• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

Who do you think will get the first official sub 5 3x3 average?

Who do you think will get the 1st Sub 5?

  • Tymon

    Votes: 82 59.4%
  • Max

    Votes: 35 25.4%
  • Ruihang

    Votes: 2 1.4%
  • Felix

    Votes: 8 5.8%
  • Fahmi?

    Votes: 6 4.3%
  • Others

    Votes: 3 2.2%
  • matty

    Votes: 2 1.4%

  • Total voters
    138
Joined
May 28, 2021
Messages
777
Location
Cubing
WCA
2022WEIP01
This thread has 2 purposes.
1. To find who most people think will become the first sub-5 ao5 in comp
2. to find a mathematical method to find out who has the highest chance of doing so.
 
Last edited:

turtwig

Member
Joined
Apr 2, 2015
Messages
656
I feel like 28-30% in the next round is definitely too high for anyone. Tymon and Max have both competed in a decent number of rounds recently so if it were that likely they would've already done it. I think something like this would be hard to compute precisely, since I think there's the factor of people getting nervous in the last 1-2 solves after 3-4 very good solves and knowing they will break the WR and a very significant time barrier (which I think makes it generally harder to break records than their time distribution would otherwise suggest), which is hard to objectively quantify.

That being said, Max probably has a slightly lower global average. The fact that he has more records or accomplishments in general may help his nerves, as well. However, I find that Tymon has more variation in his times. He seems to get a lot more sub-5's. According to the WCA stat page, he indeed has more sub-5's than Max, although that might be because he has competed more in the past year (on the other hand, Max's count includes more sub-5's done before Tymon was at the level he is now). According to Jonatan Klosko's WCA stat page, while Max has a slightly better average of 50 and 100, Tymon has a better average of 25, 12, and of course 5, lending credence to my theory that his times are more varied (which would help him get better small averages, since he can be "luckier" in the short term). For that reason, I find myself leaning towards Tymon. Also, Tymon seems to be competing more than Max, although that could easily change in the coming months.

EDIT: I did some simple simulations
I took all of Max and Tymon's solves from 2021 and 2022. For Max there were 90 (14 rounds) and for Tymon there were 175 (35 rounds). Max's mean and standard deviation were 5.88 and 0.68, approximately, and for Tymon we have 6.06 and 0.93. Indeed, Tymon average slightly slower but has much higher variance. I then generated random times assuming a normal distribution. In a sample of 10,000,000 averages, Tymon "got" 84,778 sub-5 averages while Max got 33,343. In other words, we could say Tymon has a 0.8% chance of getting a sub-5 average in any given round and Max has a 0.3% chance. This seems to support my theory that Tymon has better odds due to his highly varied performance. I find it surprising that Tymon managed to get an average so close to sub-5 in 35 rounds, given the simulation gives it <1% chance. Although, I'd say the "real" probably is probably at most 1-2% for either person given neither actually has a sub-5 average after 49 rounds.
 
Last edited:

qwr

Member
Joined
Jul 24, 2019
Messages
3,371
YouTube
Visit Channel
Your math is misguided. Poisson distributions model events happening over time at a certain rate, such as radioactive decay. Instead, we have comps occurring at known dates, and for the sake of simplicity we can model solve times as independent draws from some solving time distributions. (Classical statistical distributions assume independence which is almost certainly not the case for comp solves as time series data as we see over time solvers are improving (distribution shift), and in the short term you could have correlated good and bad days. But the independence assumption is frequently used anyway and could still give accurate results. There is a lot of theory on time series data but it's probably not worth studying unless you will do statistics actively in your career.)
 
Top