What is this Analysis?

This notebook is for examining the current ratings on boardgamegeek.com with the aim of adjusting and computing different versions of these ratings. The main goal is to compute alternative ranks to adjust for a tendency of the BGG ratings to skew towards complex and recently released games. This is not a novel contribution, but this notebook aims to explain some of the considerations behind calculating different ratings for games as well as refresh these ratings frequently off by scraping the most up to data from the BGG API.

1 Examining BGG Ratings

I’ll load in my most recent pull of games from BGG (2023-01-30 15:30:47). For the purpose of this analysis, I’ll look at games published before 2022 that have at least 50 user ratings. There are a couple of different variables I’m interested in exploring: the number of user ratings, the average rating, the geek rating (bayesaverage), and the average weight, or complexity. For the average and geek ratings, votes can range from 1 to 10, while complexity ranges from 1 to 5.

If we plot average rating of games vs their complexity, we can see that a game’s rating is heavily correlated with its complexity.

The average rating is also a function of time, as we see a general rise in the average for more recently released games.

If we size games by their number of user ratings, we can see when we start to really see the number of user ratings for games take off.

1.1 What is the Geek Rating?

Now, so far we’ve just looked at the average rating, which is simply the average of community ratings for a given game. Most people care about the Geek rating, which is combination of the community average and the number of votes. In order to get a high Geek rating, a game need to both be well rated (a high average) and have a high enough number of user ratings. We can get a sense of how games map on both of these dimensions via a visualization I have come to dub ‘the mountain’.

The geek rating is designed to capture a mix of being highly rated and popular. To achieve this, boardgamegeek used a form of Bayesian averaging, (roughly) starting every game off with a set number votes at a lower value which means that it takes a decent number of users rating the game to actually move the geek rating away from this baseline.

We can see this if we plot all games based on their average rating as well as their number ratings and size by the number of user ratings.

BGG hasn’t, to my knowledge, ever made it clear how precisely they compute the geek rating. I’ve seen other people try to reconstruct it, and the general consensus seems to be that they add a few thousand 5.5 dummy votes to every game.

I can figure out the number of dummy votes with some quick optimization, attempting to reconstruct the geek rating for games by toggling the number of dummy ratings.

dummy_ratings dummy_average mae rmse rsq
2050 5.5 0.011 0.020 0.997
2100 5.5 0.011 0.020 0.997
2000 5.5 0.010 0.021 0.997
1950 5.5 0.011 0.022 0.997
1900 5.5 0.012 0.024 0.997
1850 5.5 0.014 0.027 0.997
1800 5.5 0.017 0.030 0.997
1800 5.4 0.074 0.079 0.996
1850 5.4 0.076 0.080 0.996
1900 5.4 0.078 0.082 0.996
1950 5.4 0.080 0.083 0.996
2000 5.4 0.082 0.085 0.996
2050 5.4 0.084 0.087 0.996
2100 5.4 0.086 0.089 0.996

We actually get the closest with roughly around 2000 votes at 5.5. What are our misses when we use this as our formula? I’ve seen some people speculate it’s a function of the standard deviation.

It doesn’t look to be case, but it is interesting looking at the games where the actual geek rating is super different than my estimated version. Alien: USCSS Nostromo, The Fantasy Trip: Legacy Edition and the Binding of Isaac have really big differences, and if we look on BGG they all seem to either have a disproportionate amount of 1s or 10s. I wonder if in the Geek rating they filter out ratings from accounts that are flagged as just spamming ratings?

I’ve also seen people wonder whether they add more ratings to older/newer games. But I haven’t seen much evidence of that? Or at least, not resoundingly so.

At any rate, we should be good to go by using roughly 2000 dummy votes at 5.5.

2 Adjusting for Complexity

The geek ratings on BGG are highly influential, but they skew heavily towards very complex games. This means that “normal people” will have a hard time using the games that are highly rated on BGG.

