Scorinators (OOC Discussion)

[Osarius]

there's nothing the NSFS3 application can do that xkoranate can't do better.

Ehh... NSFS3 arguably has an easier-to-use interface.
Particularly if you want multiple conferences with inter-conference fixtures, or a knockout cup.

[Stvoto Latoli]

there's nothing the NSFS3 application can do that xkoranate can't do better.

Ehh... NSFS3 arguably has an easier-to-use interface.
Particularly if you want multiple conferences with inter-conference fixtures, or a knockout cup.

It's true, xkoranate has a steeper learning curve, but it's worth learning. It has a far deeper array of customization options, its scheduler is far better (sometimes in NSFS3 I'd see teams playing four or five games in a row on the road), and can scorinate any sport you want. Most importantly, NSFS3 is okay for doing your domestic leagues and such but it must NEVER, EVER be used during an international tournament like the NSWC, or any other tournament where RP bonuses are updated every matchday, because of its fatal fixed-max-points defect. Because of that, I always encourage people to learn xkoranate. If you have even the slightest desire to ever host something, you need to be familiar with it.

[Eastfield Lodge]

Ehh... NSFS3 arguably has an easier-to-use interface.
Particularly if you want multiple conferences with inter-conference fixtures, or a knockout cup.

It's true, xkoranate has a steeper learning curve, but it's worth learning. It has a far deeper array of customization options, its scheduler is far better (sometimes in NSFS3 I'd see teams playing four or five games in a row on the road), and can scorinate any sport you want. Most importantly, NSFS3 is okay for doing your domestic leagues and such but it must NEVER, EVER be used during an international tournament like the NSWC, or any other tournament where RP bonuses are updated every matchday, because of its fatal fixed-max-points defect. Because of that, I always encourage people to learn xkoranate. If you have even the slightest desire to ever host something, you need to be familiar with it.

Actually, that's been fixed. CH released NSFS 3.0.1a which had modifiable max points and style modifiers (IIRC). Irrelevant anyways.
*sheepishly looks at unused xkoranate folder*

[Valanora]

As Eastfield Lodge has pointed out, CH fixed the Fixed Maxpoints and modifier in NSFS3 a long time ago. So there really is no reason not to use it, if it is your preference. I personally prefer NSFS3 over xkoranate unless I am scorinating UICA like events, where the results table is very nice to have.

[Stvoto Latoli]

As Eastfield Lodge has pointed out, CH fixed the Fixed Maxpoints and modifier in NSFS3 a long time ago. So there really is no reason not to use it, if it is your preference. I personally prefer NSFS3 over xkoranate unless I am scorinating UICA like events, where the results table is very nice to have.

I should perhaps not be trusted to be neutral on the topic, as I passionately dislike the NSFS scorination formula itself. Despite the apparently commonly held wisdom that NSFS is less random, when the NSFS formula is compared to the SQIS formula, matches between two lowly ranked teams actually become more random, not less. Only matches between powerful teams at the top and minnows at the bottom become less random, so much so that they are often unrealistic eight or nine nil blowouts (which I'm also not in favor of).

Given my suspicions about the formula (which were present before the max-points fix and have not been ameliorated by the max-points fix), I confess I haven't used NSFS3 for a very, very long time, so perhaps it has become a richer scorinator with better customization options and more sports since my last use of it - I don't know. But I know this much about NSFS3: if there's "no reason not to use it" (a statement with which I'm hesitant to agree), there's also no reason to teach yourself xkoranate given its efficacy in many circumstances (like UICA- or World Cup-style events).

[Osarius]

I always encourage people to learn xkoranate. If you have even the slightest desire to ever host something, you need to be familiar with it.

This is a fair assessment. It's a more versatile tool, no doubt, and typically, it is the scorinator of choice for bigger tournaments.

I'm just saying, though... if you're just starting out with the scorinating of stuff, NSFS3 is far less daunting in certain aspects. Just might be more comfortable to use.

I'm not going to touch the formula debate, lol.

