Documenation

[Christine Lam]

When I first came up with the idea of Tetris, I had many ideas on a very large scale. I wanted to compare statistics between different versions of the game over time, and how each one was presented. Then I realised that I would have to collate all the data. Although I do enjoy playing the game, I wasn't ready to play 50 games of Tetris and then analyse each game, piece by piece, level by level.
What I did do though, was analyse 1 particular game, in detail. This meant more specific data, instead of many averages. I chose to analyse a game of my favourite version on Tetris - Windows Tetris, from the Best of Windows Entertainment Pack of the early 90's.

The data collected relates to me in that it was an actual game of Tetris which I analysed and collated data from.
I divided the stats into the 10 levels, so it would be easy to compare how each level is presented. Within each level, I tallied how many times each piece occurred.

This is so I can visualise the ratio of pieces within each individual level. This was done simply by displaying pieces which occurred more times larger than pieces that didn't occur many times.

I noticed that level 1 was a good, rounded up sample of all the pieces, with all pieces occuring many times in the time it took to form 11 lines. From there on, levels had less pieces, some levels not even featuring some pieces. I also noticed that there were outliers, in the form of levels 2 and 5. I put this down to the number of lines that had needed to be achieved to reach the next level. I cleared level 1 with a Tetris, and therefore there was a high number of lines going into the level. That meant that fewer lines were needed to progress to the next level, and less pieces were therefore required.
To determine the size of the shapes in relation to the other levels, I scaled the numbers so levels with more pieces had a larger surface area than levels which had very few pieces. These numbers were what I used to scale the pieces when I modelled the pieces. Because the size of the pieces is in relation to the game, a piece that occurred the same number of times in different levels, they would also fit the mould for the other level. Eg, The 'S' shape occured 3 times in levels 6, 9 and 10. The 'S' shape that I have cut out will fit into the moulds in any of those levels.
The numbers below represent the scale factor of each piece of each level. It was the number entered into the scale tool in Maya to come up with the size of shape.

I also tallied up how many pieces were in each level as a whole, and how many lines were achieved in each level.

These figures were used to create etches on the base of each level's mould.

Now we have an understanding of the data that I collected, I thought of ways to represent the data as interactively as possible. I see that Tetris is a game fill of puzzles, so I decided to create a cube with 10 jigsaw puzzles, each one representing a level of the game.
As Tetris is a game, I wanted the viewer to be able to pull apart a game of Tetris and see how all the pieces link together to form a game. I also wanted to show that the pieces don't always fit together, which is also an aspect of the game.

I decided to make the moulds of each level with clear perspex so the viewer can see through the levels and compare them. I used a different colour for the pieces of each level as a way to group the pieces.
Each mould consists of 2 sheets of 2mm clear perspex. The bottom layer has the layer labelled on it, as well as etched lines representing the number of lines in that corresponding level. The upper layer is the actual mould where pieces are to lay.
The moulds are to show the user where to put the pieces, which are 3mm thick and a shade of transparent perspex. The completes mounds represent the lines, shapes and significance of the level in relation to the rest of the game.

I've attached magnets to the top and bottom corners of all the moulds, so that the puzzle can be pulled apart and rearranged. This is useful for when the user may want to compare different levels on top of each other.

I had the intention of mapping the spread of each piece within each level on the pieces, in terms of clustering.

I had collected similar data for all the other levels: tetris spread.xls
Looking at each column, I had wanted the clusters or lack thereof to be shown on the pieces themselves, but I had scaled the whole puzzle down to cut down costs. I didn't think it was plausible to etch markings on the smaller pieces.