I can't for the life of me understand why commander Jim McDivitt thought he could eyeball a rendezvous - point the nose and thrust. The futility of that technique is one of the first things even amateurs like me learn about orbital dynamics. They had plenty of trajectory experts - Bill Tindall and his team plus Buzz Aldrin was an expert on rendezvous.

  • 11
    $\begingroup$ Last year, I asked Buzz Aldrin this very question. His answer was vague, something along the lines that he was too new, when the Gemini IV crew began training, to have an impact. However, he did tell me that he was more involved in the planning of Gemini 6A, the first mission to accomplish a successful space rendezvous. Maybe I'll drill Buzz about this again the next time I see him... $\endgroup$ – Digger Jan 5 '19 at 19:16
  • 8
    $\begingroup$ Another thing to keep in mind was that the Gemini program had four main objectives, three of which [to demonstrate long duration (by their standards) spaceflight, spacewalking, and space rendezvous] were to be investigated, for the very first time, by the crew of Gemini IV. One wonders how much time was spent training for the rendezvous task during what must have been a very compressed training timetable... $\endgroup$ – Digger Jan 5 '19 at 19:21
  • 10
    $\begingroup$ You're an amateur who learned about it by studying information that had to be gathered somehow. $\endgroup$ – chrylis -cautiouslyoptimistic- Jan 5 '19 at 19:22
  • 7
    $\begingroup$ I think @chrylis is saying that it only seems obvious even to amateurs today because of hindsight, and that this hindsight is based, among other things, on the experience gained by Gemini 4. $\endgroup$ – David Richerby Jan 6 '19 at 17:38
  • 2
    $\begingroup$ The "point the nose and thrust" technique isn't always futile - it's a perfectly valid option if you're fairly close to the target. $\endgroup$ – Sean May 2 '20 at 21:51

The long comment chain below this answer highlights the mis-conception that NASA astronauts as a whole did not understand the orbital mechanics of docking.

As this comment points out, the mechanics was well understood at the time, and at least one astronaut had written a thesis on the topic a few years earlier:

... Aldrins thesis about orbital rendezvous - which turned out to be an important cornerstone for NASA - is from 1963, just two years prior. So I think its at least probably that not everyone was familiar with it. I think they were knowledgeable about orbital mechanics, but the quote is about rendezvous - which can be very counter-intuitive. Aldrins thesis is about exactly that factor ;) I think its more that they didn't have the proper procedures, not that they couldn't figure it out in theory. They would probably have absolutely loved a simulator like KSP just to learn procedures ;)

My emphasis added above.

See Line-of-sight guidance techniques for manned orbital rendezvous

  • 1
    $\begingroup$ I searched the pdf of Aldrins thesis for the time interval used for computer simulations of orbital maneuvers and found only a 15 seconds interval. 15 seconds is adequate to simulate a maneuver of 15 or 30 minutes length. But to simulate the very last phase of a rendezvous with a duration of some few minutes, a much shorter computing interval should be used, about a second or even less. May be the numerical errors were too big when using so short intervals. In 1963 there were no computers doing megaflops with double precision available. I found no occurences of the words "floating point". $\endgroup$ – Uwe Jan 5 '19 at 13:58
  • 6
    $\begingroup$ @Uwe people were much better then than they are now at getting very good results with small computing power. I will not have the discussion in comments here, but to do this simulation in order to understand these effects, you can just use one-body orbits in a spherical potential using very simple math. You do not need the kind of numerical power you are suggesting. These are practically circulular orbits and 10 second timesteps with 4th order Runge-Kutta is all you need, if that. You can even throw in J2 without much extra crunching. $\endgroup$ – uhoh Jan 5 '19 at 14:04
  • 4
    $\begingroup$ @Uwe but if you are concerned, why not ask a new question about the calculations in Aldrin's thesis? That would make more space for this very interesting topic! $\endgroup$ – uhoh Jan 5 '19 at 14:07

Gemini 4 was the first unsuccessful try of a rendezvous. They sought at that times it should be possible to rendezvous from a short distance by simply thrusting towards the docking object. They had to learn it the hard way that this strategy works only on very, very short distances and in a short time.