[Ideos]

Was interested in hosting a few sporting events for my region, and decided I'd start scorinating with Xkornate. I was wondering if anyone here could point me in the direction of a guide here on NS or any external site that would give me some understanding on the scoriantor. So far, I haven't found any.

[Osarius]

Was interested in hosting a few sporting events for my region, and decided I'd start scorinating with Xkornate. I was wondering if anyone here could point me in the direction of a guide here on NS or any external site that would give me some understanding on the scoriantor. So far, I haven't found any.

this might be of some help
as a general rule, though... when it comes to scorinators, if you're stuck, just ask. someone will help you out.

[Ideos]

Was interested in hosting a few sporting events for my region, and decided I'd start scorinating with Xkornate. I was wondering if anyone here could point me in the direction of a guide here on NS or any external site that would give me some understanding on the scoriantor. So far, I haven't found any.

this might be of some help
as a general rule, though... when it comes to scorinators, if you're stuck, just ask. someone will help you out.

Alright. I have taken a look at that thread, but I'll take a second look.

[San Jose Guayabal]

Well, I saw recently that there are complains about how SQIS works for association football, mainly based on its randomness and some "crazy" scores are the main arguments coming from some users, well, there's SLIS as an alternative, but after some time, I decided to seek some balance between SQIS randomness and SLIS "Predictiveness" on this, so, I present to all of you the SGIS or Schiavone Guayabalense International Scorination (Testing and tips provided from Saintland).

Here I present you the encoding.

<?xml version="1.0" encoding="UTF-8"?>
<sport version="0.3"> 
<name>Association football—SGIS formula</name> 
<discipline>Association football</discipline> 
<event>SGIS formula</event> 
<scorinator/> 
<!-- only one scorinator is needed for this event -->
 <paradigm>sqis</paradigm> 
<paradigmOptions> 
<double type="constantA">0.11</double> 
<double type="constantB">0.08</double> 
<double type="homeAdvantage">1.22333333333333</double> 
<double type="homeAdvantageGG">0.123333333333</double> 
<int type="attacks">12</int> 
<double type="etAttackCoeff">0.293333333333</double> 
<int type="shootoutLength">5</int> 
<double type="goldenGoalProb">0.4</double> 
<double type="shootoutProb">0.65</double> 
<int type="pointValues">1</int> 
<double type="attackCoeffs">1</double> 
<double type="NSFSStyleCoeffA">2.0000000000000</double> 
<double type="NSFSStyleCoeffB">1.2000000000000</double> 
<double type="NSFSStyleExponent">-0.23705203296846</double> 
<double type="NSFSStyleOffset">0.102435335725325</double> 
<string type="tableHeaderGoalsFor">GF</string> 
<string type="tableHeaderGoalsAgainst">GA</string> 
<string type="tableHeaderGoalDifference">GD</string> 
<list type="tiebreakers"> <string>extratime</string> 
<string>shootout</string> </list> 
<list type="tiebreakerNames"> 
<string>AET</string> 
<string>pen.</string> 
</list> 
</paradigmOptions> 
<!-- the style modifier distribution is a bit wonky because negative modifiers are less likely to take effect -->
 <dataPoints name="style"> 
<dataPoint pos="0">-3</dataPoint> 
<dataPoint pos="0.4">-1</dataPoint> 
<dataPoint pos="0.5">0</dataPoint> 
<dataPoint pos="0.65">0</dataPoint> 
<dataPoint pos="1">3</dataPoint> 
</dataPoints> </sport>

Now, some tests between the three formulas.

All the tests were made with the following settings:

1 v. 0 (max points 1) for 1000 Round Robin.

                          Pld      W     D     L      GF     GA     GD     Pts 
1 Stronger               1000    862   108    30    2531    564  +1967    2694
2 Weaker                 1000     30   108   862     564   2531  −1967     198

SGIS W/Home advantage

SQIS:                     Pld      W     D     L      GF     GA     GD     Pts 
1 Stronger               1000    823   134    43    2416    624  +1792    2603
2 Weaker                 1000     43   134   823     624   2416  −1792     263

W/Home advantage

                         Pld      W     D     L      GF     GA     GD     Pts 
1 Stronger               1000    903    73    24    2811    519  +2292    2782
2 Weaker                 1000     24    73   903     519   2811  −2292     145

SLIS W/Home Advantage

And without home advantage.

