| View previous topic :: View next topic |
| Author |
Message |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 02, 2012 1:53 pm Post subject: Aug 19 - Sept 9 :: RAW 1.5 Beta Testing |
|
|
All the gory details — including the old formula — beginning Monday.
Last edited by paranoiarodeo on Thu Sep 13, 2012 11:41 pm; edited 4 times in total |
|
| Back to top |
|
 |
WhirlPlaid Lux Pacifist

Joined: 30 Aug 2007 Posts: 2427 Location: Waiting for a Cube Wars game.
|
Posted: Thu Aug 02, 2012 3:48 pm Post subject: |
|
|
| All maps? Every room? or only para hosts? |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 02, 2012 5:35 pm Post subject: |
|
|
Good questions. I haven't made any final decisions yet.
(I will be scoring kill bonuses, so yeah, überhost only.) |
|
| Back to top |
|
 |
Mumps Disease of Lux

Joined: 13 Oct 2010 Posts: 303 Location: Oz
|
Posted: Fri Aug 03, 2012 1:39 am Post subject: |
|
|
∞ Mumps reserves seats ∞  |
|
| Back to top |
|
 |
luiz zenni Lux Newbie

Joined: 23 Oct 2011 Posts: 190 Location: Brazil
|
Posted: Sat Aug 04, 2012 8:16 am Post subject: |
|
|
WOW Reserve a seat. BOOOOOOMMMBOMMMMMMM  |
|
| Back to top |
|
 |
WhirlPlaid Lux Pacifist

Joined: 30 Aug 2007 Posts: 2427 Location: Waiting for a Cube Wars game.
|
Posted: Mon Aug 06, 2012 11:08 am Post subject: |
|
|
| Those two weeks are already really booked up for me... but I will try to get in as many games as I can. The potential has me both curious and excited for the prospects of Lux. |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Mon Aug 06, 2012 12:54 pm Post subject: |
|
|
Lemme get a few things outta the way. First, I'm gonna be fairly casual about all of this. The formal textbook explanations will be saved for later. I'm just gonna toss little pieces of info out there one by one. If you have any questions, then please speak up! At the end of these two weeks, I hope no one is confused about the old formula, my new formula, and why I've made the changes that I've made. Second, my beta test month could be a disaster! Theoretically, I'm confident that my ideas and tweaks will work, but one never knows until it's actually out there, living and breathing. I can predict only so well how players themselves will adapt and game against the system. Third, to quote our beloved American Vice-President: "This is a big fucking deal." IMHO, RAW is an inextricable part of Lux's DNA, and messing with the "secret formula" is dangerously close to blasphemy. I once asked dustin for the details, so I could help improve things behind the scenes, but his only answer was: "It's not that simple." Umm, okay. Therefore, everything I'm about to explain to you was reverse engineered without help. In other words: I stared at the numbers, found the patterns, alone, and now I'm sharing that all with you. Granted, it wasn't Bletchley Park level codebreaking work, but nonetheless, I feel like I'm giving something away for free. There's something wonderful about discovering a secret on your own, and keeping that secret close. It's a fun party trick to be able to shout out scores before they're officially posted. I guess anyone will be able to do that now, huh? Oh well.
∞ paranoiarodeo sharpens his pencil ∞ |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Tue Aug 07, 2012 2:12 pm Post subject: |
|
|
Let's start with the basics:
1. Standard result.
If you've ever payed attention to the results, then these numbers should be familiar to you: +90 / +30 / +15 / 0 / -15 / -60. That's the baseline scoring for an average six human game when all players have similar RAW. I'm gonna change that slightly: +90 / +27 / +9 / -9 / -27 / -45. In other words: You'll win less for second and third, lose more for fourth and fifth, and lose less for sixth. The "overall point inflation per game" will be lowered from +60 to +45. That's a 25% reduction in the upward pressure on all scores.
(Biodeux scores reduced to 80% of baseline 'cause of the 30% cont income setting. More on that later.)
 |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Tue Aug 07, 2012 11:21 pm Post subject: |
|
|
2. Random tidbits. Original formula.
You can not lose points for second place.
(That rule was added on August 21, 2006.)
Max wins for 1st, 2nd, and 3rd: +222 / +66 / +33.
(That rule was added on March 23, 2010. Unlimited upside prior to change.)
Max loss for any place: 10%.
(That rule was added way back in the pre Lux 5.0 days. Maybe 2004?)
500 is reserved for inactive players.
(If you end up at 500, then you'll be given an extra point for 501. Same goes for first games played in "unranked" rooms.) |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Wed Aug 08, 2012 1:38 pm Post subject: |
|
|
3. Calculating the winner's score. Original formula.
(Time for some meat and potatoes. Yum.)
Again, let's assume six humans in standard Classic. No fussing with reductions 'cause of fast timers, higher cont income settings, bots, etc. (Yet.) Basically, the RAW formula uses the scores of players ahead of you, and behind you, to determine your score. When calculating how much someone wins for second, third, fourth, or fifth place, then you will win more points if higher scores place below you. Who gets killed in what order does matter. However, for first and sixth place, the final scores will be the same, regardless of kill order.
Now, if you've been playing a long time, then these numbers should be familiar to you too: +90 / +72 / +54 / +36 / +18 / +1. Those are the most common scores for winners, and as you may have noticed, they're all multiples of 18. Sooo, why is that?
(Biodeuxers: +73 / +58 / +44 / +29 / +15 / +1.)
See that nice flat line? If your opponent's RAW score is between "400 points behind" and "120 points ahead" of yours, then you receive 120 points. If you're 401 or more points ahead of someone, then you get zero. (That's what we call a hard cutoff. Ouch.) And if your opponent is 121 or more points ahead of you, then you get 121 or more points, on a linear one-to-one basis with no maximum ceiling. Sooo, you take each of your five opponents scores relative to yours, look at the graph, figure out how many points you get for each one, total up the numbers, then multiply by 0.15.
Any player who's been on a hot winning streak knows what happens: You win +90 a couple of times, then suddenly, you run up against that wall, and your wins begin dropping by 18 points. Why? 'cause your opponents, one by one, have fallen more than 400 points behind you.
Logically absurd example: If you begin a game with 899 points, and win against five humans with 500 points, then you get 120 points for all five opponents, equalling 600. Take that total and multiply by 0.15, and you get 90 points. Bingo. However, if you had 901, then you woulda won zero (+1).
Here are five recent games as a quick example of the calculation:

