Xenosim being inconsistent across browsers

[CuCN]

Using these decks in the game simulator tool.
Deck 1:

Hover over cards for details, click for permalink

Deck import code : [Select]

744 744 744 744 744 744 744 744 744 744 744 744 74c 74c 74c 74c 74c 74c 74i 74i 74i 74i 74i 74i 7n2 7n2 7n2 7n2 7n2 7n2 8pl

Deck 2:

Hover over cards for details, click for permalink

Deck import code : [Select]

55k 55k 55k 55k 55k 55k 55s 55s 55s 562 562 562 5oi 5oi 5oi 744 744 744 744 744 744 74c 74c 74c 74i 74i 74i 7n2 7n2 7n2 8pr

Deck 1 should clearly do better than Deck 2. However, this is what happens:

These numbers are typical of the values displayed on each browser. The simulator shows different results across all three, which indicates that at least two of them must be unreliable.

[Higurashi]

Before I react at all: did you also run several tests on the same browser to see if the results were anywhere near as variable? This being a combo deck, the quanta won't matter half as much as draw luck, so I'm not surprised yet.

[CuCN]

Before I react at all: did you also run several tests on the same browser to see if the results were anywhere near as variable?

I tried multiple 1000 game runs on each browser, and the winrate never varied more than 5%. (This is what I meant when I said the numbers in the picture were typical.)

[Higurashi]

Aight, here are my results in Chrome. 1 run, 100 000 games. Consistent with your results in Chrome indeed.

Player 1
Wins: 72760
Winrate: 72.76 %
Average TTW: 6.926676745464541
EMs: 1810
EM rate: 2.4876305662451896 %

Player 2
Wins: 27240
Winrate: 27.24 %
Average TTW: 7.518906020558003
EMs: 1779
EM rate: 6.530837004405286 %

[eaglgenes101]

What about opera and safari?

[dawn to dusk]

maybe run singular tests to see how the second deck wins?