H2H9 / H2H9 in numbers, or Liam's Excel Wisdom

Last H2H I started a series of write-ups incorporating data from my magical Excel sheet, I was going to spill all the beans. I ended up writing only two episodes, which was just me warming up. One of the big things I wanted to share was a ranking of most valuable H2H players of all time. Rather than slowly building up to something like that and getting swept up in the H2H madness before I can get to it, let me start with this right away, and get it out of the way.
 
Episodes
1. H2H8 Team Continuity (2018)
2. H2H Veterans (2018)
3. H2H Power Rankings (2021)
 
-------------------------------------------------------------
 

3. H2H Power Rankings

 

One of the worksheets in my Excel file is a list of everyone who has ever particiapted in H2H in the rows, with data from each H2H season in the columns. Data for each H2H season includes the draft round the player was picked in, the team the player was on in the season, and two columns per park the player has built on; one for the creator share on the park, and the vote percentage from the match. This is all the data I used to determine my Power Rankings. Some assumptions and decisions necessary to make this work:

 

1. Each H2H has the same significance 2. H2H1 is ignored due to incomplete data 3. Each match has the same significance 4. Parks that win due to the opponent forfeiting or getting DQ'ed are ignored 4. Players that win matches are valuable 5. Players with an x% share on a park are x% responsible for that park's score 6. Quality of the opponent's park is not taken into account 7. The more H2H parks you've built, the more significant your record

 

 

For example, John has a 100% share on Venetia Harbour from H2H3, which received 78.75% of the votes in its match against Kierikkengord Castle, and a 50% share on Rapa Nui from H2H4, which received 48.98% of the votes in its match against Isla Magica. John can claim half of the score from the loss, and the full score from the win.

So, having acquired all this data, to produce a ranking I have to bring it all back to a single number. Simply adding up scores is not good enough. Three losses may be more than a single win, but the goal is to determine which players are losers, and which players are winners. So the first step is to take the score from each match, and detract 50%. Losing parks now have a negative score, and winning parks have a positive score. The next step is simple: multiplying the scores with the creator shares, and adding them up.

Lastly, a weight is applied using an arbitrarily determined multiplier that results in a curve that I liked. For each park you've built, the multiplier increases, but with each extra park the increase gets smaller. If you've built on one park, the multiplier is 0.5. If you've built on three parks, the multiplier is about 1.0. For ten parks, the multiplier is about 1.4. The highest multiplier at the moment is 1.426 to be precise, for Kumba, who has the record for the most H2H parks built (13), with Liampie and Roomie trailing one park behind.
 

John example: Venetia Harbour won with 78.75% of the votes. Detracting 50% results in a 28.75% win margin. Rapa Nui has a 1.02% loss margin, or a -1.02 win margin. Venetia Harbour: 28.75 * 100 = 2875 Rapa Nui: -1.02 * 50 = -51   Sum: 2875 + -51 = 2824   John's 2824 score gets multiplied by 0.876 for a final total of 2472.

Now that we have the final numbers, we can get to the actual ranking. I will start with the top 20, and end with the bottom 5, with some additional notable players found inbetween the two extremes. The full list will be disclosed ini the next update, sometime between tomorrow and 2025.

 

