Disclaimer: I am terrible at making websites, therefore I am not trying anything fancy.
Our mechanism is a simple design where a strong base extends in line with the far hole and supports a stiff truss. At the end of this truss is the servo mechanism with an arm attached; one side of this arm is a fork-like section that catches and lifts the weight, while the other side supports a counterweight.
Our mechanism lifts the weight at a distance of roughly 2.5 inches from the axis of rotation. To lift a 1-lb weight, this requires [1lb*(16oz/lb)*(2.5 in)] 40 ounce-inches of torque. Since this quickly approaches the servo's limit and does not account for inefficiences and slack in the design, we added a one-ounce counterweight at a distance of 6 inches from the pivot. This aids the servo in lifting the weight by providing an opposite moment equal to [1oz*6in] 6 ounce-inches of torque. In theory, this means that the servo needs to only provide [40-6] 34 ounce-inches of torque, roughly 2/3rds of the servo's full output. At a distance of 1.5 inches, with the center of the servo located one inch above the weight, lifting the weight 2 inches requires [arctan(1/2.5)*2] 44 degrees of travel. These torque and travel requirements are well within the servo's limit and should pose no problems.
There are several components of our structure that make it unique. In addition to these calculations, we constructed the base and the cross-members of the truss out of 7075 aluminum, which provides added stiffness due to it's higher modulus of elasticity (the trade-off is that you cannot bend it into tight 90 degree elbows). In order to increase stiffness further and keep weight low, we used aluminum rivets to sandwich the truss members tightly together. Just so everyone could see how awesome this plan was, we specifically ordered blue rivets at no extra cost.
Close-up of our structure
The original, failed design
Front view of support truss. Note the blue rivets.