If we look at the top 100 games on BGG according to the geek ratings, we can see that all generally fall on the heavier side in terms of game complexity.

yearpublished game_id name rank bayesaverage complexity
2017 174430 Gloomhaven 1 8.44 3.89
2018 224517 Brass: Birmingham 2 8.43 3.90
2015 161936 Pandemic Legacy: Season 1 3 8.41 2.83
2021 342942 Ark Nova 4 8.29 3.71
2017 233078 Twilight Imperium: Fourth Edition 5 8.25 4.29
2016 167791 Terraforming Mars 6 8.25 3.26
2020 291457 Gloomhaven: Jaws of the Lion 7 8.23 3.62
2016 187645 Star Wars: Rebellion 8 8.17 3.74
2011 115746 War of the Ring: Second Edition 9 8.17 4.20
2017 162886 Spirit Island 10 8.16 4.05
2017 220308 Gaia Project 11 8.16 4.39
2020 316554 Dune: Imperium 12 8.12 3.01
2015 182028 Through the Ages: A New Story of Civilization 13 8.11 4.42
2005 12333 Twilight Struggle 14 8.09 3.60
2016 193738 Great Western Trail 15 8.09 3.70
2016 169786 Scythe 16 8.03 3.43
2011 84876 The Castles of Burgundy 17 8.01 2.99
2018 167355 Nemesis 18 7.98 3.42
2015 173346 7 Wonders Duel 19 7.97 2.22
2007 28720 Brass: Lancashire 20 7.97 3.86
2013 124361 Concordia 21 7.96 3.00
2016 177736 A Feast for Odin 22 7.95 3.86
2020 246900 Eclipse: Second Dawn for the Galaxy 23 7.94 3.60
2012 120677 Terra Mystica 24 7.94 3.97
2019 266507 Clank!: Legacy – Acquisitions Incorporated 25 7.93 2.74
2019 266192 Wingspan 26 7.93 2.45
2016 205637 Arkham Horror: The Card Game 27 7.92 3.52
2018 237182 Root 28 7.90 3.77
2020 312484 Lost Ruins of Arnak 29 7.88 2.89
2014 164928 Orléans 30 7.88 3.03
2018 248562 Mage Knight: Ultimate Edition 31 7.88 4.64
2018 199792 Everdell 32 7.87 2.80
2015 183394 Viticulture Essential Edition 33 7.87 2.89
2011 96848 Mage Knight Board Game 34 7.87 4.34
2019 251247 Barrage 35 7.84 4.09
2019 271320 The Castles of Burgundy 36 7.83 2.94
2015 175914 Food Chain Magnate 37 7.82 4.21
2021 324856 The Crew: Mission Deep Sea 38 7.81 2.05
2017 192135 Too Many Bones 39 7.81 3.85
2019 285774 Marvel Champions: The Card Game 40 7.81 2.90
2002 3076 Puerto Rico 41 7.81 3.27
2019 256960 Pax Pamir: Second Edition 42 7.80 3.85
2013 102794 Caverna: The Cave Farmers 43 7.80 3.78
2015 170216 Blood Rage 44 7.79 2.88
2018 247763 Underwater Cities 45 7.78 3.61
2007 31260 Agricola 46 7.78 3.64
2017 185343 Anachrony 47 7.77 4.02
2021 341169 Great Western Trail (Second Edition) 48 7.76 3.71
2017 221107 Pandemic Legacy: Season 2 49 7.75 3.26
2016 205059 Mansions of Madness: Second Edition 50 7.75 2.67

We want a list of games that isn’t so heavily skewed towards complexity. We want to “control for” the influence of complexity on the rating. That is, if we take the variation in the average ratings that isn’t explained by complexity, which games still have high ratings?

2.1 Fitting a Simple Regression

To do this, we need the residuals from a regression of average rating on complexity - these will be the variation in game ratings that are not explained by complexity. We will fit the bivariate model on games published through 2021 and inspect the results.

term estimate std.error statistic p.value conf.low conf.high
(Intercept) 5.244 0.015 348.890 0 5.215 5.274
averageweight 0.588 0.007 85.501 0 0.575 0.602

The intercept indicates the average rating of a game with a complexity of 0, which is kind of nonsensical. The coefficient indicates the effect of a unit increase in the complexity of a game on the average rating. Putting these two together tells us that a game with a complexity rating of 1 would have a rating of 5.83.

A game with a complexity rating of 5, meanwhile, would have a rating of 8.19.

r.squared sigma statistic logLik AIC BIC deviance nobs
0.29 0.73 7310.34 -19472.6 38951.21 38974.54 9410.12 17615

The model indicates that the complexity of a game explains about 29% of the variation in average ratings (R-squared, which simply is the correlation coefficient (R) we saw earlier squared). So it’s not the only thing that matters, but it has a pretty sizeable impact on the average rating and the corresponding geek rating.

2.2 Examining the Residuals

We don’t really care about the model per se, we just want to get the residuals.

The point of the residuals is that they are the variation in a game’s average rating that is not explained by complexity, meaning we will see no correlation between complexity and these residuals.