                          Pld      W     D     L      GF     GA     GD     Pts 
1 Stronger               1000    822   135    43    2360    594  +1766    2601
2 Weaker                 1000     43   135   822     594   2360  −1766     264
SGIS w/ home advantage off

                          Pld      W     D     L      GF     GA     GD     Pts 
1 Stronger               1000    796   155    49    2105    560  +1545    2543
2 Weaker                 1000     49   155   796     560   2105  −1545     302
SQIS no home advantage

                          Pld      W     D     L      GF     GA     GD     Pts 
1 Stronger               1000    878   102    20    2498    498  +2000    2736
2 Weaker                 1000     20   102   878     498   2498  −2000     162
SLIS (68) no home advantage

Now, with this tests done by Saintland with the formula that I modified, I leave this to the public opinion, any suggestions and criticisms are accepted.

[95X]

With all the recent raves about how SQIS is the best soccer/association football scorinator in NS, thought it'd be worthwhile to mention that while using SQIS to score my current domestic league season, SC Oceanside came up with three consecutive 3-0 wins.
So, perhaps there is—

95X gets pulled into a dark, IC room where a familiar-sounding, cold-toned voice says "there is nothing wrong with those results."
"Yes, Xi Foureleven," was his reply.

—nothing wrong with SQIS! Nothing at all! :)

[New Romili]

I have downloaded the NSFS2 and NSFS3 but can't get either to open. Since this topic is on scorinators I thought I would ask to see if anyone has this problem and how to deal with it.Thinking about having a cup for my region.

[Free Republics]

I have downloaded the NSFS2 and NSFS3 but can't get either to open. Since this topic is on scorinators I thought I would ask to see if anyone has this problem and how to deal with it.Thinking about having a cup for my region.

Do you have Java installed on your computer? NSFS is written in Java, so it requires Java to be installed or it won't start.

[Saintland]

From the WBC discussion thread:

The only really perfect solution would be to create a new scorinator to reduce randomness, much as Saintland/FFR did for (gridiron) football and ice hockey, but it's not as simple as just saying do it. I'm guessing baseball is harder to code because of the linescores, though I personally wouldn't be heartbroken to see those go as long as it showed the number of innings.

I have an idea for a baseball scorinator (currently leaning toward the name Baseinator) and may attempt to create it, if there's sufficient interest. In the case of baseball, I haven't been able to alter the randomness for the existing NSCF formula and it is probably impossible to produce a baseball scorinator by editing xkoranate sport files that would produce linescores or even handle extra innings properly (I have never seen an extra inning result from the xkoranate formula for baseball). Thus, a new scorinator for baseball that addressed the issues with the existing baseball scorinators (and had an option to adjust the randomness level) would have to be programmed from scratch.

The way in which existing scorinators are designed is not really conducive to producing scores for baseball that include linescores. Most, if not all, scorinators are designed around the idea that a team has a limited number of "attacks" over the course of a match with the opportunity to score a certain number of points on each attack. This is ideal for sports that have a clock, but baseball is a sport where a team gets additional opportunities to score whenever they score. With baseball, I think a system of 9 "attacks" per team is ideal, with each team receiving an additional attack whenever they score. Once a team rolls a "no score" rather than a "score," the scorinator could then output the number of successful scores as the linescore for a half-inning. For the bottom of the 9th or subsequent innings (the 9th attack for the home team), the scorinator could be hard-coded to stop if a team scores enough runs to win by 4. If the home team is ahead after the top of 9th, of course, the scorinator would need to recognize this, output an X for the bottom of the 9th, and end the game. If the game is tied after 9 innings, then the scorinator would just run additional half-innings until a team is ahead after a full inning (with the same "home team can't win by more than 4" rule hard-coded in). I believe this design would be more true to the nature of baseball and would solve the question of how to produce linescores.

