H2H7 / H2H7 Standings

  • wheres_walto%s's Photo
    Oh they're entirely made up lol I'll be able to explain better in a bit how they all work, but there is definitely a method to the madness!
  • wheres_walto%s's Photo

    Okay, I think Aggregate park scores are pretty well explained, so I'll start with the points formula using the Robber Baron parks as examples:

     

    ((Park Score * 10) - 500) * (%Built / 3)    <- remember, dividing points by 120 = wins

     

     - geewhzz/walto (Raptor) = ((85.62 * 10) - 500) * (50% / 3) = 356.2 * 16.67% = 59.37 points (0.49 wins)

     

    Gee and I get equal shares, and Raptor was worth 0.98 wins total (this is important)

     

     - robbie (World's Fair) = ((79.78 * 10) - 500) * (70% / 3) = 287.8 * 23.33% = 69.49 points (0.58 wins)

     

     - Sephiroth (World's Fair) = 287.8 * (20% / 3) = 287.8 * 6.67% = 19.19 points (0.17 wins)

     

     - Sey (World's Fair) = 287.8 * (10% / 3) = 287.8 * 3.33% = 9.59 points (0.08 wins)

     

    Overall World's Fair is seen to be good for 0.83 wins.  Robbie receives more individual points than gee and I because his increased workload outweighs the difference in Raptor and WF's park score.  This is really all the formula is doing: finding a balance between quality of the park and percentage of the park built and assigning some value (points) that can be converted into wins (done by dividing by 120, it's arbitrary but works).

     

     - BelgianGuy (Raubritter) = ((76.64 * 10) - 500) * (56% / 3) = 266.4 * 18.67% = 49.72 points (0.41 wins)

     

     - walto (Raubritter) = 266.4 * (37% / 3) = 266.4 * 12.33% = 32.85 points (0.27 wins)

     

     - RWE (Raubritter) = 266.4 * (7% / 3) = 266.4 * 2.33% = 6.22 points (0.05 wins)

     

    Raubritter was worth 0.73 wins, lower than both Raptor and World's Fair due to having a lower park score.  

     

     - FK (Archelaus) = ((65.81 * 10) - 500) * (80% / 3) = 158.1 * 26.67% = 42.17 points (0.35 wins)

     

     - Seb (Archelaus) = 158.1 * (15% / 3) = 158.1 * 5% = 7.91 points (0.07 wins)

     

     - BelgianGuy (Archelaus) = 158.1 * (5% / 3) = 158.1 * 1.67% = 2.64 points (0.02 wins)

     

    Archelaus was worth 0.44 wins

     

     - JJayMForce (Morow World) = ((74.94 * 10) - 500) * (95% / 3) = 249.4 * 31.67% = 78.96 points (0.66 wins)

     

     - RWE (Morow World) = 249.4 * (5% / 3) = 249.4 * 1.67% = 4.16 points (0.03 wins)

     

    And Morow World was good for 0.69 wins

     

     

    So with 5 parks, we can see patterns develop; let's look at what park wins indicate:

     

    Raptor = 85.62 park score, #2 out of 30 H2H7 parks, 0.98 wins

    World's Fair = 79.78, #12, 0.83

    Raubritter = 76.64, #14, 0.73

    Archelaus = 65.81, #27, 0.44

    Morow World = 74.94, #17, 0.69

     

    To estimate wins a park is worth, imagine that the top park (assuming 30 total parks made) is given the number 29, the #2 park 28, etc. until the #30 park gets 0 points.  Under those conditions, Raptor would be 28, World's Fair 18, Raubritter 16, Archelaus 3, and Morow World 13.  This is the number of parks each would have beaten this season.  Now, divide that number by 29.  The resulting number is the percentage of parks that each park would presumably have beaten based on park scores.  The top park has a perfect 1.0, while the worst has 0.0.  Take the square root of that number and you have a number very near the number of wins each park was worth.  Let's plug in the numbers:

