Results vs. Spirits

Results in Scenarios

Scenario Difficulty Completed Completion Rate Average Score Score Model Win Rate Deviation p(χ²) Win Model Differential
Blitz 0 390 54.62% 25.43 -5.664 30.51% -0.277 0.0 -1.111 -0.7173
Dahan Insurrection 4 1056 68.39% 51.92 6.436 45.27% -0.128 0.0 -0.216 0.1331
Guard the Isle's Heart 0 2579 76.69% 51.18 5.551 71.62% 0.128 0.0 0.845 1.1648
Powers Long Forgotten 1 373 58.83% 38.63 0.455 56.03% -0.021 0.4016 -0.215 0.1201
Rituals of Terror 3 226 56.36% 36.87 0.14 33.19% -0.238 0.0 -0.861 -0.1586
Rituals of the Destroying Flame 3 132 39.4% 33.54 0.064 34.09% -0.206 0.0 -0.648 -0.0063
Second Wave 0 788 62.89% 33.31 -1.4 48.98% -0.094 0.0 -0.463 -0.1741
Ward the Shores 2 385 62.81% 30.89 -3.019 23.38% -0.319 0.0 -1.648 -0.6736

Results using Layouts

Results using Blight Cards

Blight Card Completed Average Score Score Model Win Rate Deviation p(χ²) Win Model Differential
A Pall Upon the Land 1826 36.98 -1.613 47.75% -0.04 0.0006 -0.561 1.5954
Aid From Lesser Spirits 1465 37.92 -1.131 48.05% -0.056 0.0 -0.543 1.4676
Back Against the Wall 1490 36.84 -1.594 45.77% -0.062 0.0 -0.615 1.5482
Disintegrating Ecosystem 1775 32.08 -3.995 40.39% -0.095 0.0 -0.862 1.4389
Downward Spiral 2495 35.23 -2.029 49.5% -0.052 0.0 -0.468 1.5822
Erosion of Will 1815 34.54 -2.726 42.04% -0.083 0.0 -0.777 1.511
Memory Fades to Dust 2477 33.72 -2.801 44.53% -0.079 0.0 -0.69 1.5701
Promising Farmlands 1888 30.44 -4.779 33.21% -0.17 0.0 -1.238 1.0575
Tipping Point 1726 33.51 -3.198 41.89% -0.085 0.0 -0.791 1.4763

Results Using Boards

Board Completed Completion Rate Average Score Score Model Win Rate Deviation p(χ²) Win Model Differential
A 9936 67.95% 37.76 -0.707 53.3% -0.052 0.0 -0.032 0.0314
B 9951 67.33% 38.23 -0.339 54.87% -0.036 0.0 0.039 0.035
C 12526 70.84% 42.84 2.375 60.42% 0.017 0.0001 0.43 0.1602
D 10122 67.96% 38.26 -0.335 54.1% -0.045 0.0 0.043 0.0118
East 1515 60.99% 42.54 0.465 43.76% -0.141 0.0 -0.011 -0.0501
Northeast 1420 63.28% 43.4 0.735 47.25% -0.107 0.0 -0.047 -0.0556
Northwest 1256 62.3% 43.08 0.385 49.52% -0.085 0.0 0.09 0.081
West 1486 61.08% 41.09 -0.574 42.26% -0.151 0.0 -0.529 -0.2678

Results Using Options

Option Completed Completion Rate Average Score Score Model Win Rate Deviation p(χ²) Win Model Differential
No Blight Card 100 61.35% 40.36 -3.241 60.0% 0.019 0.7067 -0.504 0.6337
Not Using Events or Tokens 761 60.59% 32.76 -3.04 37.19% -0.213 0.0 -0.776 -0.5604
Skip Initial Invader Actions 544 68.77% 56.91 5.044 79.41% 0.128 0.0 0.549 -0.2052
Start of Game Growth 802 64.63% 57.68 5.688 79.68% 0.153 0.0 0.815 0.5348
Unlimited Undo 2561 65.52% 44.11 1.509 55.29% -0.046 0.0 0.149 -0.2162
Using Tokens but Not Events 2521 67.82% 36.65 -0.823 59.82% 0.002 0.8286 -0.085 -0.1532

Results

Result Completed
Win: Scenario condition 568
Lose: Scenario condition 427
Lose: Adversary condition 9127
Win: Invaders destroyed 13902
Lose: Invader deck ran out 65
Win: Terror victory 2518
Win: Sacrifice victory 72
Lose: Spirit destroyed 380
Lose: Blight pool ran out 5044

Online Multiplayer

Score Model: -0.772 Win Model: -0.059
Multiplayer? Completed Completion Rate Average Score Win Rate
Single-Player 30638 70.34% 38.78 52.97%
Multiplayer 1465 78.01% 37.6 56.66%

Terror Level

Terror Level Completed Average Score Win Rate
I 9245 21.57 5.92%
II 12266 29.31 56.41%
III 8015 32.12 88.01%
IV 2577 35.3 98.56%

Number of Players

Score Model: 0.44 Win Model: -0.016
Num. Completed Completion Rate Average Score Win Rate Dev. p(χ²)
1 21599 74.16% 38.01 51.48% -0.07 0.0
2 6609 67.03% 37.65 54.32% -0.042 0.0
3 2185 62.36% 42.01 60.37% 0.014 0.1922
4 1710 57.97% 47.77 60.35% 0.011 0.3472

Notes

Score Model

Ridge regression coefficient using all independent variables as predictors, and the score as the outcome. This corresponds 1:1 with score, e.g. a coefficient of +1 means a predicted score 1 higher than average. Ridge regression is used instead of standard linear regression because of high multicollinearity.

Deviation

The difference between the actual win rate of this combination of elements and the win rates of each of those elements on their own, averaged. A positive value means that this combination results in a win more often than would be expected.

p(χ²)

This is the significance value of a χ² test comparing the observed data and expected value. In most cases, this expected value is the mean of the win rates of each element. For team position, the expected value is an equal likelihood of winning in all positions. Lower p-values (no greater than 0.01) mean a greater likelihood that the difference is meaningful and not due to chance.

Win Model

Logistic regression coefficient using all independent variables as predictors, and whether the game was won or not as the outcome. Higher values indicate greater likelihood of winning when present in the game. For continuous variables (e.g. number of players), this number is per unit. model.

Differential

The difference between the logistic regression coefficient for this item using this Spirit, as compared to using this item in general. A negative number means that this Spirit underperforms using this item, and a positive number means that it overperforms.

Layouts

Diagrams of the layouts are on the main page. Layouts for completed games are shown as the same as what they were when the game was started, despite powers like Cast Down Into the Briny Deep.

Last updated: Mar 16, 2023, 09:52 AM PDT