Steady or Sharp?

A courier app predicts how many minutes each delivery will take. Two candidate models were tested on the same six orders:

order	1	2	3	4	5	6
actual time	34	47	29	52	41	63
Model Steady	29	51	23	57	38	69
Model Sharp	33	47	31	51	42	46

With residual $r_i = \text{actual} - \text{prediction}$ for each of the $n$ orders:

\text{MAE} = \frac{1}{n}\sum_{i=1}^{n} |r_i|, \qquad \text{RMSE} = \sqrt{\frac{1}{n}\sum_{i=1}^{n} r_i^{\,2}}

Customers shrug off an estimate that is a few minutes out, but a single badly wrong one generates complaints. So the team deploys whichever model has the lower RMSE.

What is the MAE, in minutes, of the model they deploy? Round your answer to 2 decimal places.

Steady or Sharp?

Hints