A positive residual in this case is a game that has a higher than expected average given its complexity. But, we can’t just adjust the ratings alone because it skews heavily towards games that have a highly inflated average rating due to a only having a handful of users.

yearpublished game_id name usersrated average bayesaverage .resid averageweight adj_average
2018 228370 TerroriXico 70 9.41 5.50 2.80 2.33 9.24
0 368557 HeroScape: Age of Annihilation 80 9.47 5.54 2.76 2.50 9.20
2018 198005 Revanche 50 8.73 5.54 2.60 1.50 9.04
2012 130357 Go First Dice 50 8.38 5.58 2.55 1.00 8.99
2020 315219 Black It Out 73 8.28 5.54 2.45 1.00 8.89
2018 255249 Monikers: More Monikers 325 8.28 5.92 2.44 1.00 8.88
2019 283152 Monikers: Serious Nonsense 320 8.25 5.91 2.42 1.00 8.86
2019 256885 The Fantasy Trip: Legacy Edition 396 9.31 5.73 2.40 2.83 8.84
0 207203 War Titans: Invaders Must Die! 57 9.23 5.51 2.37 2.75 8.81
2015 203560 Fall of Magic 74 8.18 5.59 2.35 1.00 8.79
2018 263351 Abstratus 79 8.15 5.58 2.32 1.00 8.76
2019 283151 Monikers: Classics 148 8.14 5.70 2.31 1.00 8.75
2012 120269 Red White & Blue Racin': Stock Car Action Game 125 8.64 5.66 2.27 1.90 8.71
2017 221248 Monikers: The Shut Up & Sit Down Nonsense Box 394 8.08 5.96 2.25 1.00 8.69
2019 284111 Drinks & Daggers 76 8.67 5.52 2.25 2.00 8.69
2016 195709 Monikers: Something Something 231 8.07 5.79 2.24 1.00 8.68
2019 259704 Wizard Thieves 50 8.45 5.51 2.22 1.67 8.66
0 210292 Moonstone 61 8.64 5.57 2.22 2.00 8.66
2015 140135 Small World Designer Edition 248 8.96 5.88 2.22 2.55 8.66
2019 249550 MegaMetroCity 62 8.81 5.54 2.19 2.33 8.63
2015 179448 Monikers: Shmonikers 305 8.02 5.86 2.18 1.00 8.62
2017 220975 History Maker Golf 122 8.68 5.66 2.18 2.12 8.62
2012 132229 Rogues to Riches 72 8.39 5.54 2.16 1.67 8.60
2020 245240 Goblin Grapple 199 8.11 5.51 2.16 1.20 8.60
2000 39331 Legends of Boxing 57 8.68 5.56 2.14 2.20 8.58

3 Examining Complexity-Adjusted Games

We will adjust these residuals using an approach similar to their Bayesian averaging methodology, adding 2000 ratings at the average of 5.5

We can put this all together to now look at games that are rated highly after adjusting for the effect of complexity. We’ll look at the top 50 games to keep it simple.

