NationStates Jolt Archive

Dance2Spaam

Dance 2 Revolution & Spaam for World Cup XXXI

Welcome to the Dance2Spaam bid! Spaam has been heavily involved with the NSWC community for over 100 years (3 RL), and successfully hosted many tournaments, including three Cups of Harmony. D2R has been RPing for over 80 years (2 RL) and is a consistent and well liked RPer, best known for winning the first Baptism of Fire. Together they will attempt to create a fun and exciting tournament for all WC nations.

The format of the qualifying competition will be 12 groups of 8, with the top two teams of each group qualifying, and the third place teams participating in a home and away fixture for wild-card entry. The final competition will be as per normal. We will also be providing a number of off days for teams to participate in friendlies. In addition there may be other special matches held, such as the Wooden Spoon match, to determine the worst team in the world. (Scorination and RP bonuses are discussed later in this thread)

Any questions anyone may have, please, feel free to ask!

<stadiums and other information to come>

Time to unveil the new Spaamanian Scorinator!

http://spaam.info/worldcup/scorinate.cgi

Give it a whirl and tell us what you think.

Features:

No limit on styles! Any real number is valid, even the uber karelan -1000000
No disadvantage to low ranked offensive or high ranked defensive teams: offense and defense win equally
Styles actually MEAN something! If both teams have max defense, 1/6 matches will be scoreless draws. If both teams have max offense, you just might see one team beating another 11:5! (Thats the biggest we've seen from simulation)


If you want any statistics, ask, and I shall generate.

So what makes this scorinator different from the myriad of others already in existance? Eg, NSFS, Leagion, Scihia/Qutars, Vil/TLC's, etc... What's the general algorithm?

And aren't the statements "Offense and defense win equally" and "If both teams have max defense 1/6 matches will be scoreless" contradictory? This states that styles will not affect the chances of winning or losing AND states that ultra defensive will mean 1 out of six matches scoreless draws.

The average number of points won is the same whether one is defensive of offensive or neither. It works by decreasing the advantage that a high ranked offensive team would have, or a low ranked defensive team.

The difference is the fact one is not disadvantaged what style one chooses, the difference between KPB ranks means nothing - its the RATIO, and that everything in this scorinator is based on mathematical principles :p

The general algorithm is this: (you would've worked it out if you looked at the code though :p)

style = 0.32 * (arctan(astyle/2) + (arctan(bstyle/2))
All this formula does is turn any style into a number between -1 and 1

z = (3 + style) * (a / (a + b)) ^ (1 - 0.3 * style) / 15
The (3 + style) part is the average number of goals per game.
The a/(a+b) is the percentage of games Team A would win - if one team has a KPB rank of 20, and the other 10, then the one with 20 should win 20/(20+10) = 2/3 of games between them, or twice as many.
The ^ (1 - 0.3 * style) part just adjusts the percentage to fix up advantages/disadvantages related to style.
And the 15 is the number of attacks.

Generate a random number between 0 and 1, and if it is less than z, score a goal. Repeat 15 times.

a/(a+b) is just a ridiculous formula. It makes it WAY harder for lower ranked teams. the difference between a team with 20 points and 10 points is ten points, but in terms of times it's just two times. But if you take that same 10 point difference between a team with 11 pts and 1 pt you get a multiplier of 11. I would argue that there's actually a bigger difference in ten points at the top than at the bottom, but teams at the bottom get unfairly punished by this system.

I think there ought to be some sort of system that in general pushes teams that RP up and teams that don't down, but this looks like it makes it very difficult for teams to move up. It seems like the opposite of progress.

a/(a+b) is just a ridiculous formula. It makes it WAY harder for lower ranked teams. the difference between a team with 20 points and 10 points is ten points, but in terms of times it's just two times. But if you take that same 10 point difference between a team with 11 pts and 1 pt you get a multiplier of 11. I would argue that there's actually a bigger difference in ten points at the top than at the bottom, but teams at the bottom get unfairly punished by this system.

I think there ought to be some sort of system that in general pushes teams that RP up and teams that don't down, but this looks like it makes it very difficult for teams to move up. It seems like the opposite of progress.
I'll describe the RP bonus system soon. You'll see that it works well.

a/(a+b) is just a ridiculous formula. It makes it WAY harder for lower ranked teams. the difference between a team with 20 points and 10 points is ten points, but in terms of times it's just two times. But if you take that same 10 point difference between a team with 11 pts and 1 pt you get a multiplier of 11. I would argue that there's actually a bigger difference in ten points at the top than at the bottom, but teams at the bottom get unfairly punished by this system.

I think there ought to be some sort of system that in general pushes teams that RP up and teams that don't down, but this looks like it makes it very difficult for teams to move up. It seems like the opposite of progress.
Before I comment on the RP bonus system, I would just like to note that I would expect a team ranked 53rd (11 KPB points) to beat a team ranked 167th (1 KPB point), a lot more than a team ranked 31st (21 KPB points) over a team ranked 53rd (11 KPB points). So your comment is rather faulty.

