What Ruleset/Algorithm is Drag Racing Using?

[MisterBurkes]

I've played Drag Racing extensively on many channels, and I still don't understand how the algorithm/ruleset works. Most recently, I tried betting on the same racer winning 30x in a row without winning, which is statistically unlikely (~1.8% chance assuming each racer has a fair chance at winning). Could you explain the rule-set programming behind this game, it's driving me a bit crazy haha.

Love the bot and the drag racing minigame, by the way

[MisterBurkes]

Happened again, bet on the same racer 28 times without winning.

My current operating theory is that the size of the bet actually reduces the likelihood of the racer winning. This would explain why my results are so skewed, because i steadily increase my bet size after consecutive losses on the same racer.

[Mitch]

Random.

[MisterBurkes]

Random.

It's purely random? So 12.5% chance per racer, no additional modifiers?

Could you perhaps add a page on Drag Racing and Arena to the wiki, in the same likeness as: http://deepbot.deep.sg/wiki/Bank+Heists&structure=Index&page_ref_id=13

[ExpertsOnline]

For drag race, there is a 20% chance that one car will breakdown at the starting line (turns grey)
For the remaining cars, there is an equal chance. Since both these events are mutually exclusive, the odds turn out to still be 12.5%

Funny thing about stats is that the probability assumes an infinite number of events. With a smaller event set, it is entirely possible for your "sample" odds to appear different.

For example, take a dice. Probability states each side has a 1 in 6 chance of coming. So if I pick a number, and throw the dice 12 times, it should appear exactly twice.
If you try that out, you get a very different result set.

I actually setup a spreadsheet to do exactly that. Generate a random number between 1 and 6 twelve times. This was the result.


Now I increase the sample size to 120 tries. This was the result.


With 12000 tries, this is the result.


As you can see, with small sample sizes, the probability 'appears' to look skewed....this is what is happening with your tests.

[MisterBurkes]

For drag race, there is a 20% chance that one car will breakdown at the starting line (turns grey)
For the remaining cars, there is an equal chance. Since both these events are mutually exclusive, the odds turn out to still be 12.5%

Funny thing about stats is that the probability assumes an infinite number of events. With a smaller event set, it is entirely possible for your "sample" odds to appear different.

For example, take a dice. Probability states each side has a 1 in 6 chance of coming. So if I pick a number, and throw the dice 12 times, it should appear exactly twice.
If you try that out, you get a very different result set.

I actually setup a spreadsheet to do exactly that. Generate a random number between 1 and 6 twelve times. This was the result.


Now I increase the sample size to 120 tries. This was the result.


With 12000 tries, this is the result.


As you can see, with small sample sizes, the probability 'appears' to look skewed....this is what is happening with your tests.

Thanks for the explanation! And yep, my sample size is only ~90, so there is definite skew assuming 12.5%.