-
Welcome to Smogon! Take a moment to read the Introduction to Smogon for a run-down on everything Smogon, and make sure you take some time to read the global rules.
-
Congrats to the winners of the 2023 Smog Awards!
"The Win Formula" -- Lights, Camera, Action!
- Thread starter DougJustDoug
- Start date
- Status
OK, so obviously this joke is ruffling feathers all over the place. IRC is bristling about it and the thread is going nuts. We need to let this simmer down a bit. We can't sustain an argumentative pace like this for the next 20 days. We'll "burn out" long before April Fools Day. Let's try to keep the schedule and not get too wrapped up in the drama right now. If people freak out over it, let them vent and just make a general defense of the idea.
My biggest worry is that there will be a widespread leak that this is a joke. Which would be a real shame, because it looks like this thing could be the ultimate prank of all-time. If we can somehow manage to keep a lid on this, and if we can somehow continue to manufacture arguments for this preposterous formula proposal AND do it with a straight face -- the payoff on this joke will be epic.
I'm crossing my fingers and hoping we don't get "haxed" ourselves!
My biggest worry is that there will be a widespread leak that this is a joke. Which would be a real shame, because it looks like this thing could be the ultimate prank of all-time. If we can somehow manage to keep a lid on this, and if we can somehow continue to manufacture arguments for this preposterous formula proposal AND do it with a straight face -- the payoff on this joke will be epic.
I'm crossing my fingers and hoping we don't get "haxed" ourselves!
Agreeing that we let it simmer down a bit before going in with more sucker punches. I still can't wait to ping some pong though.
I have a channel dedicated to programming / AI that Chris and occasionally LN hang out in. I could drop some hints in there that I'm working on a program that could be used to "auto-calculate" battles to attempt to determine the overall effect that luck has played. I obviously wouldn't want to say that now, but more along the lines of Mar-25th or so I let people know that I've been working on a top secret project the past month or so, and then explain just what it is.
What started out as something I was interested in just to see if it were possible ended up as a several week project (that leverages my current work) to work out battles in earnest.
What started out as something I was interested in just to see if it were possible ended up as a several week project (that leverages my current work) to work out battles in earnest.
I intentionally didn't cast the White Knight, because I think we need to let the joke play out a bit, and get a feel for who would be the right person to deliver the punch line of this big joke. I don't really see myself doing it -- since I think the White Knight should kinda "come from nowhere to save the day" (hence the term White Knight). I plan to be involved in this quite a bit, after it gets rolling -- so I probably won't be doing that final reveal.
But, we'll see how this progresses. We may change the ending completely and not have a White Knight. We'll certainly have a "big reveal", but we may pull it off in a way completely different than I wrote it in the script. We'll make the call later.
Since April 1st is the beginning of a new month -- I may post a red herring list of usage stats. With the following pokemon at the top of the OU list:
1. Azelf
2. PorygonZ
3. Rotom-A
4. Infernape
5. Latias
6. Flygon
7. Octillery
8. Omastar
9. Lucario
10. Scizor
This will give people in the forum something to chew on, just in case they don't visit the website (which will be in Portuguese). I'm sure many people will take great pride when they "figure it out" after deciphering the list.
HINT: For those of you that didn't get it -- read the first letter of each pokemon in the top ten...
1. Azelf
2. PorygonZ
3. Rotom-A
4. Infernape
5. Latias
6. Flygon
7. Octillery
8. Omastar
9. Lucario
10. Scizor
This will give people in the forum something to chew on, just in case they don't visit the website (which will be in Portuguese). I'm sure many people will take great pride when they "figure it out" after deciphering the list.
HINT: For those of you that didn't get it -- read the first letter of each pokemon in the top ten...
I know I'm not a cast member per se, but I love the idea of this prank and I'm a well known whinger about hax and so I think it would be good if I posted supporting the formula.
Hope that it doesnt matter that I didnt get instructions, cause it'd be kinda suss that someone who is known to hate hax so much would not offer an opinion on this.
Hope that it doesnt matter that I didnt get instructions, cause it'd be kinda suss that someone who is known to hate hax so much would not offer an opinion on this.
Welcome to the show, MS. I added you to the cast list and sent you a Casting Call PM. Break a leg!
Oh, come on, Omastar demands 7th.
Cheers Doug :)
For my role, I have gotten a longish reason as to why to implement the formula that I typed up tonight. Posting here to:
a) Receive the green light for when to post this
AND
b) Receive suggestions on how to improve it
Argument for the Formula
I have thought about this idea for a great deal now and I’ve come to the conclusion I’m a strong advocate for it. It enriches our metagame both through minimizing luck and increasing the variety present. For example, with the new hax formula, stall is a much more viable style of play. I have had massive troubles in the past where I have thoroughly outplayed my opponent only to have been done in by a lucky critical hit. Conversely, I’ve been outplayed by people running a stall team thoroughly only for my MixMence’s Draco Meteor to critical hit his Skarmory and faint it, undeservedly opening his team up to a sweep for another member even though his team is structurally more sounds then mine. I am sick of a metagame that is so centered around revenge killing and “checks” rather then the good old fashion way of being an “100 percent counter” to something, and this formula would encourage more solid team structure as well as promote the use of a relatively rare playstyle.