#1 Kumba
Score: First H2H: Last H2H: Parks built: Matches won:	16363 H2H3 H2H9 13 69%	This shouldn't be too much of a surprise; a veteran of H2H, and two times captain/champion. The reason for Kumba's high score is having only two real losses (Rowling vs Tolkien; 6%, and Zombieland; 20%), two narrow losses (which barely count), and nine (!) comfortable wins, including two 100% wins. Notable wins are Roman Vice (62%), Star Wars (85%), Aviara Cove (77%), and Terra Progressia (100%).
#2 geewhzz
Score: First H2H: Last H2H: Parks built: Matches won:	13480 H2H4 H2H7 7 86%	Another obvious top tier player. Geewhzz has never lost a game, although Kayte Ridge versus Internet City from H2H4 remains one of the closest and most controversial wins in the game. All his other parks won by a comfortable margin. Notable wins are Lenox Mall (100%), Raptor (88%) and Belmont Shores (84%). Arguably his weakest park, Harkview Heights, won by forfeit and thereby doesn't count for the Power Ranking.
#3 Liampie
Score: First H2H: Last H2H: Parks built: Matches won:	10448 H2H5 H2H9 12 75%	Another two time captain/champion. His seasons typically consist of two high calibre parks and a throwaway bonus park, resulting in a few losses. His wins include one hard-fought narrow win with Billy Wonka's (52%), one landslide win in #diamondheights (100%), and mostly comfortable wins in the 70s like Tenochtitlan (74%) and Corsair Veredian (76%). Metropolis (26%) and Aquatica (16%) are two of his three losses, with The Atlantean Ark (6%) being the biggest misfire.
#4 Pacificoaster
Score: First H2H: Last H2H: Parks built: Matches won:	8795 H2H6 H2H9 6 83%	Liam's narrow Billy Wonka win is tied to Pacificoaster's only loss, Ruigrijk (42%). In the same season, Castles-n-Coasters got a narrow win (51%), and then there are some comfortable wins, most notably Paradise Pier (72%) and Disney's American Waterfront (94%). Paradise Pier is an interesting one, being part of a four-way final. Despite not winning, it beat two of the three parks head to head and thus is a net winner.
#5 Cocoa
Score: First H2H: Last H2H: Parks built: Matches won:	8602 H2H6 H2H9 6 67%	Things didn't start out well for Cocoa, with a 32% loss in H2H6, with Worlds of Fun. DisneyPunk sits at 49.99%, and then there are four comfortable wins. The Wild West got a 100% win, and Bermuda: The Lost Colony, Neverland, and of course Forgotten Mekong are in the 70s. A solid record. Strong wins, no deep losses.
#6 Coaster Ed
Score: First H2H: Last H2H: Parks built: Matches won:	8262 H2H2 H2H4 7 71%	The only real old school player to rank so high. Coaster Ed had three epic seasons, with even his two losses (Cajamarca and Kitabasaki) being famous. Notable wins include Battlefield RCT (88%), Mars Colony Eurolai (70%), BLOOD ISLAND Adventure (66%), and Erwindale Forest (66%).
#7 BelgianGuy
Score: First H2H: Last H2H: Parks built: Matches won:	7815 H2H6 H2H9 7 71%	BelgianGuy has accumulated an impresive record in his three seasons. H2H6 resulted in two landslide wins for him, with Frankenstein (93%) and Atlantis Resort (93%), and H2H8 had BelgianGuy feature in Zerzura and Dig Site 4.
#8 dr dirt
Score: First H2H: Last H2H: Parks built: Matches won:	7212 H2H5 H2H9 7 57%	Dr dirt entered the H2H scene in the fifth season, becoming one of the most succesful replacements ever with a 71% win with Renaview Bay Park and a marvelous performance with Worlds of Tim Burton, which sadly doesn't count due to having no opposition. While Calico Canyon and Studio Ghibli may not have won, Galaxy Geeland (100%), Dig Site 4 (94%) and Zerzura (84%) certainly did.
#9 Steve
Score: First H2H: Last H2H: Parks built: Matches won:	6456 H2H3 H2H9 10 60%	I don't know how this happened. Three losses!? Steve contributed to Metropolis, Kidderbrook and Malaria losing, although admittedly the latter two are shallow losses. #diamondheights (100%) is Steve's biggest win, Tierra Aventura (88%) also carries a lot of weight, and some small shares on winning parks like ColorFlood and Corsair Veredian add up to an impressive score.
#10 Tolsimir
Score: First H2H: Last H2H: Parks built: Matches won:	5868 H2H5 H2H9 8 75%	Two losses (Pixar Animation Studios, 29% and A Year in Winkelheim, 36%) can't cancel out a sextet of strong wins, like Zerzura (84%), Asteroid Fields (78%), Mobray's Illusions (75%) and Tenochtitlan (74%). With his recent endeavours having proppeled him into an elite tier of players, Tolsimir is just a park or two away from an even higher position.

