Unit 2: the Forecast is always wrong, Video 1: Cost Estimates

Summary

In this video, Richard de Neufville explores the inherent unpredictability of forecasting, particularly in project cost estimation. Using the example of resurfacing airport runways, he highlights how factors like energy prices and labor costs lead to significant variations between projected and actual costs. De Neufville extends this discussion to more complex projects, like those at NASA, where the cost deviations are even more pronounced. He emphasizes that forecasting, even by experts, is often wrong due to unpredictable external factors, urging project managers to consider this uncertainty when planning long-term projects.

Takeaways

• 😀 Forecasts about the future are often inaccurate, even when based on seemingly simple systems.
• 😀 Even straightforward tasks, like resurfacing runways, have large cost variations despite predictable inputs.
• 😀 The ratio of actual costs to estimated costs can vary significantly, often doubling or halving the predictions.
• 😀 External factors like energy prices (e.g., oil) and economic conditions drive large fluctuations in costs.
• 😀 Labor costs can also fluctuate based on economic activity, influencing overall project costs.
• 😀 Simple activities, such as runway resurfacing, can experience cost overruns or savings due to unpredictable external factors.
• 😀 The variation between predicted and actual costs increases with the complexity of the project.
• 😀 In high-tech industries, like NASA, the cost variations are even higher, with discrepancies reaching double the initial estimates.
• 😀 Standard deviation, which measures how dispersed the actual costs are from predictions, can be as high as 50-60% for many projects.
• 😀 Forecasts are inherently wrong due to changing conditions and unexpected events that impact project outcomes.
• 😀 Despite expert knowledge, the unpredictability of the world makes long-term forecasts especially challenging.

Q & A

What is the main topic discussed in the transcript?

The main topic is the challenge of accurately forecasting project costs and the inherent errors in such predictions, even for seemingly simple projects like resurfacing runways.

Why does Richard De Neufville state that 'the forecast is always wrong'?

De Neufville points out that forecasts about future costs or performance are often inaccurate due to various unpredictable factors, such as fluctuating energy prices and economic conditions, which make it difficult to get precise estimates.

What is the significance of the 'ratio of real cost to estimated cost' graph presented in the transcript?

The graph illustrates how actual costs can differ significantly from the forecasted costs. It shows that even for a simple task like resurfacing runways, the real cost can be much higher or lower than the estimated cost, highlighting the unpredictability of forecasts.

What specific example does De Neufville use to explain forecasting inaccuracies?

De Neufville uses the example of resurfacing airport runways, where the costs of materials like asphalt can fluctuate due to factors such as energy prices and labor costs, leading to significant discrepancies between estimated and actual costs.

How much variation in cost prediction does the graph show for runway resurfacing projects?

The variation in cost prediction for runway resurfacing projects is significant, with actual costs being either half or double the forecasted costs, showing a wide range of possible outcomes.

What external factors contribute to the forecasting errors discussed in the transcript?

External factors contributing to forecasting errors include fluctuations in energy prices (e.g., oil), economic conditions (e.g., recessions or booms), and changes in labor costs (e.g., overtime during high economic activity or low bids during recessions).

How does the forecast error in high-tech projects differ from that in simpler projects?

In high-tech projects, such as those conducted by NASA, the variation in forecasted and actual costs is even higher, often up to twice the estimated cost, with a standard deviation of more than 60%, compared to 50% for ordinary projects.

What is the standard deviation mentioned in the transcript, and what does it indicate?

The standard deviation is a measure of how much the actual costs deviate from the average predicted costs. A standard deviation of 50% or more, as mentioned in the transcript, indicates a high level of uncertainty and variability in cost predictions.

What lesson should we take from the examples presented in the transcript?

The main lesson is that forecasting future costs and project performance is inherently uncertain. The world changes in unpredictable ways, and forecasts should be treated with caution, acknowledging the possibility of large discrepancies.

What is the overall message about experts and their ability to forecast project costs?

While experts are skilled in making predictions, the transcript emphasizes that even their forecasts are often inaccurate due to uncontrollable external factors, and that the complexity of forecasting is underestimated.