published game_id name geekrank adjustedrank rankdiff geekrating adjustedrating complexity
1876 521 Crokinole 54 1 53 7.72 8.12 1.24
2015 161936 Pandemic Legacy: Season 1 3 2 1 8.41 7.99 2.83
2015 178900 Codenames 116 3 113 7.48 7.96 1.27
2018 254640 Just One 144 4 140 7.41 7.95 1.05
2015 173346 7 Wonders Duel 18 5 13 7.97 7.93 2.22
2017 230802 Azul 66 6 60 7.66 7.87 1.76
2019 284083 The Crew: The Quest for Planet Nine 51 9 42 7.72 7.78 1.98
2018 244521 The Quacks of Quedlinburg 60 8 52 7.69 7.78 1.95
2014 163412 Patchwork 107 7 100 7.51 7.78 1.61
2019 266192 Wingspan 25 10 15 7.93 7.77 2.45
2020 318977 MicroMacro: Crime City 211 11 200 7.28 7.75 1.11
2009 54043 Jaipur 157 12 145 7.38 7.73 1.47
2016 192291 Sushi Go Party! 225 13 212 7.26 7.70 1.30
2016 204583 Kingdomino 250 14 236 7.21 7.70 1.22
2011 92828 Dixit: Odyssey 256 15 241 7.20 7.70 1.18
2009 46213 Telestrations 275 16 259 7.16 7.69 1.07
2018 225694 Decrypto 98 17 81 7.53 7.68 1.79
2014 165722 KLASK 246 18 228 7.22 7.68 1.07
2017 224037 Codenames: Duet 201 19 182 7.29 7.66 1.35
2020 316554 Dune: Imperium 11 21 -10 8.12 7.65 3.01
2012 129622 Love Letter 299 20 279 7.13 7.65 1.19
2008 39856 Dixit 317 22 295 7.12 7.64 1.20
2016 167791 Terraforming Mars 5 23 -18 8.25 7.62 3.26
2019 266507 Clank!: Legacy – Acquisitions Incorporated 24 24 0 7.93 7.61 2.74
2016 201808 Clank!: A Deck-Building Adventure 71 25 46 7.64 7.58 2.22
2014 160069 Ticket to Ride: 10th Anniversary 111 27 84 7.50 7.57 1.90
2019 263918 Cartographers 121 26 95 7.47 7.57 1.88
2012 128882 The Resistance: Avalon 169 28 141 7.36 7.57 1.76
2016 188834 Secret Hitler 214 29 185 7.28 7.56 1.68
1997 172 For Sale 314 30 284 7.12 7.55 1.26
2005 14996 Ticket to Ride: Europe 143 31 112 7.41 7.54 1.92
2014 147020 Star Realms 135 34 101 7.43 7.53 1.93
2018 233867 Welcome To... 145 32 113 7.41 7.53 1.83
2014 156129 Deception: Murder in Hong Kong 219 33 186 7.27 7.53 1.59
2015 156546 Monikers 292 35 257 7.14 7.53 1.09
2011 92415 Skull 404 36 368 7.02 7.53 1.12
2018 199792 Everdell 31 37 -6 7.87 7.52 2.80
2014 148228 Splendor 194 38 156 7.31 7.52 1.79
2017 174430 Gloomhaven 1 40 -39 8.44 7.51 3.89
2016 209010 Mechs vs. Minions 61 39 22 7.69 7.51 2.43
2010 68448 7 Wonders 81 43 38 7.60 7.50 2.32
2017 217372 The Quest for El Dorado 136 41 95 7.43 7.50 1.93
2018 260605 Camel Up (Second Edition) 252 42 210 7.21 7.50 1.48
2011 84876 The Castles of Burgundy 16 44 -28 8.01 7.49 2.99
2016 205059 Mansions of Madness: Second Edition 47 46 1 7.75 7.49 2.67
2019 253344 Cthulhu: Death May Die 85 47 38 7.59 7.49 2.41
2018 244522 That's Pretty Clever! 150 45 105 7.40 7.49 1.88
2016 199478 Flamme Rouge 228 49 179 7.26 7.48 1.68
2016 198773 Codenames: Pictures 381 48 333 7.04 7.48 1.24
2004 9209 Ticket to Ride 204 51 153 7.29 7.47 1.84

This gets us a list of games that is wildly different than before, and it’s a list of games that are much more palatable to “normal people”. In my mind, Crokinole taking the top spot makes a ton of sense and I will die on this hill.

3.1 Which Games Move Down?

Which games have moved up and down the most? We can look at the difference between the BGG Rank and the Complexity adjusted ranks.

We want to focus in on the games that see a positive and negative shift in particular. Let’s look at the games that are penalized the most.