The main issues after that would be figuring out the correct likelihood of scoring runs and adjusting that for the relative strength of both teams (both numbers could easily be made available as user-editable options). I'd be inclined to add style mods and park factors, but I'm a little leery (given my inexperience with designing new scorination formulas rather than just editing .xml files for xkoranate) that I might end up implementing them in a manner that could alter either the likelihood of victory or the margin of victory and thus imbalance the scorinator. I'd imagine that it would be fairly straightforward to automate the series group stage format and the best-of-series format as well (making them easier for potential hosts who are so inclined).

Any thoughts on the idea of a new baseball scorinator in general or the proposed design in particular?

[Kaboomlandia]

Still being annoyed by unranked Czesc winning Group H and making it to the semifinals with a roster and no RP, I would support Saintland's proposal.

[Gregoryisgodistan]

We definitely need a new scorinator, but if you're incorporating it into xkoronate, I don't think you'd be able to do best-of or series things, since I don't think xkoronate supports them. It would have to be a standalone program. Which is fine, but then it's more complicated.

[Commerce Heights]

I haven't been able to alter the randomness for the existing NSCF formula

Assuming you mean NSFS, the way the formula works is that in each half-inning, the batting team has a probability of (baseScoringProb + baseScoringCoeff × skill difference) to score its first run, and then (extraScoringProb + extraScoringCoeff × skill difference) to score each additional run. The half-inning ends when the team fails to score. Knowing that, I’d think you should be able to adjust it to your liking.

and it is probably impossible to produce a baseball scorinator by editing xkoranate sport files that would produce linescores or even handle extra innings properly (I have never seen an extra inning result from the xkoranate formula for baseball).

You could get a linescore of some sort if you really wanted to, but it wouldn’t handle the ninth inning correctly; only the nsfs-baseball paradigm can do that (to the extent that it does). The xkoranate formula for baseball is intentionally set to run extra innings as a “quietTieBreak” because it doesn’t have a nice way to report just the number of innings without enumerating the result of each extra inning.

With baseball, I think a system of 9 "attacks" per team is ideal, with each team receiving an additional attack whenever they score. Once a team rolls a "no score" rather than a "score," the scorinator could then output the number of successful scores as the linescore for a half-inning.

This sounds familiar…

For the bottom of the 9th or subsequent innings (the 9th attack for the home team), the scorinator could be hard-coded to stop if a team scores enough runs to win by 4.

Oh, you want it to be accurate, do you?

[Saintland]

So, based on CH's post, I've gone in and played around with the variables for the NSFS baseball and I think a .08 or .09 (as opposed to the default of .05) is probably ideal for the coeffs. That cuts the likelihood of a team with a rank+bonus of 0 defeating the team at max points to around 20%, which I think would probably cut the likelihood of such teams making it to the playoff rounds.

I'm surprised that it actually works the way I had proposed and had been under the impression that it was a modified version of the NSFS soccer scorinator, so I'm a little surprised (and embarrassed) that I had inadvertently reinvented the wheel without knowing that I had done so.

Here's the modified scorinator file:

Code: Select all

<?xml version="1.0" encoding="UTF-8"?>
<sport version="0.3">
   <name>Baseball—Adjusted NSFS3 formula</name>
   <discipline>Baseball</discipline>
   <event>Adjusted NSFS3 formula</event>
   <scorinator /> <!-- only one scorinator is needed for this event -->
   
   <paradigm>nsfs-baseball</paradigm>
   <paradigmOptions>
      <string type="resultStyle">periods</string>
      <int type="resultWidth">1</int>
      <double type="baseScoringProb">0.2815</double>
      <double type="baseScoringCoeff">0.09</double>
      <double type="extraScoringProb">0.4095</double>
      <double type="extraScoringCoeff">0.09</double>
      <list type="tiebreakers">
         <string>extratime</string>
      </list>
      <list type="tiebreakerNames">
         <string></string>
      </list>
      <int type="forceTieBreak">1</int>
      <int type="quietTieBreak">1</int>
      <int type="periods">9</int>
      <string type="tableHeaderGoalsFor">RF</string>
      <string type="tableHeaderGoalsAgainst">RA</string>
      <string type="tableHeaderGoalDifference">RD</string>
   </paradigmOptions>
</sport>