I assume in the Spaamanian Scorinator, rank = KPB point and not the rank eg 1st.

First up... all these new-fangled RP bonus methods and whatnot are IMHO too complicated and subjective. Ours will be a lot simpler, less subjective, and should work well.

We are not going to judge quality of RPs indepth, because when it comes down to it, its not going to matter when random numbers come into it, and a good RP to some ppl can seem like a bad to others. The RPing is about fun, and all the RP bonus is for, is to encourage RPing. If someone makes a post, they should get a reward. However, we will differentiate between two types of RPs: lazy and normal.

This is an example of a lazy RP: http://forums.jolt.co.uk/showpost.php?p=11627739&postcount=558

This is an example of a normal RP: http://forums.jolt.co.uk/showpost.php?p=11628156&postcount=561

Lazy: http://forums.jolt.co.uk/showpost.php?p=11627535&postcount=553

Normal: http://forums.jolt.co.uk/showpost.php?p=11627538&postcount=554

A normal RP will get a fixed RP bonus which is the same for all teams. This bonus will obviously help lower ranked teams more, thanks to the ratio method of scorination. A lazy RP will get a fraction of this RP bonus (for example, 1/4). Thus it is worth it to put that little bit more work in to get at least four times more RP bonus. Of course, the RP bonus will be cumulative, and thus have more of an effect towards the end of qualification. And no more than one RPbonus per matchday (to prevent Spaaming :D)

Now, how will this affect qualification? Here we are going to assume 15 groups, 8 teams per group, top two qualify. Simulations are over 10000 groups (over half a million games). Please note that though this data comes from simulation, we reserve the right to fine-tune variables before the WC starts.

If noone RPs...
seed %qual
1 95.8
2 80.3
3 20.0
4 3.8
5 0.1
6 0.0
7 0.0
8 0.0
Fairly realistic numbers.

If everyone RPs...
seed %qual
1 93.6
2 74.3
3 23.7
4 6.8
5 1.2
6 0.3
7 0.1
8 0.0

Lower ranked teams have a much bigger chance of qualifying, to a slight detriment to the top two seeds.

If only one team in a group RPs...

Seed 1 2 3 4 5 6 7 8
1 97.7 94.7 90.9 92.3 95.3 96.2 96.3 96.1
2 77.4 88.1 63.1 69.8 76.0 78.3 79.4 80.3
3 20.7 14.5 43.0 16.1 18.4 19.0 19.9 19.6
4 4.4 2.8 2.0 21.9 3.3 3.1 3.2 3.5
5 0.1 0.1 0.1 0.1 7.1 0.1 0.1 0.1
6 0.0 0.0 0.0 0.0 0.0 3.3 0.0 0.0
7 0.0 0.0 0.0 0.0 0.0 0.0 1.3 0.0
8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.5

You can see if you're the only one in your group to RP, you have a MUCH bigger chance of qualifying (over twice if you're a third seed, over 5 times if you're a fourth seed, etc...)

So we have just shown that the RP bonus DOES have a significant effect, while staying 'realistic'.


And Jeruselem: Yes. I call it rank cos there's no better word for it :p

This is an example of a normal RP: http://forums.jolt.co.uk/showpost.php?p=11628156&postcount=561

Lazy: http://forums.jolt.co.uk/showpost.php?p=11628156&postcount=561



While you may find other RP calculation methods as overly complicated, I fail to see how this method of considering RP, where the same post is deemed to fall into both the Normal & Lazy categories, is any more clear cut than the more detailed methods which involve the hosts taking more time to actually read the RPs and judge them as opposed to taking a glance at their sustinance and classifying them in a binary on or off-type grading system.

:)

Yeah I realised it about a minute before you and fixed up the post :p

The hosts do have to read them, but it won't take an hour to do scorination each day like your method. Plus, can we have stats like mine for your system? :)

Yeah I realised it about a minute before you and fixed up the post :p

The hosts do have to read them, but it won't take an hour to do scorination each day like your method. Plus, can we have stats like mine for your system? :)

Scorination doesn't take an hour in either method, but I'd like to think on a good day we're all reading everyone's RPs that they are putting time and effort into, not just the hosts. As for the stats, I suppose If I coded some more I could output some stats, but They are the same basic sims that I've posted in our bid thread, and in the end the scenarios are so extreme they would hardly be duplicated in the actual cup anyway :)

I think we need more information about the RP bonus... its all rather airy-fairy. We wouldn't mind knowing the actualy scorination formula as well.

As for your sims, its hard to actually tell what the results would be. We know basically nothing about it apart from its this wonderful new method that will admittedly benefit RPers. I think some sort of comparison is in order...

I think we need more information about the RP bonus... its all rather airy-fairy. We wouldn't mind knowing the actualy scorination formula as well

Indeed. Please post them Spaam. And by the formula I mean with RP bonuses added in.

And as for the post about 21 to 11 being bigger than 11 to 1 is this.