Last edited by paranoiarodeo on Thu Aug 09, 2012 3:47 pm; edited 4 times in total |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Wed Aug 08, 2012 4:49 pm Post subject: |
|
|
4. Calculating the sixth place score. Original formula.
Same idea as first, but kinda mirror imaged.
If a player placed above you is more than 400 points ahead, then you lose nothing against them.
If they're between 120 behind and 400 ahead, then you lose 120 points.
If they're more than 120 behind, then you lose the full amount.
And since the standard result is -60 instead of +90, then we multiply the sum by 0.1 instead of 0.15.
(Remember there's a 10% max loss cap.)
Examples:

Last edited by paranoiarodeo on Thu Aug 09, 2012 3:48 pm; edited 1 time in total |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Wed Aug 08, 2012 5:43 pm Post subject: |
|
|
(Footnote.)
dolla asked me a good question: "Where's the formula?" Here's the thing, folks: "There's no formula, per se." Nothing I can scribble up on the chalkboard, like algebra class in high school. No simple, glorious unified equation that spits out an answer like 42. Sooo, if that's what you're looking for, or expecting, then I'm gonna disappoint. The closest thing you'll get is a 6x6 grid, with those "between -400 and 120" or "between -120 and 400" ranges implicit, as seen in the two major graphs posted so far. I'll do my best to make that make sense when the time comes.
(Just wait until we get to the bot game scoring. Oy.) |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 09, 2012 1:30 pm Post subject: |
|
|
5. More random tidbits. Original formula.
If you lose points for third place, then your loss is cut in half.
(Last five years: 28861 losses for third place outta 362453 FH games in the last 5 years. Roughly 8%.)
All numbers are rounded up in every circumstance.
(Also known as a ceiling function.)
There's some "imperfect decimals" in the formula caused by "floating point imprecision."
(For example: Standard 30% Biodeux win should be +72, but it's +73. Why? 'cause the result is reduced by 19.999...% not 20%.) |
|
| Back to top |
|
 |
dollabillz Lux Troll

Joined: 03 Feb 2009 Posts: 1050
|
Posted: Thu Aug 09, 2012 3:11 pm Post subject: |
|
|
 |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 09, 2012 3:29 pm Post subject: |
|
|
 |
|
| Back to top |
|
 |
dollabillz Lux Troll

Joined: 03 Feb 2009 Posts: 1050
|
Posted: Thu Aug 09, 2012 3:35 pm Post subject: |
|
|
 |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 09, 2012 4:31 pm Post subject: |
|
|
I updated the two major graphs up above with colors.
Extra points to anyone who can figure out the "one thing" I've left out.
∞ paranoiarodeo dares Luxworld to solve the riddle ∞
(Hint: The multiplier for second through fifth places is the same.) |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 09, 2012 4:32 pm Post subject: |
|
|
ps. You've got until Saturday afternoon US time.
(Bueller? Bueller? Bueller?) |
|
| Back to top |
|
 |
el toro Met Dustin

Joined: 28 Jul 2004 Posts: 4648 Location: chi-town
|
Posted: Thu Aug 09, 2012 5:41 pm Post subject: |
|
|
| uh bacon? |
|
| Back to top |
|
 |
engie Lux Dynamo

Joined: 02 Aug 2010 Posts: 641 Location: Tejas
|
Posted: Thu Aug 09, 2012 7:03 pm Post subject: |
|
|
For 2nd-5th, you need to sum the "120"s and then ADD another 120 to each score. Then, you multiply each of the sums by 0.0625 (or 1/16)
Ta-da!  |
|
| Back to top |
|
 |
engie Lux Dynamo

Joined: 02 Aug 2010 Posts: 641 Location: Tejas
|
Posted: Thu Aug 09, 2012 7:04 pm Post subject: |
|
|
| (except you add a negative 120 for 5th place, since it's a negative sum) |
|
| Back to top |
|
 |
imapickle is a pickle

Joined: 10 Aug 2009 Posts: 619 Location: in the jar next to the marmalade
|
Posted: Thu Aug 09, 2012 7:04 pm Post subject: |
|
|
Dunno how to do symbols on here so sorry for the latex notation. If I understood the graphs properly the winners score change formula for the old system is
\Delta x_1 = 0.15 [\sum_{i=2}^{6} 120 \Theta(x_i -x_1 +400) + (x_i -x_1-120)\Theta(x_i - x_1 -120) ]
where \Theta(x) = 0 if x<0>0 is the Heaviside step function.
Here we should take
\Theta(0) =0 since the hard cut-off hits you when the difference is equal to 400.
Edited. Should be correct now.
Last edited by imapickle on Thu Aug 09, 2012 8:19 pm; edited 4 times in total |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 09, 2012 7:04 pm Post subject: |
|
|
∞ paranoiarodeo sounds the wrong answer buzzer ∞
(You compare your guess against old game results? Nope.) |
|
| Back to top |
|
 |
imapickle is a pickle

Joined: 10 Aug 2009 Posts: 619 Location: in the jar next to the marmalade
|
Posted: Thu Aug 09, 2012 7:06 pm Post subject: |
|
|
| engie wrote: | | (except you add a negative 120 for 5th place, since it's a negative sum) |
Engie, if you must make minus sign errors, pls try to make an even number of them. |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 09, 2012 7:12 pm Post subject: |
|
|
Protip: Scroll back through these results, and look for FHs with cont incomes increases of 10% or less.
(Hosted by para or iggy.) |
|
| Back to top |
|
 |
WhirlPlaid Lux Pacifist

Joined: 30 Aug 2007 Posts: 2427 Location: Waiting for a Cube Wars game.
|
Posted: Thu Aug 09, 2012 7:17 pm Post subject: |
|
|
| paranoiarodeo wrote: | Protip: Scroll back through these results, and look for FHs with cont incomes increases of 10% or less.
(Hosted by para or iggy.) |
No! I refuse to! I only want to play my games... I prefer to let para do all the work. |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 09, 2012 7:33 pm Post subject: |
|
|
| ∞ paranoiarodeo does all the Lux geek work so Whirly doesn't have to ∞ |
|
| Back to top |
|
 |
imapickle is a pickle

Joined: 10 Aug 2009 Posts: 619 Location: in the jar next to the marmalade
|
Posted: Thu Aug 09, 2012 7:54 pm Post subject: |
|
|
According to that example, and using the same cutoff at 400 etc, the second place score change formula for the old system should be
\Delta x_2 = [ - 120\Theta(x_2 -x_1 -400) - (x_2 - x_1 + 120)\Theta(x_2 -x_1 + 120)] /20+ 3[ \sum_{i=3}^{6} 120 \Theta(x_i -x_2 +400) + (x_i-x_2-120)\Theta(x_i - x_2 -120) ]/40
where \Theta(x) is the Heaviside step function.
Edit:
The conversion factor is 1/20 for negative contribution (i.e. from the 1st place player) and 3/40 for +ve contributions to give (120*4)*(3/40)-120*(1/20) = 30, the baseline for 2nd.
This would be for classic settings. As i understand, bio just has a different overall factor in the raw change based on cont increase etc.
Formulae for 3rd-6th follow similary.
Last edited by imapickle on Thu Aug 09, 2012 10:00 pm; edited 7 times in total |
|
| Back to top |
|
 |
engie Lux Dynamo

Joined: 02 Aug 2010 Posts: 641 Location: Tejas
|
Posted: Thu Aug 09, 2012 7:55 pm Post subject: |
|
|
Attempt #2!
1. Determine # of points to be deducted from each person who finish ahead of you... (deduct 120 for RAW that is 120 above to 400 behind, etc) and determine the number of points that need to be added (inverse of point deductions as para described earlier.)
2. Sum up these points.
3. Next, add the sum to the average points added/deducted for that game. (for example if the point additions and deductions were: 120, 120, 120, 200, and 214 the average would just be these numbers added together and then divided by 5.)
So, so far we have (Sum+Average)
4. Now, multiply (Sum + Average) * 0.0625
5. Finally, round to the nearest whole number and add ONE point. |
|
| Back to top |
|
 |
paranoiarodeo Lux Moderator


Joined: 17 Jul 2005 Posts: 8194 Location: Perched Upon the Eternal Throne of Lobotomia
|
Posted: Thu Aug 09, 2012 7:56 pm Post subject: |
|
|
∞ paranoiarodeo bzzzts ∞
(Strike two.) |
|
| Back to top |
|
 |
|