Am i missing something about special tile placement?

Wumpushunter

If everything was weighted the same scarlet witch should place her auto countdown 1.5% chance in each spot on the board. But over an observational period of 500 placements she did favor the corners by far over this amount. And even the outside edge of the board was overly favored. The inner 28, 29, 36 and 37 square was actually under represented by a large amount.

Less accurate but still concerning is the chance that her tile will appear in a situation in which it can be automatically matched away right after placement. In 500 observed cases her tile could be matched away directly in about 36% of the times.

Is their weighting I am not aware of in special tile placement? Is it official?

meekersX

As far as I know, it is completely random. I wouldn't be surprised if their code somehow does not ensure this though.

I have two immediate thoughts. First, the placement would be 1/64 = 1.5625% in each spot, which rounds up. Second, I'm curious what exactly the distribution you're seeing is. 500 trials probably isn't enough to establish significance unless there is something majorly wrong with the algorithm.

Also, I don't think it's too surprising that a placed tile can often be matched away. There's 8 surrounding tiles, and colored tiles occur at higher frequency than team-up tiles. This is much harder to control for, as the distribution of colors on the board can be majorly skewed depending on matches made and abilities used.

Talestummy

You also have to factor in that it creates the countdown on a blue tile.

DeNappa

Talestummy said:

You also have to factor in that it creates the countdown on a blue tile.

No, it creates a blue countdown tile. It doesn't place a countdown on a blue tile.

Also, @Wumpushunter, for that data to be 'complete' i think it should not only record at what spot the countdown ends up, but also (per data point) a list of spots on the board where it can't be placed because of specials. And I don't know how you could do calculations or estimations with such a set.

An alternative would be to record only instances of countdown placements on 'clean' boards but I don't know if that would be an accurate representation.

Talestummy

Thanks, looks like I misread that badly.

Dormammu

Wumpushunter said:

Less accurate but still concerning is the chance that her tile will appear in a situation in which it can be automatically matched away right after placement. In 500 observed cases her tile could be matched away directly in about 36% of the times.

For this data to have any meaning you have to account for how much blue is on the board. This tends to happen  to me as well, but I often don't have a use for blue on Switch teams and don't match it - meaning my boards are usually heavy on blue.

vinsensual

I always thought it was weird how she creates the blue countdown, every other similar special tile creation I can think of, a basic tile of that color has to exist on the board otherwise nothing will happen. Maybe it's a nod to how powerful switch can be in the comics.

TPF Alexis

vinsensual said:

I always thought it was weird how she creates the blue countdown, every other similar special tile creation I can think of, a basic tile of that color has to exist on the board otherwise nothing will happen. Maybe it's a nod to how powerful switch can be in the comics.

I believe 3* R&G can do that with their blue Countdown as well, tho I haven't actually used that power in ages, so I don't remember for sure. It may have just been a paradigm that the devs were toying with for a bit, but decided not to pursue.

Punisher5784

vinsensual said:

I always thought it was weird how she creates the blue countdown, every other similar special tile creation I can think of, a basic tile of that color has to exist on the board otherwise nothing will happen. Maybe it's a nod to how powerful switch can be in the comics.

That is because SW's Blue CD used to be Beast's Blue CD. Originally for a whopping 8 AP, Beast would create a Blue CD that would either be matched away or when it resolved, the special tiles would match away because they were blue. Eventually Beast got a buff and his lousy Blue CD mechanic shifted to new character SW a month or two later. One character's poor ability made it an awesome ability for another character... damn I am MPQ old lol

Lucifier

Wumpushunter said:

In 500 observed cases.

500 observed cases, that's a lot not to mention un-observed cases (stop using her).
if you need purple generation use 3* gambit 0/5/5 or 0/0/5 instead. (if for charlies angles, from my experience gambit works better than SW).

PenniesForEveryone

Did you account for times where the tile was placed in between two or more other blue tiles, immediately making a match and disappearing?

Jaedenkaal

Wumpushunter said:

If everything was weighted the same scarlet witch should place her auto countdown 1.5% chance in each spot on the board. But over an observational period of 500 placements she did favor the corners by far over this amount. And even the outside edge of the board was overly favored. The inner 28, 29, 36 and 37 square was actually under represented by a large amount.

This is only correct if there are no other special tiles on the board, as others have pointed out. I'm also not sure if it is allowed to replace Team-Up tiles, which is another consideration. Also, there's a 34.4% chance of the tile being created on the edge of the board, assuming no special and/or team-up tiles, which is often (and frustrating) enough to be noticeable.

I'm also not convinced that 500 trials is enough to draw any real conclusions. It's a large number in human experience, but not as far as statistics are concerned.

Wumpushunter said:

Less accurate but still concerning is the chance that her tile will appear in a situation in which it can be automatically matched away right after placement. In 500 observed cases her tile could be matched away directly in about 36% of the times.

Well that depends entirely on the number of other Blue tiles on the board, as well as the number of non-blue match-4s and match-5s, and this state is non-random, because both teams make non-random matches that control the distribution of tiles on the board. This is especially true if Team-Up tiles are ineligible targets. So I don't think anyone should draw any conclusions from this particular observation.

Wumpushunter said:

Is their weighting I am not aware of in special tile placement? Is it official?

There has been no official comment about this (and I have no idea why you'd expect otherwise), and I have no reason to believe that the Arcane Incantation tile is placed using any algorithm other than a generic one that selects a tile at random given a set of targeting restrictions.

Wumpushunter

I use her often I will continue with a larger sample size.