In this experiment, you will be playing a simple computer game where you control a lunar lander spacecraft. The game interface is shown on the right. It is important that you do NOT refresh or close the page. Otherwise, the experiment will remain incomplete and you will not have a chance to continue.

Goal. The game consists of two phases. In the first phase, you just need to stabilize the lander, i.e., keep it upright. The goal of the second phase is to safely and quickly land the spacecraft on the ground, as close as possible to the landing pad between the two flags. A landing is safe if the main body of the spacecraft does not hit the ground. Examples of successful and unsuccessful landing attempts are shown and explained below.

After each round a score will appear on the right. You will get a green check for a successful stabilization (in the first phase) or landing (in the second phase) and a red cross for an unsuccessful one. On average, scores close to 200 and above indicate success.

Even if the lander crashes or goes out of the field of view, you may still get slightly better scores if the lander is close to the landing pad or the collision is mild.

Control Input. You will be using three arrow keys to control the lander. At any time, at most one control input is active, i.e., you are not allowed to fire multiple engines at the same time. If you press more than one key, only the last one will be active. The figure below shows these controls.

Auto Pilot. Finally, we have an auto pilot implemented. The auto pilot takes over the control completely once it is active, but it becomes active only automatically, so you do not have control over when it will be on or off. You can see the state of the auto pilot (on or off) at the top right corner of the game screen (see the figures below).

You will be playing the game for 20 rounds. In some of the rounds, auto pilot can take over the control often and in some it may not interfere at all. It is even possible that auto-pilot be off during all 20 rounds. In the rounds when the auto pilot occasionally becomes active, it becomes active only when you provide some suboptimal control inputs to correct your actions. Regardless of the auto pilot's being on or off, you MUST keep providing control inputs to the lander.

At the end of some rounds, auto pilot may demonstrate you why it interfered by showing you some replays. This is a chance to improve your skills controlling the lander. Specifically, the auto pilot is going to tell you a bad sequence of 3 actions you took, and show you what would happen if it did not interfere. Next, it is going to show you an ideal trajectory that follows the 3 actions it recommends. An example of replays is shown in the figures below.

The first 7 rounds are the first phase, where the goal is to keep the lander upright. You will have 16 seconds in each round in the first phase. In the second phase (13 rounds), you will have 48 seconds to land the space craft in each round. The round automatically ends when the attempt is complete, either successfully or unsuccessfully.

You will be compensated \$3.75 for this study. We are compensating at the rate posted on Amazon Mechanical Turk, i.e. \$0.25 per minute (\$15/hour). Payment is arranged via Amazon Mechanical Turk. We will be offering bonuses for workers who score highly on the games. Workers will have a chance to earn up to \$6.25 total. So make sure you try your best to do well on these games! Now, please complete the 2-question quiz below to start the experiment.

1. What does it mean if the "Auto Pilot" indicator is green (active)?

2. Which of the following is true?

Correctly answer the questions above and agree the terms to activate the login button.
Please enter your AMT Worker ID which can be found on the top left corner of the AMT website.
Your AMT Worker ID: