atduskgreg

• Graph the rest of the games
• For (174) graph from black's POV two sets of centipawn scores (for the same moves): => high search depth vs low search depth => we're looking for "false hope"
• Python script that starts from an empty game, plays through best move for black and white and logs the centipawn score for each played move (graph that) (ruby code for reference for basic architecture: https://gist.github.com/atduskgreg/8951065)

atduskgreg

• Graph of successfully detected boomerang: http://chesscomputer.tumblr.com/post/77384413487/graphed-above-from-whites-pov-boomerang-w
• Graph of non-boomerange: http://chesscomputer.tumblr.com/post/77386675090/richter-amateur-174-b-h8h1-cp-274-w-g1h1-cp
• False Hope Detection: when full-depth and limited-depth search predicts the same set of moves, but limited-depth scores them higher
• Graph of centipawn score over the course of the game shows moments of drama (crossings, big changes, changes of direction, etc.). We can detect these. Automatic highlight reel generation?
• Working on for this week: => graph full game from start to finish => graph the rest of the boomerang games => start automatic graphing process

atduskgreg

class LoopingFile {
  String[] lines;
  int currentLine = -1;
  
  LoopingFile(  String nameOfFile){
    lines = loadStrings(nameOfFile);
  }
  
  int currentLineNum(){
    return currentLine;

atduskgreg

• Meet Shannon
• Shannon update on automation progress
• Twitch! Chess network stream of world championship match: http://www.twitch.tv/chessnetwork/b/481728593
• We should use Twitch to prototype (when we're ready). April? May?
• Parallels with Philip Tan's work on Starcraft
• Discussion of time in chess

atduskgreg

* Produce a PGN log of each game that we play (save as a file)
* Find a PGN "validator"
* Also, log the info output from the engine (annotated with PGN move number)
* Convert to fsm

atduskgreg

"A lot of my interest in Apocalypse Now was carried over into Star Wars. I figured I couldn’t make that film because it was about the Vietnam War, so I would essentially deal with some of the same interesting concepts that I was going to use and convert them into space fantasy, so you’d have essentially a large technological empire going after a small group of freedom fighters or human beings."

-- George Lucas, quoted in The Making of Star Wars

atduskgreg

Webserver that responds to
GET http://localhost/boomerang?f=asdfasdfasd/asdfasdfasdf/asdfasdfasfasdf/ b - -
=> csv with centipawn time series from white's pov

atduskgreg

import gab.opencv.*;
import processing.video.*;
import java.awt.Rectangle;

int minBlobArea = 500;
float smoothAmount = 0.8; // 1.0 is never move, 0.0 is don't smooth

OpenCV opencv;
Capture cam;

atduskgreg

				p = []
				lastMouse = false;
				function setup(){
					resetDrawing();
				}

				function draw(){
					if(mouseIsPressed()){
						p[p.length-1].push({x: mouseX, y: mouseY});
						lastMouse = true;

atduskgreg

# -*- coding: utf-8 -*-

# 1. Generate 100 points x uniformly distributed between 0 and 1, and let y = 2+3x+ζ,
#    where ζ is a Gaussian random variable with a standard deviation of 0.5. Use an
#    SVD to fit y = a + bx to this data set, finding a and b.

import numpy as np

# x is uniformly distributed
x = np.random.random(100)