Radjaïdjah Blog

Stacker is a popular arcade game designed and manufactured by LAI games and commercialized by entertainment distribution companies^[1]. It is a successor of Claw cranes.

The goal of the game is to align rows of moving blocks on top of each other. A player who can stack eleven rows will win a minor prize, which is usually low in value, sometimes lower in value than the amount of money the player paid to play the game. A player who gets to the top row wins a major prize. The major prize varies from machine but will often include game consoles, cellular phones, and MP3 players.

The gameplay is the following. There is a row of three cubes which move side to side on the screen, at the first row. When the player pushes the start/stop button, the row of squares will stop. Then, another row of three moving squares appears above the previous row, moving faster than the one before it. If the squares do not align directly above the previous set, any overhanging squares will be removed. If the player misses completely, the game is over. The number of available squares is automatically reduced to two, then one, during the game. The goal is to consistently get the squares directly above the previous set, "stacking" them to the minor prize and ultimately major prize levels.

In a typical game, what is observed is the following. The player easily manages to get to the minor prize level, refuses it and carries on. At this point three consecutive successful moves lead to winning a major prize. But from now on, the unit to align with the stack is a single block. Usually if the player is focused enough, he achieves to stack the next block and the penultimate one, however stacking the last one fails by a very small fraction. As a result the player loses his coin. And because he was so close to win, he might want to try again, with the same outcome.

The most interesting part is the last step, the one leading in general to a game loss.

Why does the so far very skilled player lose at the last step? Is it the pressure of winning a phone? The answer is: because the last step is different. It looks the same as the previous ones, but it is not.

Pressing the stop button while the last single block is displayed at the correct position does not necessarily leads to a win. In fact most of the times it leads to a loss, and a posteriori the block is displayed mis-aligned with the stack. Only in certain specific cases a win will be obtained.

LAI Games Stacker operator's manual describe all the internal parameters the owner can set to configure the machine. One of these operators is "P10 = Skill Setting (Major price)", whose description is (Default 8) (Adjustable 1 – 10) This option sets the Skill level for players to reach the Major Prize level, as listed in the table below. As this is a skill game the win rate is only the approximate rate for each difficulty setting. Follows then the corresponding options, with P10 = 1 implying approx 1 win in 20 games, whereas P10 = 10 implying approx 1 win in 800 games. This parameter does not affect the displayed game at all, "P08 = Cube Speed" is a different parameter.

In practice, Stacker behaves as the following. For the sake of simplicity, let's imagine that all players are super skilled (they always press the stop button when the last block is correctly aligned), and that P10 = 8 (approx 1 win in 400 games). Instead of having all players win, only 1 over 400 will win. How to achieve this? Once could propose to throw a 400-side dice to decide if a game is a win, but the skill aspect of the game would be instantaneously lost. A more elegant way to do it is to define ''concretely winning states" as a subset of "appearantly winning states", like the following pseudo-code does:

if block is well aligned:
    if internal_timer_in nanoseconds % 400 == 0:
        proceed to winning
    else if stop_button was pressed before block_duration / 2:   
        move displayed_block to previously_displayed_block
    else:
        move displayed_block to next_to_be_displayed_block
    proceed to losing

This is compatible with purported LAI's statement that Stacker is "100% a game of skill and although it is very difficult, every game played can be a winning game".

However, this theoretically skill game is in practice a luck game, so Stacker is practically a gambling machine. This great video comes to the same conclusion for another arcade game, Cyclone.

Psychologically, having the player close to winning but finally losing is part of the design of this arcade game. Stacker operator's manual describes the machine as a bright and attractive display, simple but exciting game play and a real “Ahh! Just missed” feeling. Indeed the player can see that he was very close, and the belief that the loss is due only to a very tiny amount of bad luck triggers the envy to replay with the confidence to be more accurate next time and win. In reality it can be said that the progressive feelings of self-esteem building with the initial successes, frustration while losing by a near-miss, and hope while deciding to replay, have been entirely engineered by the game designers. You see a game that you control, you live a game who controls you.

