "in line" for the hp increments to be twice as big as well.

Honestly the passive damage argument is mostly wrong. You're not guaranteed to lose 6/48 when you switch in to rocks or take 3/48 to sandstorm. A full health zard with odd max HP will lose 23/48 due to all pixels rounding up (except 47-> 48). You can argue that 5/48 is guaranteed to die to rocks but then again so is 12%.

The player who needs to determine whether or not they survive passive damage has access to their current hp/ max hp, which they should always use regardless of HP accuracy. I've probably played on cart more than I have on PS and I've never counted pixels to determine my next move lol. It's always through current/max hp.

The only good argument for 48ths is staying true to the cart imo.

this is wrong. not technically wrong per se, in that zard will lose 23/48 if it switches in on SR from full, but that doesn't change the fact that a zard at 24/48 will always die to SR and a zard at 25/48 never will. (or a garde at 6/48 vs 7/48, etc.) It has to do with rounding.

If this gets implemented then we had better change sleep clause to reflect carts as well. I wanna see Randomuser69 lose instantly the moment he tries to put two of my mons to sleep.

Trying to simulate cart mechanics is fine, but don't act like it's ok to follow them sometimes and not others. We don't have mandatory timers (stall would be less common if we did), we don't have accurate sleep mechanics, and we don't show things in 48ths. I don't see why the percentages are an issue if we continue to change things for our own benefit anyway.

But like I said, change what you want. Just make sure that we consider all the other modifications we have made that have actual effects on games.

there's no need to be a butt. obviously we can vary from cart if it's significant enough to be worth it, but cart implementability should be something worth trying for. it's a weighing of trade-offs, it's not all or nothing.