• 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!

MultiBLD Regulations.

cubeshepherd

Member
Joined
Feb 18, 2018
Messages
2,258
WCA
2016STEE01
First off sorry for the bump, but I had a quick question regarding MBLD namely: Why do the scrambles cubes have to be in a square of 8 or something akin to that?
I ask because I do not see how it changes anything, but I do not know if I am missing something or not. I am also asking because I am going to be competing this weekend in MBLD for the first time this weekend and I just wanted to see why the regulations are as they are, because when I attempt 10 cubes I like to have 2 groups of 4 cubes and one of 2, but I do not think that, that is allowed. (Or am I wrong on that). Thanks in advance for you help.
 

GenTheSnail

Member
Joined
Mar 18, 2016
Messages
2,249
Location
Illinois, USA
WCA
2016GEEN01
YouTube
Visit Channel
First off sorry for the bump, but I had a quick question regarding MBLD namely: Why do the scrambles cubes have to be in a square of 8 or something akin to that?
I ask because I do not see how it changes anything, but I do not know if I am missing something or not. I am also asking because I am going to be competing this weekend in MBLD for the first time this weekend and I just wanted to see why the regulations are as they are, because when I attempt 10 cubes I like to have 2 groups of 4 cubes and one of 2, but I do not think that, that is allowed. (Or am I wrong on that). Thanks in advance for you help.
So, once you start your attempt, you can re-arrange your cubes in any shape you want. If you feel like it, you can even make a cube tower (though it is unadvised and you may get dirty looks from your delegate).

Those regs are for setting up the cubes before they are uncovered. This was clarified (I believe) because shivam bansal was getting his mbld attempts with his cubes already organized into his preferred shape, while Kamil Przybylski was getting his cubes in a big heap.
The reg change was to make sure that no elite cubers got an advantage over another.

E:Also, maybe the oqoa blind thread would be a slightly better place to ask this? but here is totally fine
 
Last edited:

cubeshepherd

Member
Joined
Feb 18, 2018
Messages
2,258
WCA
2016STEE01
So, once you start your attempt, you can re-arrange your cubes in any shape you want. If you feel like it, you can even make a cube tower (though it is unadvised and you may get dirty looks from your delegate).

Those regs are for setting up the cubes before they are uncovered. This was clarified (I believe) because shivam bansal was getting his mbld attempts with his cubes already organized into his preferred shape, while Kamil Przybylski was getting his cubes in a big heap.
The reg change was to make sure that no elite cubers got an advantage over another.
O' I see. Thank you very much for the clarification on that. That makes me a little happier now that I can have my shapes that I like:). Thank you once again for you help.
 

abunickabhi

Member
Joined
Jan 9, 2014
Messages
6,687
Location
Yo
WCA
2013GHOD01
YouTube
Visit Channel
I recalculated the WCA MultiBLD positions based on the points=2*solved-missed formula. If anyone is interested in the full data set then message me, otherwise here is the R code:

Code:
library(data.table)

# 0. Get the filtered data. ----
results <- fread(
  input = "WCA_export_Results.tsv",
  sep = "\t",
  select = c("competitionId", "eventId", "pos", "best", "personId", "roundTypeId")
)

mbld <- results[eventId == "333mbf" & best > 0]

# Clean up.
rm(results)

# 1. Create a function. ---------------------------------------------------
ufnPointsFunction <- function(solved, missed, seconds, weighting = 1){

  if(solved < missed || (solved == 1 & missed == 2)) {
    return(-1)
  } else {
    new_value = solved * weighting - missed
    new_dd = 99 - new_value
    new_mm = ifelse(missed < 10, paste0("0", missed), as.character(missed))
    new_format = paste0(new_dd, seconds, new_mm)
    return(as.numeric(new_format))
  }
}

# 2. Extract the key info from the "best" column, ignoring old style MBLD. ----
# Make an ID based on the competitionId and round:
mbld$group <- paste(mbld$competitionId, mbld$roundTypeId, sep = "_")

mbld$competitionId <- NULL
mbld$roundTypeId <- NULL

# Use the README in the WCA Export to find out what to do.
# Use substring functions to get the points:
mbld$points <- 99 - as.numeric(substr(mbld$best, 1,2))

# Get the number of missed cubes from the end:
mbld$cubesMissed <- as.numeric(substr(mbld$best, nchar(mbld$best)-1,nchar(mbld$best)))

# Finish off cube numbers.
mbld$cubesAttempted <- mbld$points + mbld$cubesMissed*2
mbld$cubesSolved <- mbld$points + mbld$cubesMissed

# Get the time in seconds.
mbld$seconds <- substr(mbld$best, 3,7)

# 2. Try combinations of our function. ------------------------------------
# Calculate the new points and the new position in the comp with that new score.
mbld$solved_x2 <- ufnPointsFunction(mbld$cubesSolved, mbld$cubesMissed, mbld$seconds, 2)
mbld[,pos_x2:=rank(solved_x2),by="group"]

# 3. Compare who's positions have changed. --------------------------------
output <- data.frame(

  PersonId = mbld$personId,
  Comp_Round = mbld$group,
  Result = sprintf("%s/%s", mbld$cubesSolved, mbld$cubesAttempted),

  Points_x1 = mbld$points,
  Pos_x1 = mbld$pos,

  Value_x2 = mbld$solved_x2,
  Points_x2 = mbld$cubesSolved * 2 - mbld$cubesMissed,
  Pos_x2 = mbld$pos_x2,
  Change_x2 = ifelse(mbld$pos == mbld$pos_x2, "no", "yes")

)

final = output[!(output$Change_x2 == "no"),]

Some interesting findings:
  • Maskow wins WC2017, not Shivam.
  • Not counting the DNFs (as the data wasn't available) the formula change affects the positions of 2324/6332 MultiBLD results in the latest import (19/Oct/2017)
  • Of the 2324 results, 1299 results improved in their positions in that competition/round, 1025 got worse.
Those with these kinds of results were the most likely to see their positions downgraded in competition:

Score Count
1: 2/4 62
2: 3/4 61
3: 3/5 57
4: 3/6 55
5: 4/6 53
6: 4/7 41
7: 4/5 39
8: 5/8 38
9: 6/10 36
10: 1/2 33
11: 5/7 31
12: 4/8 27
13: 5/10 25
14: 6/9 21
15: 5/9 20
16: 7/10 19
17: 5/6 17
18: 7/12 15
19: 6/11 13
20: 8/12 13

These kinds of results were likely to get better positions at comp:

Score Count
1: 1/2 228
2: 2/2 218
3: 2/3 163
4: 3/3 127
5: 3/4 52
6: 4/4 51
7: 5/5 47
8: 2/4 42
9: 4/5 36
10: 6/6 27
11: 6/7 27
12: 5/6 25
13: 3/5 21
14: 7/8 21
15: 7/7 17
16: 9/10 17
17: 4/6 16
18: 5/7 13
19: 8/8 13
20: 10/10 12

By the looks of things, I think results with greater accuracy are still being rewarded with a system based on a higher reward for more solved cubes. Lots of small results are affected, but they can go either way.

The biggest winner is:
Gianfranco Huanqui

Who would have seen his 16/30 jump from 73rd position to 39th at WC2017.

EDIT: Wrong figures. Please report any bugs you see.
Thats one heck of an analysis, a new a partly different way of looking at MBLD rankings, but since it doesn't change rankings that much , except for a few cases like ,Gianfranco Huanqui, the system overall for evaluating points should remain unchanged.
 
Top