Let's conclude with LAI Games' motto: Fun and Profits Go Hand-in hand.

Voici une image.

Cette image affiche son propre hash md5, d'où son titre de hashquine en hommage aux quine de Gödel Escher Bach, programmes qui affichent exactement leur propre code source.

PS : si vous trouvez une collision de SHA256, RIPE160, RIPE160(SHA256), ou SHA256(SHA256), il est encore temps de clamer votre récompense. Pour SHA-1, c'est trop tard...

Suite de l'article Robozzle (parlant de LightBot et Robozzle), sur les nouvelles méthodes pour apprendre à coder.

Publiée en français aux éditions 404, la série Les codeurs de l'ombre (VO : Secret Coders) (T1 T2 T3) raconte les aventures d'une élève arrivant dans une nouvelle école. C'est une série initiatique sur l'apprentissage de certains concepts de programmation (binaire, variables, fonctions).

Les éditions 404 ont également publié des variations littéraires des escape games, des escape books, sortes de livres-jeux dont le but est de s'echapper du livre. Par exemple : Prisonnier des Morts.

Tynker est également une approche ludique intéressante pour apprendre à coder.

Enfin, pour s'initier à la logique sous pression, signalons Keep Talking and Nobody Explodes (KTaNE pour les intimes), un jeu coopératif dont le but est de désamorcer des bombes en équipe. Le manuel de déminage a été traduit en français.

Tout cela est très prometteur, si vous connaissez des initiatives francophones similaires n'hésitez pas à les signaler en commentaire.

...ou comment s'initier à la programmation via des jeux sérieux (serious games).

Lightbot est un petit jeu de logique qui propose aux plus jeunes et aux débutants d'appréhender des concepts de programmation comme les instructions, les branchements conditionnels, les boucles, etc, sans avoir à taper du code. Lightbot est un jeu de logique, un casse-tête dont le fonctionnement est basé sur des concepts de programmation. Le jeu est très esthétique, l'objectif de chaque niveau est d'aider un robot à allumer des cases au moyen de commandes.

Jouer en ligne à Lightbot.

Application version junior (4-8 ans) pour smartphone/tablette : Android, iOS.

Application version standard (9 ans et plus) pour smartphone/tablette : Android, iOS.

Robozzle est un jeu plus avancé créé par Igor Ostrovsky, dont le principe est très bien expliqué par le CANOPÉ d'Amiens : étant donné un terrain (constitué de cases carrées bleues, rouges, vertes ou "vides"), comportant une liste d'objectifs (des étoiles), et une position et direction initiale pour un robot, il faut fournir un programme pour le robot qui le fasse passer par tous les objectifs, en un nombre limité (1000) d'étapes et sans sortir du terrain ni passer par une case "vide".

Le programme du robot est organisé en plusieurs "fonctions", nommées F1 à Fn (avec n variant selon le niveau), et comportant chacune un nombre de pas de programme limité, selon le niveau. Sans inscription pour jouer directement ou avec inscription pour créer vos propres puzzles. Le "langage" du robot est constitué des instructions suivantes (selon le niveau, toutes les instructions ne sont pas nécessairement disponibles) :

• Avancer
• Tourner de 90° à droite
• Tourner de 90° à gauche
• Appeler une des fonctions (le robot gère une pile d'appels classique)
• Fixer la couleur de la case actuellement occupée par le robot à une des trois couleurs de terrain.

Toute instruction peut de plus être conditionnée par une des trois couleurs de terrain, auquel cas le robot n'exécutera cette instruction que s'il se trouve sur une case de la couleur correspondante.

Quelques niveaux standards / rigolos : Simple Loop (màj : ne semble plus exister), , Simple Cross,Flooder (màj : ne semble plus exister), Cube Extreme, Turn around.