The circumference of a low Earth circular orbit with a height of 200 km is 41,286 km. A distance of 100 m is only 2.4 parts per million of the full orbit. Hard to believe at that times that about 100 m is still too far away for "point the nose and thrust". How close is neccessary for a successful "point the nose and thrust" rendezvous would be another good question.

If we look at the orbit period of about 90 minutes, 1 minute seems to be too long for such a "point the nose and thrust" maneuver. But how short is short enough, some seconds?

An exact digital simulation of such a rendezvous maneuver may be done using a simple single personal computer these days but could not be done in real time using the largest available computers in 1965 lacking a sophisticated and fast graphic display.

But they were successful with Gemini 6A only a half year later.

For more information see this very similar question.

  • 2
    $\begingroup$ "500 feet"; as a longtime KSP player I say that's a pretty bad failure of point and burn; anything less than 2km suffices and KSP's planet is smaller yielding much faster diversion. On the other hand, no docking radar is annoying. $\endgroup$ – Joshua Jan 6 '19 at 1:30
  • 2
    $\begingroup$ @Joshua, KSP also has much higher delta-V budgets than NASA. $\endgroup$ – Mark Jan 6 '19 at 2:09
  • $\begingroup$ @Mark: Will 15 m/s do? That's about what I'd use. $\endgroup$ – Joshua Jan 6 '19 at 3:51
  • 1
    $\begingroup$ @Joshua Also, docking in KSP is often done at much higher speeds. When the space shuttle docked with the ISS, it would approach at 0.03 m/s (nasa.gov/pdf/593865main_AP_ST_Phys_ShuttleODS.pdf), while, in my KSP playing, I usually approach the rendez-vous at around 20+m/s, depending on the situation, and even make the final approach for docking 50x faster than the shuttle. $\endgroup$ – Jarred Allen Jan 6 '19 at 6:26
  • $\begingroup$ @Joshua, the total delta-V expenditure during rendezvous for Gemini 6A was about 80 m/s, including the plane change, phase change, and establishing an intersecting orbit. The portion corresponding to your "15 m/s" is probably the "braking maneuver", at 19.8 m/s, although it might be the "terminal phase" maneuvers, at 38 m/s. (Terminal-phase maneuvers were initiated at a distance of 320 km.) $\endgroup$ – Mark Jan 6 '19 at 7:16

First, the Gemini IV maneuver was station-keeping, not rendezvous. Since the target was the just-separated upper stage, the two spacecraft were already rendezvoused, and point-and-burn would have worked if they'd done it properly.

According to the Gemini IV mission report, the main causes of station-keeping failure were a mix of procedural mistakes and inadequately-aggressive maneuvering:

Review of these figures shows that the velocity increments applied through 00:09:21 g.e.t. succeeded in reducing the separation rate, but left a residual rate of 1.5 ft/sec away from the launch vehicle. As a result, the range from spacecraft to launch vehicle increased to 0.84 nautical mile and the range rate increased to 6.5 ft/sec by 00:30:25 g.e.t. when corrective action was initiated...


At this point (00:52:00 g.e.t.) a relative velocity of 8 ft/sec normal to the line of sight existed. This velocity propagated into a separation distance of 1.6 nautical miles and a separation rate of 17 ft/sec by the time corrective action was initiated at 01:05:30 g.e.t. The corrective thrust applied was insufficient...


It appears that if a procedure had been followed that required the crew (1) to initially establish a clearly perceptible closing rate with the target at all times and (2) to again establish a perceptible closing rate any time the range became larger than several stage II lengths, then the closeup station-keeping goal could perhaps have been achieved.


During the station-keeping exercise, the critical nature of rate determination was demonstrated. After separation, following the four thrusts back toward the launch vehicle, a rate of 1.5 ft/sec away from the stage II existed, whereas, a rate toward it should have been established. The range was approximately 1800 feet at this time. Later, at the point of closest approach, an 8-ft/sec rate existed, normal to the line of sight, which should have been removed.


The ability of a flight crew member to determine rates of the target even in daylight is considerably impaired without a stable background or familiar objects in the foreground. At night, the ability to determine rates depends on the relative distance between two reference lights if they are both visible. If only one light is visible, the flight crew member's judgement depends on his ability to measure the intensity of the light, and, if this one light is flashing, the task becomes very difficult.