name complexity geekrating adjustedrating rating_diff
Through the Ages: A New Story of Civilization 4.42 8.11 6.83 -1.27
On Mars 4.67 7.75 6.48 -1.27
Gaia Project 4.39 8.16 6.92 -1.24
Lisboa 4.58 7.71 6.48 -1.22
Mage Knight Board Game 4.34 7.87 6.66 -1.21
Food Chain Magnate 4.21 7.82 6.68 -1.15
Through the Ages: A Story of Civilization 4.18 7.64 6.51 -1.13
The Gallerist 4.26 7.69 6.56 -1.12
Twilight Imperium: Third Edition 4.26 7.53 6.40 -1.12
Twilight Imperium: Fourth Edition 4.29 8.25 7.15 -1.11
War of the Ring: Second Edition 4.20 8.17 7.08 -1.09
Trickerion: Legends of Illusion 4.25 7.51 6.45 -1.06
Dominant Species 4.04 7.59 6.53 -1.06
Spirit Island 4.05 8.16 7.10 -1.06
Kanban: Driver's Edition 4.36 7.27 6.21 -1.06
Terra Mystica 3.97 7.94 6.89 -1.05
Barrage 4.09 7.84 6.80 -1.04
Mage Knight: Ultimate Edition 4.64 7.88 6.86 -1.02
Die Macher 4.31 7.08 6.06 -1.01
Anachrony 4.02 7.77 6.78 -1.00
Brass: Birmingham 3.90 8.43 7.44 -0.99
Kanban EV 4.33 7.67 6.69 -0.98
Brass: Lancashire 3.86 7.97 6.98 -0.98
Alchemists 3.92 7.43 6.45 -0.98
A Feast for Odin 3.86 7.95 6.99 -0.96
Antiquity 4.31 7.15 6.20 -0.95
Mombasa 3.90 7.54 6.59 -0.95
Maracaibo 3.90 7.75 6.80 -0.94
Advanced Squad Leader 4.74 7.03 6.09 -0.94
Madeira 4.30 6.94 6.00 -0.94
Caverna: The Cave Farmers 3.78 7.80 6.86 -0.94
Gloomhaven 3.89 8.44 7.51 -0.94
Robinson Crusoe: Adventures on the Cursed Island 3.81 7.61 6.68 -0.93
Caylus 3.80 7.55 6.62 -0.93
Age of Steam 3.89 7.45 6.52 -0.93
1830: Railways & Robber Barons 4.16 7.24 6.31 -0.93
Ora et Labora 3.89 7.35 6.42 -0.93
Le Havre 3.73 7.70 6.77 -0.92
Sid Meier's Civilization: The Board Game 3.90 7.12 6.20 -0.92
Vinhos 4.21 6.91 5.99 -0.92
Great Western Trail 3.70 8.09 7.18 -0.91
Go 3.94 7.31 6.41 -0.91
Fields of Arle 3.87 7.63 6.73 -0.90
Root 3.77 7.90 7.00 -0.90
Feudum 4.60 6.79 5.89 -0.90
Teotihuacan: City of Gods 3.77 7.62 6.73 -0.89
Tzolk'in: The Mayan Calendar 3.67 7.73 6.84 -0.89
Vinhos: Deluxe Edition 4.02 7.48 6.60 -0.88
The Republic of Rome 4.35 6.83 5.95 -0.88
Star Wars: Rebellion 3.74 8.17 7.29 -0.88