     

    Square Root of Percentage of Parks Beaten = square root (# of parks ranked lower / total # of parks -1)

     

    Raptor = SQRT (0.97) = 0.98, compared to 0.98 wins using the other method

    World's Fair = SQRT (0.62) = 0.79, 0.83

    Raubritter = SQRT (0.55) = 0.74, 0.73

    Archelaus = SQRT (0.10) = 0.32, 0.44

    Morow World = SQRT (0.45) = 0.67, 0.69

     

    So it's not perfect, but one way of thinking about wins is to think of them as the square root of the percentage of parks the park would be predicted to defeat.

     

     

    Let's move on to skill scores.  The formula for those is:

     

    Wins / Total Percentage Built on Parks

     

    So if we replace wins with "percentage of parks the park would be predicted to defeat", we get this:

     

    % of matchups the park would win / % of the park built

     

    Remember, though, that for individual players, we are using the original wins formula rather than the one derived from the total parks, and that percentage of parks built should be in decimal form.  Let's plug in the Barons again:

     

    wheres_walto = 0.77 wins / 0.87 share (87% of parks built, 50% Raptor, 37% Raubritter) = 0.88 skill

     

    geewhzz = 0.49 / 0.50 = 0.99

     

    robbie92 = 0.58 / 0.70 = 0.83

     

    Sephiroth = 0.17 / 0.20 = 0.83

     

    Sey = 0.08 / 0.10 = 0.83

     

    BelgianGuy = 0.44 / 0.61 = 0.72

     

    RWE = 0.09 / 0.12 = 0.72 (keep in mind these are rounding errors)

     

    FK = 0.35 / 0.80 = 0.44

     

    Seb = 0.07 / 0.15 = 0.44

     

    JJ = 0.66 / 0.96 = 0.69

     

     

    Hopefully that makes sense and it can be seen that a value of 1 or greater (possible because the formula doesn't perfectly match the explanation) is exceptional while a value of 0 would represent building any percentage of a park that wouldn't win any matches (0 / x = 0).

     

     

    Okay, so moving on to the Madden scores.  These are based near equally on skill and experience.  Skill = the average skill scores for each season of H2H participated in; Experience = the total sum of wins across all seasons.  

     

    Madden Score = (SQRT(SQRT(Total Wins * (Skill ^1.5))) * 65) + 12

     

    Skill raised to the 1.5 power is an attempt to massively reward players with skill above 1 (for example, this causes 1.1 to become 1.15) by separating them from the rest of players (a skill of 0.8 ^1.5 becomes 0.72).  Wins is multiplied by Skill, where higher numbers are better for both, and then a double square root (or ^0.25) is taken to more closely regulate the numbers by shifting them closer to the limit of 1.  Multiplying by 65 and adding 12 is entirely for scaling and both numbers are arbitrary.  Let's again plug in the Barons, instead only focusing on wins multiplied by skill ^1.5:

     

    walto = 1.50 total wins * (0.87 average skill ^1.5) = 1.50 * 0.81 = 1.22

     

    geewhzz = 3.75 * 0.90 ^1.5 = 3.75 * 0.85 = 3.20

     

    robbie = 1.74 * 0.82 ^1.5 = 1.74 * 0.74 = 1.29

     

    Sephiroth = 0.17 * 0.83 ^1.5 = 0.17 * 0.75 = 0.12

     

    Sey = 0.57 * 0.86 ^1.5 = 0.57 * 0.79 = 0.45

     

    BelgianGuy = 1.20 * 0.79 ^ 1.5 = 1.20 * 0.71 = 0.85

     

    RWE = 0.09 * 0.72 ^1.5 = 0.09 * 0.61 = 0.05

     

    FK = 1.29 * 0.47 ^1.5 = 1.29 * 0.32 = 0.41

     

    Seb = 0.07 * 0.44 ^1.5 = 0.07 * 0.29 = 0.02

     

    JJ = 0.66 * 0.69 ^1.5 = 0.66 * 0.58 = 0.38

     

     

    Again, it's a formula that attempts to reconcile the number of wins with the quality of parks.  Robbie has more wins than me, but I have a slightly higher skill score, and as a result our Madden ratings are very similar.  By contrast, FK has slightly more wins than BG, but BG has the higher skill score, and thus there is a larger gap in their ratings.

