First Five, more analysis

I looked at how many games total each network won during a given round, and came up with the following:

First 5 Rankings.PNG

The first thing to notice is that there is a LOT of variability here. Individual standings are all over the map.

That being said, the top standings seem to be dominated by the deeper networks. By the end of the 5th round, the depth of the top 5 networks is 4,5,4,5,2. These are also all smaller networks, either being 100 or 250 nodes wide.

If I had to guess about the reason, I would say that a deeper network provides an advantage overall, but that a wider network takes longer to train. If that is the case, then I would expect to see the larger networks begin to dominate over time.

Here’s more evidence of the above, grouping the networks by width and depth:bywidthanddepth

There are clear trends here by the fifth round: The deeper the network, the more likely to win; also, the wider the network, the more likely to lose.

Again, I suspect this is only because wider networks are harder to train. I’m curious to see how this works out over the next several rounds.

Oh, and the Smart Random player? She’s still dominating, enough that she’s only lost one game over the first five rounds, which was probably a fluke.



Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s