This list makes a lot of sense to me - the ratings themselves are still pretty good for some of these games, but these are all very, very heavy games. A lot of Vital Lacerda showing up in here, which to me epitomizes the disconnect between Geek Ratings and ‘ratings for people who fun games’. The former is heavily slanted towards people who relish complexity, whereas the latter is slanted towards games that provide heavy bang for their buck in terms of their weight.

3.2 Which Games Move Up?

The list of games that go up the most is on the other hand very light, party games. Even with the boost from being simple, many of these games still aren’t rated that highly, though some notable ones (Monikers, MicroMacro, KLASK, Just One) end up near the top of the overall list.

name complexity geekrating adjustedrating rating_diff
Exploding Kittens 1.09 5.86 6.53 0.67
UNO 1.11 5.32 5.94 0.62
Cards Against Humanity 1.19 5.63 6.24 0.61
The Game of Life 1.17 4.33 4.93 0.60
Exploding Kittens: NSFW Deck 1.07 5.96 6.56 0.60
Escape the Dark Castle 1.19 6.38 6.97 0.59
Spot it! 1.03 6.41 6.99 0.58
The Mind 1.07 6.68 7.24 0.56
Zombie Dice 1.10 6.08 6.64 0.56
Apples to Apples 1.16 5.66 6.21 0.55
Just One 1.05 7.41 7.95 0.54
Jenga 1.11 5.56 6.09 0.53
Sushi Go! 1.16 6.91 7.44 0.53
Dixit 1.20 7.12 7.64 0.52
Telestrations 1.07 7.16 7.69 0.52
Love Letter 1.19 7.13 7.65 0.52
Virus! 1.05 6.27 6.79 0.52
Skull 1.12 7.02 7.53 0.51
Rhino Hero 1.02 6.67 7.18 0.51
Jungle Speed 1.14 6.29 6.80 0.51
Joking Hazard 1.03 6.03 6.53 0.50
ICECOOL 1.04 6.58 7.08 0.50
Dixit: Odyssey 1.18 7.20 7.70 0.50
Incan Gold 1.11 6.69 7.20 0.50
5-Minute Dungeon 1.11 6.63 7.13 0.50
Yahtzee 1.18 5.40 5.89 0.49
Sorry! 1.15 4.66 5.15 0.49
No Thanks! 1.14 6.95 7.44 0.49
Kingdomino 1.22 7.21 7.70 0.49
Loopin' Louie 1.05 6.52 7.00 0.48
Pass the Pigs 1.06 5.38 5.86 0.48
Qwixx 1.11 6.66 7.14 0.48
6 nimmt! 1.20 6.82 7.30 0.48
Codenames 1.27 7.48 7.96 0.48
Tsuro 1.23 6.54 7.03 0.48
Battleship 1.22 4.74 5.22 0.48
Chutes and Ladders 1.03 3.63 4.11 0.48
Cockroach Poker 1.11 6.56 7.05 0.48
Candy Land 1.10 3.82 4.28 0.47
MicroMacro: Crime City 1.11 7.28 7.75 0.47
Phase 10 1.27 5.13 5.59 0.47
Point Salad 1.16 6.99 7.46 0.47
Taboo 1.21 6.13 6.60 0.47
Can't Stop 1.15 6.72 7.18 0.47
Animal Upon Animal 1.04 6.53 7.00 0.47
KLASK 1.07 7.22 7.68 0.46
L.L.A.M.A. 1.06 6.25 6.71 0.46
Wavelength 1.12 7.01 7.47 0.46
Twister 1.08 4.76 5.22 0.46
Guess Who? 1.11 4.94 5.40 0.46

