Successful landings on the Moon happened 50 years ago. Technology (satellites, computer, ML) has come a long way during this time, so why did Israel and India fail recently to land a probe there?

Is the chance of successfully landing a manned mission higher than for a non-manned mission?

Does the US have some secret insight into landing on the Moon?

  • 11
    $\begingroup$ They, and Russia, have had more practice $\endgroup$
    – user20636
    Commented Oct 1, 2019 at 21:00
  • 25
    $\begingroup$ We spent more.... $\endgroup$ Commented Oct 1, 2019 at 21:01
  • 27
    $\begingroup$ NASA software engineers don't come from an Agile "fear of failure is a bad thing" background? $\endgroup$
    – user20636
    Commented Oct 1, 2019 at 22:54
  • 28
    $\begingroup$ There's also the very small sample size. Even with a good survival rate, 2/2 failures is not statistically remarkable. (and why does the sample not include the Chinese landings?) $\endgroup$ Commented Oct 1, 2019 at 23:48
  • 11
    $\begingroup$ @JCRM NASA did plenty of fail fast and early In fact, I expect they did far more - and still they ran into some... lets say "interesting" software faults. $\endgroup$
    – Baldrickk
    Commented Oct 2, 2019 at 14:49

4 Answers 4


Does the US have some secret insight into landing on the moon?

Yes: fail early and often.

The US developed experience with uncrewed landings first, before attempting crewed landings in the Apollo program; those earlier programs had a very high failure rate.

The first US lunar spacecraft were in the Ranger program, which was simply attempting to hit the moon while taking photographs all the way down, and didn't achieve that goal until Ranger 7. The first two Rangers didn't even leave Earth orbit. Ranger 4 was completely inert after separation from its launcher, but it at least hit the moon.

Following Ranger was the Surveyor program, which attempted soft landings. Two out of seven of the US Surveyor missions crashed (#2 and #4).

By comparison, the ISRO lunar program has been very successful; the Chandrayaan-1 and Chandrayaan-2 missions both put spacecraft into lunar orbit, and the landing attempt of the latter came very close to succeeding. I haven't seen a thorough explanation of the landing failure, but my guess is that the root cause was a hardware fault.

Israel's Beresheet was likewise nearly successful, and this was the first spacecraft developed by that team. The failure in this case appears to have been a gyroscope sensor malfunction, which could happen to any spacecraft at any time, but in this case happened at a crucial moment in descent when there wasn't enough time to recover. There's nothing fundamentally wrong with the design, and the team didn't do anything wrong.

Then of course, there's also China's lunar exploration program, with three successful robotic landings (Chang'e 3, Chang'e 4, Chang'e 5), along with rovers and (as of December 3, 2020) a sample-return launch, and apparently no major failures.

Is the chance of successfully landing a manned mission higher than for a non-manned mission?

I would say so, for three major reasons.

  • The US went from uncrewed impact programs to uncrewed landing programs to crewed landing, and incorporated many hard lessons learned in the earlier programs. Anyone else contemplating a crewed lunar landing is very likely to take a similar path.

  • Crewed spacecraft and launchers are held to higher safety standards and provided with more redundant backup options in virtually all spacecraft systems.

  • Automatic systems plus humans can solve many more critical problems than either automatics alone or humans alone; this finding was one of the most important results of the X-15 spaceplane program.

  • 22
    $\begingroup$ You could add that the Soviet programme also had plenty of failures. $\endgroup$
    – gerrit
    Commented Oct 2, 2019 at 10:56
  • 34
    $\begingroup$ @Barmar They have. Chandrayaan-1 was a near total success on the first try at a cost of ~US$56 million. The orbiter portion of Chandrayaan-2 was also successful. Between just those two missions they achieved more than the entire Ranger program did at a cost of ~US1B in today's money (albeit with US instead of Indian labor costs). Despite the disappointment of failing to successfully soft-land, the ISRO's lunar program has been extremely successful so far. $\endgroup$ Commented Oct 2, 2019 at 19:53
  • 6
    $\begingroup$ In addition, pretty much everyone involved in the US Apollo program is retired or dead now, and none of the parts or equipment it used are being made for any purpose anymore. Lots of it nobody knows how to make anymore. So even the US at this point would likely require the same amount of time (or more) if they wanted to do it again. $\endgroup$
    – T.E.D.
    Commented Oct 3, 2019 at 2:16
  • 6
    $\begingroup$ @Barmar Being twenty times as money efficient isn't good enough for you? And it's not like the US is going to give everyone all of that knowledge just like that. Don't forget that the Space race was as much about showing off your ability to build rockets (delivery platforms for nuclear weapons) as anything else. This was, and still is, extremely high-tech, with very few chances of recovery. There's no repair shops in Moon's orbit. $\endgroup$
    – Luaan
    Commented Oct 3, 2019 at 8:25
  • 7
    $\begingroup$ @Barmar "Everything was new technology then, failure was expected." Exactly. If Armstrong and Aldrin had crashed into the moon, people would have said they were heroes who died advancing the frontiers of humanity. If astronauts died on the moon today, people would say, "FFS, we did that 50 years ago." $\endgroup$ Commented Oct 3, 2019 at 20:25

While engineering and available technologies have greatly advanced since the 50's and 60's, safely landing something on the moon is still a highly technical feat with a critically long list of potential failure points.

