Statistical Analysis of Cockfights

Speculations about how to win cockfights in Sleeping Dogs tend to be unsupported. Before completing this guide, my assumption is that the winner of a cockfight is absolutely random. Let's see how this compares to the data we analyze.

Introduction

I've started with the North Point Cockfight arena. My initial guesses were based on How to win cockfights in sleeping dogs By Prαidαs.

TL;DR

I'm unable to refute the idea that cockfights are pure chance. You're probably better off just using the save scumming strategy to get the gambler achievement. Save when you win, load and run back when you lose. Also, be aware that if your net losses outweigh your net winnings, the gambler achievement won't update until you're back in the green. It isn't broken... it's just a bad design.

My experience with Praidas's method.

At north point, I came in with 60k, left with 60k after 60 guesses using that method.

North Point

Raw Data

Data are sorted by the winner column.

Winner vs. Loser Table (Winner in left column, loser on top row)

Matchups are color coded.
Text is:
Green - Clearly won more
Red - Clearly lost more
Yellow - Matchup is pretty even.

Effect of Wins on Winner

I've tallied up the games where the winner had higher wins, equal wins and less wins than its opponent. The values on top represent the first bird minus the second bird.

Conclusions on this North Point data

R.I. Leghorn is twice as likely to win as lose. I determine this by comparing the winner vs. loser table, where the R.I. Leghorn wins 2/3 of the time against both opponents.

Number of wins is bird breed (color) dependent. As you can see, the Jersey and R.I. Leghorn can have as many as 5 more wins than their opponent, while the Buckeye and Croad never have 3 more wins than the opponent.

Jersey is more likely to win when it has exactly 2 more wins than the opponent. Looking at the Jersey's wins, about half are from when it has 2 more wins than his opponent, but the exact number of wins he has in those instances ranges from 3-6. Of all the times it lost, only 2 were from when it had exactly 2 more wins than the opponent.

A best guess by breed alone based on this data is below. I've demarcated the winner with <bracket-majigs>.

<Croad> vs. Jersey
Croad vs. <R.I. Leghorn>
Buckeye vs. Jersey = Tie
Buckeye vs. <R.I. Leghorn>

North Point Second Run

Raw Data

Winners vs. Losers Table

Thoughts based on the second run in North Point

The winners vs. losers table is troubling. It seems that nothing held true in this second set of data. Jersey and Croad were tied. R.I. Barely lost to Croad, whereas I previously predicted that it would win 2/3 of the time. Buckeye clearly beat Jersey, which should have been a tie. Last, but not least, Buckeye and R.I. were nearly tied as well.

This second run makes me feel that the results are likely to be completely random, or perhaps based on something like how much money you've already earned. Either one is very disappointing.

[Garnyd]

I can gamble with confidence now
thank you

[hemi | Данил]

10/10 work man glad I can find gems like this every now and then

[magazine permafrost]

my cock are the best

[mr.stevens]

But seriously, nice guide, even if RNG screws you over sometimes.

[mr.stevens]

I know tai-chi, mothafucka!

[SquarelyCircle]

ZeloutBurn - I've statistically proven that all cocks are identical. ;) Nice try.

[ORadhy]

my cock are the best

[* Cpl *]

I got best bird