Mini-Smash Poll Results

Mini Smash Poll results


Recently, I asked the Smash speculation community what would they rather have: two newcomers and two veteran fighters, or four completely new fighters. After a couple of days, I decided to close the poll and compile the results. After deleting the joke votes, and the votes that did not follow directions (voted for the same newcomer multiple times), I’m left with 705 responses (40 responses were deleted).

Here’s what I found:

People would rather have two newcomers and two veteran fighters.

image (13)

Veterans request rate (Only for those who picked “A”):

image (14)

Newcomers rate (For those who picked either “A” or “B”)



image (18)


Only A: Only B: Overall:
King K. Rool 34% 53% 42%
Isaac 24% 31% 27%
Inkling 11% 19% 14%
Shovel Knight 10% 16% 14%
Dixie Kong 9% 18% 13%
Shantae 5% 21% 12%
Krystal 7% 14% 10%
Wonder Red 6% 15% 9%
Rayman 6% 13% 9%
Bandanna Dee 6% 13% 9%
Banjo and Kazooie 4% 13% 8%
Paper Mario 5% 8% 7%
Ridley 4% 14% 7%

Characters were more likely to be requested when more choices were available. This is intuitive, and was expected. King K. Rool (+19%), Dixie Kong(+9%), Shantae (+16%), Wonder Red (+9%) Banjo and Kazooie (+9%), Ridley (+10%) all saw significant increases when more choices became available. I ordered the above chart in total % desired, so it’s easy to see that for some characters, if the choices were limited their actual demand decreases. If the newcomers were ranked purely on a hypothetical scenario of “one or another” then the order becomes mixed up.

Effects of Sharing with Support Threads:

I asked others to share the poll, and I quickly realized that the poll was being shared with support groups. I decided to see how strong each other support groups were by comparing the pre-share request rate to the post share request rates. Essentially, how much of an impact does each of the support threads have on a particular character? Of course, this is common sense. If you ask someone who is on SmashBoards if they like SmashBoards, the clear majority will say “yes” (with the exception of a few rebels!). However, this is something that the community needs to understand is an issue with polling, and something I mentioned in my “Dangers of Online Polling” article. With this poll, I was able to directly record bias and the ‘echo-chamber effect’ of the community. The following is ordered by the timing of when the poll was shared with the groups. If I’m missing a share, please include a comment and I will update this post.

Feel free to scroll down to see the data in chart format.

The poll was shared on NeoGaf early on. NeoGaf is often thought of as extremely pro-Wonder Red in the speculation community. Pre-share, Wonder Red held as 12% request rate overall. Post share, Wonder Red had a 9% request rate. This is probably due to NeoGaf’s forum structure–mainly a lack of “character support threads”. Also, the poll was shared extremely early on, meaning it had a lot of time to compete — and other fans to compete with.  

Link to when the poll was shared with the King K. Rool support thread on SmashBoards. Between the time the poll was opened, and shared on the King K. Rool support thread, 359 votes were recorded. Out of all of these votes, K. Rool held an overall 31% request rate. For respondents who picked “B” from the start to when the King K. Rool thread was shared, K. Rool had a 38% request rate. For respondents who picked “A”, for the same time period, King K. Rool held a 27% request rate. Therefore, he was slightly more requested among respondents who wanted four newcomers instead of two veterans (by 11%). This is not surprising as more choices = more chances for a character to appear.

From the time the poll was shared with the King K. Rool support thread to the end of the poll, K. Rool held a 50% request rating. Essentially, the K. Rool support thread was able to bump up K. Rool’s support by 9%.

Here’s when the poll was shared with Bandana Dee support thread on SmashBoards. Before the poll was shared with the Bandana Dee supporters, Bandana Dee held a 6% request rate. After it was shared, Bandana Dee’s support jumped up to 10%. Therefore, just like King K. Rool, sharing the thread in a support thread had a noticeable impact. Bandana’s Dee support thread was able to jump up by 4%.

The poll was also shared with the Rayman support thread on SmashBoards. Before the share, Rayman had a 8% request rate. After the share, his support was 13%. Therefore, sharing with the support thread has had a noticeable impact. Overall, his support increased by 1% by the share.   