And, to illustrate what it looks like over a full season, here's the 1st Acedonian domestic baseball season (Actual Results) re-scorinated with the exact same ranks I used to scorinate that season nearly 2 years ago:

Example Season w/ Adjusted NSFS
                         Pld     W    L     RF    RA    RD 
1 Falcons                154   103   51    803   541  +262
2 Knights                154    94   60    752   634  +118
3 Bees                   154    86   68    731   621  +110
4 Bulls                  154    80   74    690   635   +55
5 Green Sox              154    71   83    603   735  −132
6 Storks                 154    70   84    682   675    +7
7 Tigers                 154    57   97    611   837  −226
8 Raptors                154    55   99    562   756  −194

[Recuecn]

Helloo, everybody,

I'm working on scorinating the next season of my domestic league using xkoronate. My league has 2 tiers, the first of which has 16 teams and the second of which has 20. I made the tiers separate groups in the same event so I could scorinate them at the same time. However, because of the different number of teams, one tier has more match-days than the other. I wanted to scorinate just the first third of the season: 10 games in the top tier and 12 games (rounded) in the bottom tier. If I ask to scorinate 12 match-days, the second tier does exactly that, with each team playing games. However, in the first league, the teams have all played different numbers of games, anywhere from 8 to 11. Only 6 games are played every match-day, whereas I'd prefer that 8 games are played every match-day, with a match-day being skipped entirely here or there.

So is there any way to get around this besides scorinating every tier completely separately?

At this point I'm pretty sure there's not, so this really is misery looking for company more than a real question. Thanks anyways, though.

[Gunsovolk]

Yeah you'll have to scorinator each tier separately.

[Gregoryisgodistan]

I unofficially rescorinated the group stage of WBC 33 using the modified formula Saintland created and the RP Bonuses as of the final day of that stage - I didn't feel like going back and editing it 10 times, one for each day. This will show us if it really would have been less random, and if so, how much less. It does not factor in RP Bonus earned after the conclusion of the group stage, as I did that separately.

Here are the results.

Group A                          Pld    W   L    RF   RA   RD   Pts 
1 Gregoryisgodistan                 10    8   2    55   29  +26    24
2 Darmen                            10    8   2    52   37  +15    24
3 Royal Johnstone                   10    5   5    52   50   +2    15
4 Semarland                         10    4   6    42   52  −10    12
5 Northwinds                        10    3   7    44   58  −14     9
6 Neyosil                           10    2   8    25   44  −19     6

  Group B                          Pld    W   L    RF   RA   RD   Pts 
1 Losthaven                         10    7   3    62   44  +18    21
2 Coreyo                            10    7   3    46   41   +5    21
3 Nova Anglicana                    10    6   4    52   42  +10    18
4 Coconut Isle                      10    5   5    53   61   −8    15
5 South Yerfilag                    10    3   7    43   53  −10     9
6 The Kytler Peninsulae             10    2   8    36   51  −15     6

  Group C                        Pld    W   L    RF   RA   RD   Pts 
1 San Jose Guayabal               10    8   2    58   35  +23    24
2 The Royal Barangay              10    6   4    44   44    0    18
3 Southern Democratic States      10    5   5    48   37  +11    15
4 Free Swiss States               10    5   5    36   41   −5    15
5 Frenline Delpha                 10    4   6    29   37   −8    12
6 Firebolt                        10    2   8    24   45  −21     6

  Group D                          Pld    W   L    RF   RA   RD   Pts 
1 Organized States                  10    7   3    51   45   +6    21
2 San Llera                         10    6   4    34   39   −5    18
3 Milchama                          10    5   5    39   30   +9    15
4 Silkjd                            10    4   6    51   43   +8    12
5 Freelands of Natives              10    4   6    33   40   −7    12
6 Vettrera                          10    4   6    48   59  −11    12