Pour apprendre l'empilement et la récursion : Recursion Review, Learning Stack, Boomerang, Second kind of memory, Limit yout stack!.

Plus difficiles mais qui en valent la peine : learning the call stack, Power-of-2 generator.

Jouer en ligne à RoboZZle (version javascript), il y a aussi une app : RoboZZle Droid.

Fix the Factory est un jeu similaire, développé par Lego, où un robot doit traverser une salle en contournant les différents obstacles.

Une autre façon visuelle de programmer est la programmation diagrammatique, du style de Lego Mindstorms ou de LabView.

Un jeu de plateau similaire et plus guerrier : Roborally.

Scratch, enfin, est un environnement pédagogique complet développé par le MIT dédié à l'apprentissage de la programmation.

Source : CANOPÉ Académie d'Amiens

Sans trop s'avancer, toute personne qui a lu le livre Gödel, Escher, Bach de Douglas Hofstadter a essayé de filmer un écran ou de mettre face à face deux miroirs se reflétant à l'infini.

Une expérience de pensée étendue consiste à fabriquer une boîte dont l'intérieur ne serait constituée que de matériau réfléchissant (des miroirs), et à imaginer ce que quelqu'un sur place verrait avec toutes ces réflexions (on peut aussi penser aux labyrinthes de miroirs dans les parcs d'attraction).

En mode cubique, une telle boîte kaléidoscopique est facile à fabriquer logiciellement^[1] avec un traceur de rayons tel povray. Vue de l'extérieur, une apparence naturelle de cube innocent :

Mais à l'intérieur, on ajoute quelques bougies pour avoir une source de lumière, et un univers sans fin est créé :

Voilà pour la théorie. Reste à voir ce que ça donne en pratique.

Mise à jour (août 2013) : pour la pratique, voir ici.

The famous internet RFC 2795 (The Infinite Monkey Protocol Suite) states that a monkey hitting keys at random on a typewriter keyboard for an infinite amount of time will almost surely type a given text, such as the complete works of William Shakespeare.

A modern variation of this monkey-fed typewriter consists of programs outputting random text sequences from a words database and a set of rules called grammar. Such programs are nicknamed generators.

Examples of generators in the area of science are SCIgen who produces meaningless computer science papers, or Mathgen who creates nonsensical math papers. The latter provides even the source code to generate full-size random e-books featuring custom-named authors, which can then be printed out at Lulu.

One can think that target journals for publication of such gems should be world-class journals such as the Journal of Irreproducible Results or the Journal of Universal Rejection, where no failure is to expect.

But it happens that from time to time so-called random papers are accepted in real-world publications. For example, a paper entitled Deconstructing Access Points has been accepted in The Open Information Science Journal, whereas Independent, Negative, Canonically Turing Arrows of Equations and Problems in Applied Formal PDE has been accepted in Advances in Pure Mathematics. As far as conferences are concerned, the talk Rooter: A Methodology for the Typical Unification of Access Points and Redundancy has been accepted at the WMSCI2005 and the talk Towards the Simulation of E-Commerce has been accepted at the 2008 International Conference on Computer Science and Software Engineering.

This highlights that some slimy journals or conferences are ready to accept fees to publish irrelevant non-peer-reviewed papers while it may be deduced that some authors are ready to pay (institutional) money in exchange for easy CV bullets.

It also reminds of Georges Perec's Experimental demonstration of the tomatotopic organization in the Soprano (pdf) and of the Sokal affair^[1], which was more of a critic of post-modernism.

Update (March 2015): a more recent hoax in sociology has been ployed and unveiled by Arnaud Saint-Martin and Manuel Quinon after their postmodern-funny-gibberish paper entitled Automobilités postmodernes : quand l'Autolib' fait sensation à Paris has been accepted and published in Sociétés, a journal whose editor-in-chief is the French iconic sociologist Michel Maffesoli.