3.3 Movement in BGG’s Top 100

Let’s restrict to games inside the BGG Top 100 and see how these games are affected.

name geekrating adjustedrating rank_diff
Decrypto 7.53 7.68 81
Azul 7.66 7.87 60
Crokinole 7.72 8.12 53
The Quacks of Quedlinburg 7.69 7.78 52
Clank!: A Deck-Building Adventure 7.64 7.58 46
The Crew: The Quest for Planet Nine 7.72 7.78 42
Cthulhu: Death May Die 7.59 7.49 38
7 Wonders 7.60 7.50 38
Mechs vs. Minions 7.69 7.51 22
Wingspan 7.93 7.77 15
7 Wonders Duel 7.97 7.93 13
Lords of Waterdeep 7.61 7.42 9
Dominion: Intrigue 7.53 7.36 3
The Lord of the Rings: Journeys in Middle-Earth 7.53 7.35 3
Mansions of Madness: Second Edition 7.75 7.49 1
Pandemic Legacy: Season 1 8.41 7.99 1
Clank!: Legacy – Acquisitions Incorporated 7.93 7.61 0
Everdell 7.87 7.52 -6
Dune: Imperium 8.12 7.65 -10
Raiders of the North Sea 7.53 7.29 -15
Terraforming Mars 8.25 7.62 -18
Marvel Champions: The Card Game 7.81 7.45 -24
Lost Ruins of Arnak 7.88 7.46 -26
The Castles of Burgundy 8.01 7.49 -28
Aeon's End 7.64 7.30 -37
The Castles of Burgundy 7.83 7.41 -38
Viticulture Essential Edition 7.87 7.43 -38
Gloomhaven 8.44 7.51 -39
Blood Rage 7.79 7.38 -41
The 7th Continent 7.62 7.24 -45
Concordia 7.96 7.43 -47
Gloomhaven: Jaws of the Lion 8.23 7.46 -50
Architects of the West Kingdom 7.57 7.21 -56
Orléans 7.88 7.35 -63
Brass: Birmingham 8.43 7.44 -64
Five Tribes 7.63 7.21 -70
Nemesis 7.98 7.36 -71
Pandemic Legacy: Season 0 7.72 7.24 -76
Scythe 8.03 7.31 -86
Star Wars: Rebellion 8.17 7.29 -100
Eclipse: Second Dawn for the Galaxy 7.94 7.24 -102
Twilight Struggle 8.09 7.26 -104
Obsession 7.56 7.10 -120
Race for the Galaxy 7.64 7.13 -120
Tainted Grail: The Fall of Avalon 7.59 7.10 -126
Arkham Horror: The Card Game 7.92 7.18 -132
Puerto Rico 7.81 7.16 -134
Pandemic Legacy: Season 2 7.75 7.13 -141
Puerto Rico 7.59 7.06 -143
Great Western Trail 8.09 7.18 -148
Beyond the Sun 7.57 7.04 -150
El Grande 7.56 7.03 -154
Star Wars: Imperial Assault 7.70 7.10 -155
Grand Austria Hotel 7.64 7.04 -170
Power Grid 7.72 7.06 -177
Twilight Imperium: Fourth Edition 8.25 7.15 -177
The Voyages of Marco Polo 7.61 7.00 -180
Spirit Island 8.16 7.10 -203
War of the Ring: Second Edition 8.17 7.08 -212
Too Many Bones 7.81 7.01 -215
Android: Netrunner 7.66 6.97 -221
Battlestar Galactica: The Board Game 7.55 6.93 -223
Eldritch Horror 7.53 6.92 -229
Clans of Caledonia 7.70 6.96 -237
Root 7.90 7.00 -239
Underwater Cities 7.78 6.97 -245
A Feast for Odin 7.95 6.99 -250
Agricola (Revised Edition) 7.63 6.92 -253
Lorenzo il Magnifico 7.53 6.89 -258
Brass: Lancashire 7.97 6.98 -260
Pax Pamir: Second Edition 7.80 6.94 -272
Kingdom Death: Monster 7.70 6.91 -279
Agricola 7.78 6.92 -290
Keyflower 7.56 6.85 -309
Gaia Project 8.16 6.92 -320
Terra Mystica 7.94 6.89 -336
Caverna: The Cave Farmers 7.80 6.86 -342
Paladins of the West Kingdom 7.68 6.84 -345
Tzolk'in: The Mayan Calendar 7.73 6.84 -361
Mage Knight: Ultimate Edition 7.88 6.86 -362
Troyes 7.53 6.77 -392
Maracaibo 7.75 6.80 -398
Through the Ages: A New Story of Civilization 8.11 6.83 -405
Eclipse 7.65 6.78 -410
Barrage 7.84 6.80 -423
Le Havre 7.70 6.77 -426
Anachrony 7.77 6.78 -435
Teotihuacan: City of Gods 7.62 6.73 -463
Fields of Arle 7.63 6.73 -471
Robinson Crusoe: Adventures on the Cursed Island 7.61 6.68 -526
Kanban EV 7.67 6.69 -533
Food Chain Magnate 7.82 6.68 -570
Caylus 7.55 6.62 -604
Mage Knight Board Game 7.87 6.66 -604
Mombasa 7.54 6.59 -635
The Gallerist 7.69 6.56 -714
Dominant Species 7.59 6.53 -744
Through the Ages: A Story of Civilization 7.64 6.51 -813
Lisboa 7.71 6.48 -872
On Mars 7.75 6.48 -883