The most amount of points you can without qualifying for the world cup proper is just about 9.5. Two decent qualifying runs and one lucky slip in to the cup can get you to 11 pretty easily. But to get to 21 you have to consistently make the cup and consistently get points off the top teams a much more difficult task. So in some sense it's much harder to make the jump from 11 to 21 than it is to make the jump from 1 to 11. Even if you were to accept that it was equally hard to jump from 11 to 21 as to jump from 1 to 11, the A/(A+B) formula is still unfair. There is just no way that the difference between a team with 11 KpB and 1 KpB is more than 5 times larger than the difference between a team with 21 KpB and 11 KpB.

Indeed. Please post them Spaam. And by the formula I mean with RP bonuses added in.

And as for the post about 21 to 11 being bigger than 11 to 1 is this.

The most amount of points you can without qualifying for the world cup proper is just about 9.5. Two decent qualifying runs and one lucky slip in to the cup can get you to 11 pretty easily. But to get to 21 you have to consistently make the cup and consistently get points off the top teams a much more difficult task. So in some sense it's much harder to make the jump from 11 to 21 than it is to make the jump from 1 to 11. Even if you were to accept that it was equally hard to jump from 11 to 21 as to jump from 1 to 11, the A/(A+B) formula is still unfair. There is just no way that the difference between a team with 11 KpB and 1 KpB is more than 5 times larger than the difference between a team with 21 KpB and 11 KpB.
If you read the descriptions, you'd see I have given the formula, with RP bonuses. Unlike V/TLC.

As for all the crap you're spouting... you've been here, how long? And you're trying to tell us how it works? Do you know how EASY it is to get 1 KPB point? 1 KPB point is about 3.5 points in qualifying, or a win and a draw. Or about 10 points in the BoF. If you CAN'T get 1 KPB point, then you are honestly a crap team. The second worst team in the World Cup qualifiers gets more than 1 KPB point. And again, this is supposed to reflect RL. I would expect Korea to beat the British Virgin Islands about 5 times more than Poland to beat Korea. Which is what we're talking about.

Though I like the 12 groups of 8 way of running a World Cup, surely a cup of such a size is a little dated, considering that the last two cups have had 122 teams in them, as opposed to the 98 that this group structure would permit. If more than 98 sign up, will you be willing to expand the qualifying format to allow more teams in or will you rigidly stick to the limits and prevent any more nations from joining up?

If we really get that many by the end of the finals, then yes. Considering we're only in the low 70s in the moment, we thought 98 teams would be plenty, and allow for playoffs (which we both like).

If we have more, we could upgrade it to 12 groups of 9 (I like byes, though not everyone is crazy about them), 12 groups of 10 (though that would take a long while), 20 groups of 6 (holey moley!), or basically whatever the community thinks would be the best number.

I like the looks of things here except for the fact that the upper two seeds get an enormous advantage to qualify over their counterparts. Trying to keep the hegemons in place, eh? I know that you may want to keep out "surprise" non-ranked qualifiers like Quakmysomethingother, but third and fourth seeds get little love. :D

And what about lower seeds who are serious about the WC and RP consistently and well? Surely they should deserve a little more than the pittance of a 0.5%/1.3%/3.3% chances of qualifying (and those chances are when they are the only ones RPing...). One particular player I know who is doing such in this cup is Errinundera, and Iansisle had some great RPs in WC29 too when he came back as unranked.

Perhaps you could rank RP Bonuses on an exponential graph. For example, arrange them on a graph with lower limit 0.5 and upper limit 1.0; the graph should give a little more weight on RPs to the lower teams. The top seeds, who (usually) have so many KPBs so that RPing is really not really necessary, could have the bonus at 0.51, second seeds, 0.56, thirds 0.68, fourth, 0.74, fifth, etc. etc. until you reach the bottom seeds at 0.999999999... or whatever other limit you could make. I don't think the values listed come out to a perfect exponential, but you know what I mean. Of course, if the bottom seed or any other player for that matter posts a "lazy" RP, then their bonus could be minimalized.

Of course, you could do the bonus by rank in the same way, in case the top seed is in the 20s or something.

Very good points!

Firstly, if an max RP unranked team plays a max RP Starblaydia (top ranked team in the world), Starblaydia would only win 4/5 times. A max RP unranked team should be better than a non RP 40th ranked team. If you're in the top 32 but don't RP, you shouldn't complain about being beaten by an RPing team outside the 100.

Second, the percentages mean that on average we'd have 14/15 first seeds qualifying, 11/15 second seeds, 4/15 third seeds, and 1/15 fourth seeds qualifying. So 25% of the qualifiers are upsets. And if any of the top two seeds don't RP, then that just benefits the lower seeds.

Third, we really don't want any first-timers qualifying. But if its their second cup, they have a chance. Quakmybush (I know he's a special case), was a third seed this cup!

Hope this answers your questions :)

...Firstly, if an max RP unranked team plays a max RP Starblaydia (top ranked team in the world), Starblaydia would only win 4/5 times...

Aren't you trying to win votes in this thread? :D

Oh fine. HOWEVER, if Starblaydia faced an non-RPing unranked team, they would win around 96% of their games, lose none, and not concede any goals :p