The poll was also shared with the Krystal support thread on SmashBoards. Before the share, Krystal was requested at a 7% rate. After the poll was shared, Krystals support jumped up to an amazing 66%! The poll would be closed shortly after it was shared with the Krystal support thread (Only 36 votes were recorded). However, Krystal was able to climb 3% by the share.  

The poll was also shared with the Dixie Kong support thread on SmashBoards. Before the share, Dixie was being requested at a 13% rate. After the share, Dixie support rate increased to 15%. Just like Krystal, only a number of votes were recorded after the share (45 votes), but the last minute bump was able only able to marginally increase her chances.

Lastly, Shovel Knight’s position might be affected by the rumor circulating around. Comparing the data to previous polling efforts might prove this.

Here’s the data in graph/ table format.
image (16)

Name: Pre-Share % Post-Share % Change (Pre and Overall): Overall
King K. Rool 31% 50% 11% 42%
Dixie Kong 13% 15% 0% 13%
Krystal 7% 66% 3% 10%
Wonder Red 12% 9% -3% 9%
Rayman 8% 13% 1% 9%
Bandanna Dee 6% 10% 3% 9%

The poll was also shared with the Snake support thread on SmashBoards. Overall, Snake held a 48% request rate. For those respondents who only picked “A”, Snake had a 78% request rate. Snake’s support decreased as time went on (-9%), which was either affected by me egging on Wolf and Snake supports against each other, the Wolf fanbase being overall stronger, or the possibility that the poll being shared with the Snake support thread had no impact on his performance.

The poll was also shared with the Wolf support thread on SmashBoards. Overall, Wolf had a 43% request rate. For respondents who only picked “A” Wolf had a  76% request rate. Before the share, Wolf had a 75% request rate with respondents who picked “A”. After the share, the support jumped up by 5% to 80%. Therefore, a noticeable jump was observed. Comparing the Snake numbers post-share for Wolf (66%) and pre-share for Wolf (42%) reveals that Wolf had little to no impact on Snake’s position in the ballot.

Anyhow, that’s all for today. I’m still working on collecting Japanese results. Please be patient, as I’ve been extremely busy with other projects. I’ll get back into it very soon.  

PushDustIn deals with data everyday. Someday he’ll finish his transition to full robot, and take over the Smash community. Until then, you can follow him on Twitter.

Share this!


  1. Characters were more likely to be requested when more choices were available.

    Well, to be fair, it may reflect more than that. It may just mean “People who prefer having only newcomers are more likely to vote for X”, “People who support X are likely to be supporting another vet as well”, etc. Also, the “more likely to be requested”, might just be because “you are given more votes for them”, as in, well, you get only TWO votes for A and FOUR votes for B. Then the popularity might just be amplified through those characters who are generally more popular and is just placed as a filler. Halving the result (assuming you didn’t already) may reflect a bit on this:

    There are those general top choices (K Rool, Isaac) who *might* just be filler choices when one only wants to vote for their veterans to be back and nothing else, but we won’t know how much there will be.

    Then there are those whose rate in B are more than twice of that in A before halving: Shantae, Wonder Red, Rayman, Bandanna Dee, Banjo and Kazooie, Ridley, which may reflect that their supporters, more likely than not, don’t actually want any more veterans if they can choose.

    And there are those that I suppose are affected due to another character in the franchise being a vet, and that’s obviously Krystal, who has roughly the same weight (B halved) for both choices.

    Again these are only possibilities, and may not really reflect what’s going on, but yeah. It’s not that simple tbh.

    1. I’m more interested about some situational changes though. Say, what if the number of characters in is decreased to 3, or 2, or 1 while the other is already confirmed to be Wolf, etc? Would it change the popular one’s choice or what?

      Also when having these “multiple choice” votes, I think it’s better to give a “rank” to them as well, seeing if they just put a popular character on the last for the sake of filling out or “for the people who campaigned for them”, or really supporting the character.

    2. That’s true there’s a lot of ways the data can be interpreted. For the % I took the total number of responses and divided it by the number of mentions. For your halving bit, are you talking about halving “B” in order to create two–more comparable sets of data to “A”?

Leave a comment below!