Purpose: This study aimed to compare estimations of critical speed (CS) and work completed above CS (D′), and their analogies for running power (critical power [CP] and W′), derived from raw data obtained from habitual training (HAB) and intentional maximal efforts in the form of time trials (TTs) and 3-minute all-out tests (3MTs) in recreational runners. The test–retest reliability of the 3MT was further analyzed. Methods: Twenty-three recreational runners (4 female) used a foot pod to record speed, altitude, and power output for 8 consecutive weeks. CS and D′, and CP and W′, were calculated from the best 3-, 7-, and 12-minute segments recorded in the first 6 weeks of their HAB and in random order in weeks 7 and 8 from 3 TTs (3, 7, and 12 min) and three 3MTs (to assess test–retest reliability). Results: There was no difference between estimations of CS or CP derived from HAB, TT, and 3MT (3.44 [0.63], 3.42 [0.53], and 3.76 [0.57] m · s−1 and 281 [41], 290 [45], and 305 [54] W, respectively), and strong agreement between HAB and TT for CS (r = .669) and CP (r = .916). Limited agreement existed between estimates of D′/W′. Moderate reliability of D′/W′ was demonstrated between the first and second 3MTs, whereas excellent reliability was demonstrated for CS/CP. Conclusion: These data suggest that estimations of CS/CP can be derived remotely, from either HAB, TT, or 3MT, although the lower agreement between D′/W′ warrants caution when using these measures interchangeably.