 

Missing out on the top 10, but still among the most valuable H2H players ever:

 

#11 Fisch	Was on the H2H5 winning team, had a not so good H2H6 aside from Sea of Sagas, a fruitful H2H7, and another excellent H2H8.
#12 Magnus	Left the H2H scene with Park Edda (43%), but had otherwise never lost a match before, carrying his teams in times of need.
#13 Twisted	Never lost a match, has two great wins in Kilika (90%) and Xlapak (84%).
#14 Turtle	Had a H2H career filled with bad luck, but turned it around in H2H6 having one of the best individual seasons ever.
#15 J K	Terra Progressia, Tierra Aventura and Anheuser Busch Park & Zoo are enough to compensate for his only loss; The Amazing Adventures of Spiderman (31%).
#16 Raven-SDI	Two wins and no losses: Rapper's Delite (60%) and Clock Tower (78%). Nothing too memorable, but a win is a win.
#17 K0NG	Two wins and no losses, but unlike Raven-SDI, K0NG has to share his wins. Very memorable: Belmont Shores (84%) and Frankenstein (93%).
#18 wheres_walto	Fantasy Park with Strong Narrative hurts a bit, but Avatar (97%), Cedar Point's Raptor (88%) and Erlebnispark Raubritter (64%) are more than enough to make up for it.
#19 Micool	What was Assault on Earth Base Gamma Gamma!? Anyway,  three wins. Who cares about Assault on Earth Base Gamma Gamma?
#20 Emergo	Only one H2H season, and another of the best individual H2H seasons ever. Alice in Wonderland narrowly lost, but Azuris City (77%) and Mont Saint Michel (83%) comfortably won.

 

A small selection of players outside the top 20:

 

#21 Ablaze	Notable because he won only one of his three matches. Still a high ranking, due to the losses (Atomkraftwerk and Elfwood) only narrowly lost, and Livewires (78%) was quite convincing.
#24 AvanineCommuter	Mictlan lost (28%), Mandala, Arcanis Mineralis and of course Le Rêve Parapluie won.
#32 Loopy	One of the most productive H2H players (8 parks), with an even split between losses and wins. For every Alumwell Bay there is a Xlapak, and for every Hanson there is a Kilika.
#34 Sulakke	Sucked during H2H5, but in his three seasons since he's produced one winning park per season, like Carreira da India (72%), Lijiang (68%) and Forum Caeleste (65%). 61% win in H2H9?
#38 saxman1089	Holds the record for most parks in a single season, and interestingly his team won the entire season. 3 wins, 3 losses, but the wins outweigh the losses.
#46 robbie92	Such a promising start in H2H5, with World's Fair and Worlds of Tim Burton, and then a dip with two disillusioning losses in H2H6 and H2H7. Thankfully, robbie92 stuck around and avenged himself in H2H8 with Tubiao Action Park (73%).

 

And now onto the very bottom of the list... You'd expect to see some obscure names here; players at the fringe of the community who happened to get drafted, built on a losing park, and disappeared into insignificance. However, remember that the more parks have been built by a player, the heavier the weight. Therefore, the bottom of the list are not the insignificant losers, but the significant losers. Context is kinder to them than the numbers, and today we're talking about the numbers.

 

#216 nin	Some bad luck? Four losses, and only two not too crushing victories. Pridelands stands out with 59%, as well as Roman Vice with 62%, but nin only has a small share on that park.
#217 Roomie	Mr. Sawyer himself! One of the most productive H2H players (12 parks), and all of them sucked! Nah, not actually. But six heavy losses, three narrow wins and only two ancient convincing wins, and that's enough to get down here.
#222 Six Frags	Another productive player: 10 parks, and all of them sucked! Nah, not actually. The rushed nature of his H2H5 parks do him no favours though. Rowling vs. Tolkien and Heaven's End make it clear that Six Frags and high fantasy are not a great match, statistically speaking. Meanwhile, Six Frags excels with parks like Forum Caeleste and Mont Saint Michel, having won matches convincingly. For every loss, there is a win.