In short, by the time the crew established a trajectory towards the target, a considerable lateral velocity had built up, and they had no reference point to judge that velocity against. If they'd established a closing trajectory immediately after separation, they would have been able to use the upper stage as a reference; if they'd approached from a different direction, they would have been able to use the Earth or the background star field as a reference.

(And for you KSP players, the Gemini FDI doesn't come with target and anti-target markers.)


The technical and organizational causes of the Gemini IV station-keeping failure are discussed in this conference paper. John Goodman, “A Cautionary Tale of a Secret, a Small Team, an Accelerated Schedule, and the Gemini IV Station-Keeping Failure,” 43rd Annual AAS Guidance and Control Conference, Breckenridge, CO, January 30 to February 5, 2020.

paywalled here

Below is a summary, the paper has a lot more detail.

Expertise resident in Project Gemini in relative motion and mission planning that could have resulted in a successful station-keeping attempt was not applied to the mission due to the secrecy surrounding the EVA and station-keeping activity and the very short schedule to prepare for the flight. Personnel in the NASA MPAD and the NASA Crew Safety and Procedures Branch that developed relative motion control techniques for the Gemini VI rendezvous mission were not involved in the planning for the Gemini IV station-keeping exercise.

Knowledge of the station-keeping exercise was limited to a small team of crew support personnel who were apparently not knowledgeable in the control of relative motion. The Gemini engineering simulator at the McDonnell plant in St. Louis had not yet been modified to support Gemini procedure development and crew training for relative motion. The lack of a station-keeping capability in the engineering simulator meant that engineers could not verify the station-keeping concept of operations before it was presented to the crew. This prevented the crew support personnel (and the crew) from testing their mental model of what station-keeping would require. Simulator execution of the proposed procedures might have revealed the technical causes of the station-keeping failure, and the need for inertial line-of-sight control and a collimated reticle. The first time the station-keeping procedure was tested was during the flight of Gemini IV. This violated a cardinal rule of spaceflight operations: never train a crew on an unverified procedure.

The crew understood that inertial line-of-sight control was required to establish and maintain an intercept trajectory with the second stage so that station-keeping could be established, but that technique required a star field or collimated reticle to serve as a reference to an inertial frame. The crew did not perform inertial line-of-sight control since they did not have a cue for thrusting orthogonal to the line-of-sight. They did have a cue for thrusting along the line-of-sight based on visual observation of the opening or closing rate. Therefore, they attempted to overcome orbital mechanics by brute force, by thrusting along the line-of-sight, until they became concerned about propellant consumption. This was the only option the crew had to attempt to establish station-keeping, given the limited information available (a window view of the second stage).

Comments about the “non-intuitive” nature of orbital mechanics, “trying to fly it like an airplane,” or “ingrained instincts” do not explain the crew’s actions. These simplistic explanations sound reasonable to people not familiar with the control of relative motion in orbit, but are misleading and not correct. The crew had a better understanding of orbital mechanics than writers have assumed.

There is no evidence that the Gemini IV station-keeping failure was a major learning incident that influenced the development of rendezvous techniques that were flown on Gemini VI-A, or on later Gemini and Apollo missions. It was a learning experience for the small team of crew support personnel that planned the station-keeping exercise, but it was not a learning experience for NASA MPAD and NASA Crew Safety and Procedures Branch rendezvous personnel who were planning the Gemini VI rendezvous mission.

The real “discovery” of Gemini IV was that it is easy to underestimate the risks and complexity of a task that has not been done before, particularly when a small team of support personnel is under schedule pressure and does not possess the right knowledge and simulation tools, and the personnel with the right knowledge are not included in mission planning.

  • 2
    $\begingroup$ Can you summarize? Otherwise this is fairly useless to those who don't have access. $\endgroup$ – Organic Marble Jan 11 at 4:42
  • 1
    $\begingroup$ Wonderful edit, thank you! Just fyi if some of these are direct quotes from the paper you can just add a > at the beginning of each paragraph; that indents it and let's us know it's a quote from your source and not your own words. If these are your own summaries in your own words, then this is fine as-is. $\endgroup$ – uhoh Jan 11 at 23:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.