     

    If I were to list the teams by skill instead of their Madden score, they would look like this:

     

    Heaven's Atlas

    Fisch, 0.85 skill

    Sulakke, 0.83

    dr dirt, 0.80

    Steve, 0.78

    Liampie, 0.78

    Tolsimir, 0.69

    FredD, 0.64

    bigshootergill, 0.54

    Coasterbill, 0.54

    Ride6, 0.52

     

    Hurricanes

    Pacificoaster, 1.06

    Xcoaster, 1.01

    RCT2day, 0.95

    thirteen, 0.95 (their only park was Pridelands)

    Kumba, 0.90

    Arjan v 1, 0.89

    nin, 0.83

    Shotguns?, 0.78

    G Force, N/A

    Rofl, N/A (negative scores)

     

    Italian Stallions

    Stoksy, 0.81

    Rene 0.77

    gdb, 0.72

    Jonny93, 0.72

    Loopy, 0.65

    trav, 0.61

    Poke, 0.56

    Dirk Pitt, 0.56

    ][ntamin22, 0.56

    Alex, N/A

     

    Manual Laborers

    djbrcace1234, 1.05 (he's built on Circus Circus, Bermuda, and Jerusalem, never more than 5%)

    Cocoa, 0.88

    Austin55, 0.85

    Louis!, 0.75

    JimmyLaessig, 0.72

    navalin, 0.72

    PizzaWurscht, 0.72

    inthemanual, 0.58

    disneylandian192, 0.49

    Lightkeeper, 0.42

     

    Robber Barons

    geewhzz, 0.90

    wheres_walto, 0.87

    Sey, 0.86

    Sephiroth, 0.83

    robbie92, 0.82

    BelgianGuy, 0.79

    RWE, 0.72

    JJayMForce, 0.69

    FK+Coastermind, 0.47

    Seb, 0.44

     

    The Rat Pack

    AvanineCommuter, 0.93

    Milo, 0.64

    ottersalad, 0.63

    turbin3, 0.61

    Roomie, 0.60

    Maverix, 0.58

    csw, 0.58

    In:Cities, 0.54

    5dave, 0.54

    SSSammy, 0.51

     

     

     

    And just for fun, the last 2 H2H champion rosters looked like this:

     

    H2H6 Heaven's Kitchen

    wicksteed, 1.07 skill, 66.90 Madden score

    Pacificoaster, 1.06, 92.87

    Turtle, 0.96, 89.69

    AvanineCommuter, 0.93, 71.42

    Sulakke, 0.83, 69.74

    Steve, 0.78, 77.88

    Liampie, 0.78, 94.46

    chapelz, 0.72, 71.60

    camcorder22, 0.71, 58.99

    FK+Coastermind, 0.47, 64.08

     

    H2H5 Hurricanes

    Kumba, 0.90, 98.66

    Fisch, 0.85, 83.01

    J K, 0.83, 89.99

    nin, 0.83, 84.58

    robbie92, 0.82, 81.29

    dr dirt, 0.80, 71.21

    JDP, 0.72, 62.58

    Gwazi, 0.67, 62.66

    Casimir, 0.63, 55.84

    disneylhand, 0.55, 55.88

  • djbrcace1234%s's Photo

    I'm an anomaly! Take that, bush haters!

  • Austin55%s's Photo

    the 4 way final will make this a bit tougher to decide

Tags

  • No Tags

Members Reading