 

 

#224 FK+Coastermind
Score: First H2H: Last H2H: Parks built: Matches won:	-5026 H2H5 H2H9 6 50%	FK's record is not all shit, but his wins (Monstrocity, 62%; ColorFlood, 58%; Park Guell, 55%) don't weigh quite as much as his losses. Fantasy Park with Strong Narrative destroys FK's rating, and Archelaus and Delta Desert Bays do him no favours either. But like Tolsimir, this is a player that has undergone quite a transformation in recent times. I'd be surprised to find FK in the bottom half of the list a few months from now.
#225 Cedarpoint6
Score: First H2H: Last H2H: Parks built: Matches won:	-5279 H2H3 H2H9 5 20%	One legendary win (Kayte Ridge, 51!) is nice, but the win margin is super tight, and it's definitely not enough to cancel out deep losses like Feira do Flamengo (27%), Mysterious Forets (27%), and Snowdrift (12%). Allegheny Adventures from H2H8 could've resulted in his first comfortable win, but sadly the opponents forfeited, and thus nothing was gained!
#226 trav
Score: First H2H: Last H2H: Parks built: Matches won:	-5993 H2H4 H2H9 10 17%	Sea of Sagas (80%) saves trav from being all the way at the bottom of this ranking. All his other parks have performed horribly. Notable losses are Cyleal (13%), A Spring Fair in Clovelly (13%), and Worlds of Fun (32%). A H2H career of 15 years filled with either bad decisions or bad luck. Can he turn the tide in H2H9?
#227 Drew
Score: First H2H: Last H2H: Parks built: Matches won:	-7205 H2H2 H2H3 6 0%	Drew got close to a win, once, and ironically this became the only draw in H2H history; Sierra Glen. His other parks are mostly forgotten. La Barranca Roja Grande is the only one I can picture right away, but Rushville, Adventura Espana or Mishashta Heights? I need to google those.
#228 Ride6
Score: First H2H: Last H2H: Parks built: Matches won:	-7589 H2H3 H2H7 10 30%	It's tragic that such a respectable builder has to sit at the bottom of this list. Ride6 has a share on the Pyramid Connection (narrow 51%) win, has a high profile solo win with Eternal Springs City (52%), and ironically his biggest win only has a small share of his work: Seven Wonders of the Ancient World (76%). Parks like Siege at Castle Grijs, The Lost Samurai, Calico Canyon, Pueblo Canyon Boardwalk and Below Bermuda are all reasonably memorable parks, but none of them won. Alumwell Bay got exactly 0 votes, further dragging Ride6 down. Certainly some bad luck involved here, excelling at the wrong times.

These rankings are certainly strange. I think you may need to sightly reevaluate your formula when I'm the fifth best player ever and cp6 is the fourth worst. I guess i get what you're going for but its clearly missing some important factors. I actually wonder if a narrow loss should actually be valuable in some way. Or rather, opponent quality needs to factor in.

Agreed. We both know CP6, Ride6 or FK for example are great players, but their H2H records are not very good. If these were ELO rankings, they'd also be hurt.

 

That said, looking forward to do a V2 at some point. But one issue is having no data that's directly indicative of a park's quality, I only have its quality relative to the opponent.

But one issue is having no data that's directly indicative of a park's quality, I only have its quality relative to the opponent.

 

 

Why not use the park's accolade score? Even if the park loses, a player's ability to build something high quality should factor into the score somehow. There are tons of losing H2H parks out there that would have beaten other parks in the same round but against a different park.

 

Additionally, you could calculate the average accolade score of each round and factor in a multiplier based on how much higher or lower a winning or losing park is from the average score. Lost World would have likely destroyed every other R1 park, but it happened to go against the one entry that could edge it out just slightly in both votes and accolade score. 

When I sign up for H2H10 after a 2.5 year hiatus I can say I'm the 18th best head to head player ever

Thank you Liam for putting this together.

 

I agree the ratings will never be perfectly fair, nevertheless this is what H2H is about. Without the drama of high-quality matchups, forfeits and low quality parks winning matchups the whole contest would lose its character. 

While the ratings applied by Liam do not take park quality into account, they take into account what matters in H2H: votes and wins.

 

On a side note: Surprised seeing myself #12.

My record sucks, surprised I'm not near the bottom.

I'm curious how my ratings would work. Haha. I've never lost in a round robin park but my percentage of building is... well. lol This is cool, Liampie. Thanks for sharing.