Similarly, bulky stat-uppers that use moves such as Bulk Up and Calm Mind aren’t used in lieu of their more offensive cousins in Tail Glow, Swords Dance, Dragon Dance and Nasty Plot. Moves like Calm Mind and Bulk Up are much better to sweep with (except for possibly Dragon Dance, due to the extra speed.) due to padding your defenses as well as your offenses. Stuff like Calm Mind Jirachi, Celebi and Suicune as well as Bulk Up Machamp and Gallade are personal favourites of mine, as they are very hard to OHKO as it is, and with even + 1 or + 2 in a stat, it becomes nigh impossible. However, because of rubbish like an Infernape’s Grass Knot Critical Hit my + 6 Suicune when all he has left is a Heatran and Scizor happening, I have been deterred from them. This new hax formula will give this stagnant metagame much needed diversity in terms of a whole new set of stat uppers becoming competitively viable as well as the almost dead art of stall.
This and the fact that it takes the rating of the players in question involved to help determine the winner means that the guys who are better players will always win their matches.
For all those opposing the implementation of the formula, isn’t minimizing luck to give the more skilled players the best chance of winning what Smogon has been striving to do all these years? We don’t play pokemon competitively to hope to get a lucky crit, we play so that we can beat our opponents through the skill of our tactics. Just something to think about for everyone.
For my role, I have gotten a longish reason as to why to implement the formula that I typed up tonight. Posting here to:
a) Receive the green light for when to post this
AND
b) Receive suggestions on how to improve it
Argument for the Formula
I have thought about this idea for a great deal now and I’ve come to the conclusion I’m a strong advocate for it. It enriches our metagame both through minimizing luck and increasing the variety present. For example, with the new hax formula, stall is a much more viable style of play. I have had massive troubles in the past where I have thoroughly outplayed my opponent only to have been done in by a lucky critical hit. Conversely, I’ve been outplayed by people running a stall team thoroughly only for my MixMence’s Draco Meteor to critical hit his Skarmory and faint it, undeservedly opening his team up to a sweep for another member even though his team is structurally more sounds then mine. I am sick of a metagame that is so centered around revenge killing and “checks” rather then the good old fashion way of being an “100 percent counter” to something, and this formula would encourage more solid team structure as well as promote the use of a relatively rare playstyle.
Similarly, bulky stat-uppers that use moves such as Bulk Up and Calm Mind aren’t used in lieu of their more offensive cousins in Tail Glow, Swords Dance, Dragon Dance and Nasty Plot. Moves like Calm Mind and Bulk Up are much better to sweep with (except for possibly Dragon Dance, due to the extra speed.) due to padding your defenses as well as your offenses. Stuff like Calm Mind Jirachi, Celebi and Suicune as well as Bulk Up Machamp and Gallade are personal favourites of mine, as they are very hard to OHKO as it is, and with even + 1 or + 2 in a stat, it becomes nigh impossible. However, because of rubbish like an Infernape’s Grass Knot Critical Hit my + 6 Suicune when all he has left is a Heatran and Scizor happening, I have been deterred from them. This new hax formula will give this stagnant metagame much needed diversity in terms of a whole new set of stat uppers becoming competitively viable as well as the almost dead art of stall.
This and the fact that it takes the rating of the players in question involved to help determine the winner means that the guys who are better players will always win their matches.
For all those opposing the implementation of the formula, isn’t minimizing luck to give the more skilled players the best chance of winning what Smogon has been striving to do all these years? We don’t play pokemon competitively to hope to get a lucky crit, we play so that we can beat our opponents through the skill of our tactics. Just something to think about for everyone.
Ok wow. In case I didn't make myself clear: wow.
I don't get why translate the front page to portuguese but if Haze and Thumpsday want some silly slangs or wording contact me.
Edit: maybe you guys could talk over making this thread public after the joke is revealed. To make people realize how much effort was put into a simple April Fools or something.
I don't get why translate the front page to portuguese but if Haze and Thumpsday want some silly slangs or wording contact me.
Edit: maybe you guys could talk over making this thread public after the joke is revealed. To make people realize how much effort was put into a simple April Fools or something.
I like that argument MS. As for when to post it, I think you should post it later today or tomorrow. Since there is plenty of negative sentiment in the thread, I think some general "hax-is-bad theory" like you have written, would be a help. If some of the extras and drive-by heroes want to toss a few comments in after MS, I think it would be appropriate.