It’s important to remember that this is meant to be a ranking list for someone who sees complexity as a negative moreso than a positive. Terra Mystica and Gaia Project might be great games for someone like me who is into the hobby, but would it rank highly for someone who isn’t that keen on games? Probably not.

4 Adjusting for “The Hotness”

The list of complexity adjusted games is pretty great for recommending games to a beginner, but it’s not as helpful for people who have been in the hobby longer. If you do like more complex games, what list should you look at? The Geek list as it currently stands is still probably good, but its not without its problems.

4.1 Why is the BGG top 100 full of recently released games?

The Geek Ratings are heavily skewed towards games that have been released in recent years: all of the top 10 games were released after 2015, and only 5 of the top 50 were released prior to 2010. Either we truly seeing the pinnacle of game design in the last 10 years (this is possible), or it means that the BGG users have a tendency to seek out and rate the hotness, which skews the list towards recent games.

game_id name rank bayesaverage
2017
174430 Gloomhaven 1 8.44
233078 Twilight Imperium: Fourth Edition 5 8.25
162886 Spirit Island 10 8.16
220308 Gaia Project 11 8.16
192135 Too Many Bones 39 7.81
185343 Anachrony 47 7.77
221107 Pandemic Legacy: Season 2 49 7.75
2018
224517 Brass: Birmingham 2 8.43
167355 Nemesis 18 7.98
237182 Root 28 7.90
248562 Mage Knight: Ultimate Edition 31 7.88
199792 Everdell 32 7.87
247763 Underwater Cities 45 7.78
2015
161936 Pandemic Legacy: Season 1 3 8.41
182028 Through the Ages: A New Story of Civilization 13 8.11
173346 7 Wonders Duel 19 7.97
183394 Viticulture Essential Edition 33 7.87
175914 Food Chain Magnate 37 7.82
170216 Blood Rage 44 7.79
2021
342942 Ark Nova 4 8.29
324856 The Crew: Mission Deep Sea 38 7.81
341169 Great Western Trail (Second Edition) 48 7.76
2016
167791 Terraforming Mars 6 8.25
187645 Star Wars: Rebellion 8 8.17
193738 Great Western Trail 15 8.09
169786 Scythe 16 8.03
177736 A Feast for Odin 22 7.95
205637 Arkham Horror: The Card Game 27 7.92
205059 Mansions of Madness: Second Edition 50 7.75
2020
291457 Gloomhaven: Jaws of the Lion 7 8.23
316554 Dune: Imperium 12 8.12
246900 Eclipse: Second Dawn for the Galaxy 23 7.94
312484 Lost Ruins of Arnak 29 7.88
2011
115746 War of the Ring: Second Edition 9 8.17
84876 The Castles of Burgundy 17 8.01
96848 Mage Knight Board Game 34 7.87
2005
12333 Twilight Struggle 14 8.09
2007
28720 Brass: Lancashire 20 7.97
31260 Agricola 46 7.78
2013
124361 Concordia 21 7.96
102794 Caverna: The Cave Farmers 43 7.80
2012
120677 Terra Mystica 24 7.94
2019
266507 Clank!: Legacy – Acquisitions Incorporated 25 7.93
266192 Wingspan 26 7.93
251247 Barrage 35 7.84
271320 The Castles of Burgundy 36 7.83
285774 Marvel Champions: The Card Game 40 7.81
256960 Pax Pamir: Second Edition 42 7.80
2014
164928 Orléans 30 7.88
2002
3076 Puerto Rico 41 7.81

Why do newer games so quickly climb the geek list? I would argue that this due to the fact that the geek rating doesn’t use enough votes to start its Bayesian average. 2000 votes or so was probably a good prior for when board games had a much smaller audience, but as BGG and the hobby has grown, newer games rapidly attract enough user ratings that they manage to quickly overtake this prior.

If we plot the average and median user ratings for games published in each year, we can see how more recently published games draw lots of user ratings (though this does start to taper off for games in the last 1-2 years, though we would expect this to go up as these games accumulate user ratings).

This isn’t a problem per se, but it does mean that two games might actually be pretty similar in quality, but the one released more recently draws lots of user ratings, and so it gets a big boost to its rating on the Geek list.

4.2 Is the Hotness a Problem?

I’m not going to argue that this is a fundamental flaw in the rating that the site uses. But I do think it has changed the meaning of the rating over time. Presumably, the idea behind using the Bayesian average is to indicate games that combine both popularity and a high rating. On some level, we’d like the ratings to be that intersection of ‘these are games that have been well received and are very popular’.

If we back to that plot of game averages plotted against logged user ratings, this basically means finding games that are at the top of the ‘mountain’. But I’ve highlighted the games in the top 100 to show that this isn’t quite the case with the geek rating. Popularity ends up mattering less than we might want it to, given the relationship between the two components.

4.3 Amending the Formula

Fortunately for us, we can adjust for this in a pretty simple way by upping the number of baseline 5.5 votes for every single game. Rather than using the 2000 or so that BGG uses, we’ll toggle the number of baseline ratings at various thresholds up to 100,000 votes and see how the rankings start to change. Games that maintain a high ranking at each threshold will be games that have both a large number of user ratings and a high average. Games that are sensitive to the number of baseline user ratings will be ones that are ‘inflated’ in the current ranking system.

What do we find? We can make a table of games with their rankings at different thresholds. There’s no getting around the fact that Gloomhaven is going to be a top game, it has an extremely high average with a lot of votes. The same goes for Terraforming Mars, which has an extremely high number of user ratings with a high average - if we set the number of baseline votes over 25k Terraforming Mars becomes the top game of all time.

There’s no correct number of votes to add, but it is interesting to see how games are affected by having more votes. If we add 100k votes we basically end up with a list that penalizes recent games to enter the top 100 (Gloomhaven Jaws of the Lion, Marvel Champions, Rising Sun, Underwater Cities, Dune Imperium) and provides a boost to some of the pillars of the board game renaissance Pandemic, Ticket to Ride, Pandemic, 7 Wonders, Puerto Rico, Agricola. Man, evidently Terraforming Mars is really good?

4.4 Top Hotness-Adjusted Games

What are the top games if use a prior with 100k votes rater than 1k?

Adding 100k votes gives us the list of games that more or less maps to the evergreen games of the last two decades, and actually shows a decent balance of complex and simple games. I would argue that is is a stronger list to look at for people who are interested in the hobby but aren’t necessarily going to be interested in chasing the hotness.

Going back to ‘the mountain’ visual, what games does this set of rankings end up prioritizing?

Hooray! That’s about what I was hoping to see.