  Group E                          Pld    W   L    RF   RA   RD   Pts 
1 Equestrian States                 10    8   2    60   34  +26    24
2 Abanhfleft                        10    7   3    60   42  +18    21
3 Flardania                         10    5   5    51   52   −1    15
4 The Peoples Emirate               10    4   6    36   50  −14    12
5 Mapletish                         10    3   7    47   61  −14     9
6 Quetzala                          10    3   7    34   49  −15     9

  Group F                          Pld    W   L    RF   RA   RD   Pts 
1 Patistan                          10    7   3    66   45  +21    21
2 Schiltzberg                       10    6   4    55   46   +9    18
3 United States of Devonta          10    6   4    46   44   +2    18
4 Nassau-Hessen                     10    6   4    48   48    0    18
5 Borxinoia                         10    3   7    44   55  −11     9
6 Banija                            10    2   8    41   62  −21     6

  Group G                          Pld    W   L    RF   RA   RD   Pts 
1 Maklohi Vai                       10    6   4    58   42  +16    18
2 St Saratoga                       10    6   4    54   50   +4    18
3 The Fair Republic                 10    5   5    43   45   −2    15
4 Gregory Llama Land                10    5   5    46   49   −3    15
5 The Sword Bloke                   10    4   6    37   42   −5    12
6 Southwest Eastnorth               10    4   6    39   49  −10    12

  Group H                          Pld    W   L    RF   RA   RD   Pts 
1 Cosumar                           10    6   4    42   25  +17    18
2 Czesc                             10    6   4    38   33   +5    18
3 Sacheyenya                        10    6   4    41   42   −1    18
4 Ko-oren                           10    5   5    50   38  +12    15
5 Kaboomlandia                      10    4   6    39   52  −13    12
6 United Russian Soviet States      10    3   7    29   49  −20     9

It's worth noting that in this scenario, only one unranked team advanced to the Round of 16 without posting a roster or RP. A few teams rank-coasted, but if people think that's an issue they can give a higher RP Bonus. That being said, quite a few unranked teams would have advanced to the Round of 16 while only posting a roster. It seems like this formula is good at weeding out teams with a rank+RP Bonus of 0, but when they have very low but positive ranks, it's not as effective.

Let's take a look at a 100-game simulation, min 0, max 1, between Team A (rank 0.8 stupid smiley) and Team B (Rank 0.1).

Pld     W    L     RF    RA    RD    Pts 
1 Team A                 100    66   34    614   351  +263    198
2 Team B                 100    34   66    351   614  −263    102

Team A only wins about 2/3 of the time. Using the existing NSFS3....

Pld     W    L     RF    RA    RD    Pts 
1 Team A                 100    59   41    506   390  +116    177
2 Team B                 100    41   59    390   506  −116    123

So that's not much of a difference, it amounts to about one extra win in 15 games. With most tournaments having a shorter group stage than that, it might very well not make a difference. Now let's take a look at an average team versus someone who posted a roster but nothing else. Give Team A a rank of 0.5.

Modified formula:

Pld     W    L     RF    RA    RD    Pts 
1 Team A                 100    61   39    466   369   +97    183
2 Team B                 100    39   61    369   466   −97    117

Original:

Pld     W    L     RF    RA    RD    Pts 
1 Team A                 100    57   43    499   433   +66    171
2 Team B                 100    43   57    433   499   −66    129

Now it's a difference of about one game in 25.

In other words, this is certainly an improvement, but randomness will still happen in short group stages.

[San Jose Guayabal]

As seeing that there's issues with Xkoranate SQIS formula, there's an alternative to change those values and have something that can be a balance between the NSFS 3.0.1 formula and SQIS formula, I'm open to suggestions and develop a new formula that can be balanced.

Cheers and thanks beforehand!

[Rejistania]

Because of that, I always encourage people to learn xkoranate. If you have even the slightest desire to ever host something, you need to be familiar with it.

Or just use leagion.

[Le Grout]

Could someone quickly explain the difference between Points-Based and Sequential scoring in NSFS?

What are the benefits of using each and in what circumstances?

Thanks

[Doppler]

What can i use to scorinate olympic events? I am in a desperate need for one. Thanks!