After a quick look at a list of moon missions, it appears that the US alone has had more launch failures than India and Israel's combined attempts.

When all a mission failure takes is to have a valve's response time in space being a handful of percentage out of spec, it becomes easy to see how a limited number of attempts might not add up to all that many successful missions...

The more launches you have, the more direct data you can gather, and the more collective knowledge and experience an organization has to draw on for future missions, which translates into fewer issues leading to critical mission losses.

  • 11
    $\begingroup$ From another source, I get that smaller rockets imply longer trips, which implies longer cosmic radiation, which implies a higher risk of electronic failure. Sometimes you just have to pick chances and have a heap of funds to build a big rocket. $\endgroup$
    – George W.
    Commented Oct 1, 2019 at 22:49
  • 4
    $\begingroup$ On the other hand, a manned mission has many more parts that might fail. $\endgroup$ Commented Oct 2, 2019 at 21:06
  • 10
    $\begingroup$ Manned missions have a higher rate of success because you don't risk putting a person in a spacecraft until you are sure you can get it right. $\endgroup$
    – Seth R
    Commented Oct 2, 2019 at 22:09
  • 5
    $\begingroup$ @Beanluc, the CO2 scrubber on Apollo 13 gets all the attention, but the crew jury-rigged a whole lot of other things (eg. a procedure for aligning the spacecraft without using the alignment computer). $\endgroup$
    – Mark
    Commented Oct 2, 2019 at 23:03
  • 4
    $\begingroup$ @Beanluc not at all. For example, all US human spaceflight dockings to date have been flown manually. $\endgroup$ Commented Oct 3, 2019 at 1:27

You could also put it this way: The secret insight of the US was to use the following procedure:

  1. Be the richest country in the world.
  2. Over a 13 year period, spend an amount equal to 4.5% of your gross domestic product as of the year you started working.

The recent missions that didn't make it were operating on much lower budgets. The technological improvements were what allowed them to try at all.

Note: In 1960, the GDP was \$543 billion. The total cost of Apollo from 1960 to 1973 was \$25.4 billion, about 4.5% of $543 billion. Of course, the money was spent over time and the 1960 GDP is just used as a reference. Originally I said "Over 8 years" because I was going from 1961 to 1969, but the budget accounting goes from 1960 to 1973.


"Technology (satellites, computer, ML) has come a long way during this time"

That may be true, but there's so much you can take with you! These modern rockets are nowhere close to the lifting capacity the Americans used. Which means you can't do the landing, stabilizing with so much brute force to keep your vehicle steady. The less weight is available to you, the less energy you can take with you. The less stabilizing you can do using sheer force, the more precise you have to control your landing. The more complicated routines needed, the more measurement and sensor technology to install. The more complex the programming, the more complex your systems, both hard- and software. The more complexity, the more can go wrong, etc.

Engineering with "unlimited" resources allowing you "a lot of energy" and "a lot of weight" allows the engineer to use more "raw" and "basic" solutions which is not comparable with high-tech engineering if you have only limited resources and limited power and limited weight-fuel.

To visualize this for those among you who are no engineers:

Imagine you have a water reservoir and the only task for the engineer and later operator is to either lock the reservoir or open it to drain the levels.

You basically just need one big switch: open or close.

Not much can go wrong there.

But now you add more tasks (only tasks to open up more or less without really interfering each other):

  • Opening % based on demand of farmers needs down the river
  • Open % based on demand of power production
  • Open % based on demand for fresh water down the river

And now you're going to add more tasks that actually interfere with each other: opening up more for farmers' needs, but the turbines produce more energy into the power grid, the energy into the power grid is based on voltage, current, and frequency, more water = higher frequency (a standard defines the boundaries between which frequencies it can variate). You're now getting to the maximum of that boundary, but farmers still need more water, so you have got to either limit the turbines or reroute water through different flows.

See what is going on here? The more complexity, the more switches needed, the more the engineer and operator have to consider.

And now I'm going to tell the engineer and operator additional tasks:

The original design the switch controlled a 1 meter radius water flow pipe with 1.5 cm stainless steel. But in the new design the engineer should limit the pipe to 0.3 meter and the thickness of the pipe to 0.4 cm and please below 350 kg per meter pipe. So steel is now out of the question, getting more complicated.

Also just opening one pipe for the operator might not do the trick. You need to handle a multitude of switches even for the same task based on the limits of weight, available flow, etc.

See how complex it is getting already with such an extreme simple comparison?

So just because the USA could do it with old technology and their extreme heavy lifting capacity doesn't mean you can copy it or do it easier just because you have more technology at hand, if at the same time you're going to do it with a lot more restrictions than US engineers had.

And the lifting weight for any space operation is the absolute most daunting restriction for any engineer in that field.

  • $\begingroup$ As an example of substituting brute force for complexity, consider the Ascent Propulsion System from the Apollo landers: it had to work, because there was no abort mode possible if it failed. It's the closest thing you can get to a "single-switch" rocket engine: no gimbals, no throttle, no ignition system, no pumps, just a pair of valves and pressurized tanks of fuel and oxidizer that ignite on contact. Efficiency was about 70% that of a more advanced design. $\endgroup$
    – Mark
    Commented Oct 4, 2019 at 22:35

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