We need to get some more Mad Scientists into the mix. I'll PM a cue to Caelum and Brain. Also I know Obi is planning to jump in here somewhere. We need to get some more tangible suggestions for improving the formula. We'll kick off Act 2 on the 16th, so we need to get some more foundation in over the next two days.
Another small script change -- later on, when we announce the date this will go on the server, I think we should announce March 28th as the go-live date. Then on the 28th, I'll come up with an excuse to delay the implementation.
By doing this, I think we will completely throw off any people that might suspect this whole thing is "strangely coincidental with April Fools Day". No one will possibly suspect it, if all the lead-in talk is centered around March 28th. So, when Act 3 kicks off -- expect to see March 28th widely touted as "the day".
We need to get some more Mad Scientists into the mix. I'll PM a cue to Caelum and Brain. Also I know Obi is planning to jump in here somewhere. We need to get some more tangible suggestions for improving the formula. We'll kick off Act 2 on the 16th, so we need to get some more foundation in over the next two days.
Another small script change -- later on, when we announce the date this will go on the server, I think we should announce March 28th as the go-live date. Then on the 28th, I'll come up with an excuse to delay the implementation.
I'll keep doing that for a few days...
You get the idea.
By doing this, I think we will completely throw off any people that might suspect this whole thing is "strangely coincidental with April Fools Day". No one will possibly suspect it, if all the lead-in talk is centered around March 28th. So, when Act 3 kicks off -- expect to see March 28th widely touted as "the day".
Actually, I think you should implement it March 28th. The reason for this is that A. Delaying it to April 1st seems way too obvious to me, and people will be suspicious, and B. It usually takes a couple days for a metagame to settle. We'd be able to see people's initial reactions in addition to their secondary reactions after living with it for a couple days. Who knows, maybe some non-badged opponents will actually LIKE the new metagame? (or go ballistic.)
Here's what I'm thinking of posting:
A bunch of people have been linking me to this thread, asking me what I think about it. I'm not really sure how good of an idea it would be to use this to determine who wins just yet, however, I do have some ideas on how to make it better.
One of my favorite games of all time is Lords of Magic. You build up armies and fight other armies. To fight, you have two options; you can enter the battle and micromanage your armies, or you can press "Auto calc". Auto calc compares the relative strength of the two armies and determines which side wins (it also determines the damage done to the survivors). Perhaps we could do something similar to that here?
I'm thinking of running through each turn several times to see just what the "expected" outcome is. Of course, it would be too computationally intensive to simulate the full battle, but you can do some short cuts that don't make sense in any given turn, but in the aggregate approximate averaging out thousands of battles.
Some people claim that making a "Pokemon AI" like this would be too hard, but that's only true if the bot has limited information. The combinatorial explosion of Pokemon turns actually results from random numbers and imperfect information. In any given turn, there are actually only 81 possible moves max (usually less), multiplied by the effects of luck. If we can find a way to "normalize" this luck, that could be used as part of a larger scheme to simplify the calculations to make this sort of predictive algorithm effective.
What I'm thinking is using this to find the expected outcome of a few turns, determining who is in the advantage those turns ahead, and combining that with their relative rating to find out who "should" win. The amount that this differs from what actually happened determines the effect luck has had on the game.
This all depends on how efficient of an AI algorithm we can make, of course.
A bunch of people have been linking me to this thread, asking me what I think about it. I'm not really sure how good of an idea it would be to use this to determine who wins just yet, however, I do have some ideas on how to make it better.
One of my favorite games of all time is Lords of Magic. You build up armies and fight other armies. To fight, you have two options; you can enter the battle and micromanage your armies, or you can press "Auto calc". Auto calc compares the relative strength of the two armies and determines which side wins (it also determines the damage done to the survivors). Perhaps we could do something similar to that here?
I'm thinking of running through each turn several times to see just what the "expected" outcome is. Of course, it would be too computationally intensive to simulate the full battle, but you can do some short cuts that don't make sense in any given turn, but in the aggregate approximate averaging out thousands of battles.
Some people claim that making a "Pokemon AI" like this would be too hard, but that's only true if the bot has limited information. The combinatorial explosion of Pokemon turns actually results from random numbers and imperfect information. In any given turn, there are actually only 81 possible moves max (usually less), multiplied by the effects of luck. If we can find a way to "normalize" this luck, that could be used as part of a larger scheme to simplify the calculations to make this sort of predictive algorithm effective.
What I'm thinking is using this to find the expected outcome of a few turns, determining who is in the advantage those turns ahead, and combining that with their relative rating to find out who "should" win. The amount that this differs from what actually happened determines the effect luck has had on the game.
This all depends on how efficient of an AI algorithm we can make, of course.
- Status