Imagine that you believe that the purpose of your life is to be the best marathon runner in the world, or even just an above average marathon runner. Despite the fact that you are committed to the mission, imagine that running is not particularly enjoyable to you, and your training regime is tough, but being a better marathon runner than others is such a noble goal you persist. Everyday, you train, painstakingly learning new running techniques, building stamina, pushing your limits. When you are not training, you are thinking about running. You dream of marathon running in your bed, think of it in the bathroom, conjecture about it on the way to work. Even if its detrimental to your happiness (although there is a high correlation between your self-perceived notion of how good of a marathon runner you are and your happiness), marathon running, you claim, is even more important to you than what you give up to obtain it. You routinely make sacrifices in order to run faster, whether that be avoiding certain foods, or forgoing leisure activities, or spending hard earned money on training. In fact, marathon running is so important to you, that you would almost say that everything in your life: your job, your relationships, your possessions, are really just a tool to increase your ability to be a good marathon runner.

Now, imagine every time you run, you are forced to do so in a VR environment. You are running in real life, but you are wearing a headset which work to manipulate your sensory, auditory, and tactile processing. As a result you are unable to gain any meaningful information about the speed at which you are running the marathons, or if  you are making progress towards becoming a better runner or not. You are also not allowed to directly race with anyone else. Additionally, you are aware that many people are biased about how great of a marathon runner they they are. Many people who by consensus are slower marathon runners think they are great. Furthermore, even John Lipid, widely agreed to be the most overweight and least athletic man who ever lived (and whom rarely ran at all) claimed, and appeared to believe, that he was a great marathon runner in his day. In fact there is little ostensible correlation between self-perception of marathon running ability and real ability. Indeed, in your culture, there are many strong incentives to view oneself as a good marathon runner despite the reality of the manner.

Now, of course, you see that you seem to work a lot harder than others at marathon running, but your best might simply not be good enough. Afterall, how many of the greatest marathon runners tried more hard than their opponents, but simply lost to an individual who had more talent? How many people devoted their lives to practicing running techniques which later turned out to be ineffective or even counter productive? Indeed, you are deeply aware that many of the training techniques proposed for marathon running running are flawed and in fact, can reduce your overall speed, and even the very best techniques are very difficult to pull off, requiring very intricate techniques that are hard to understand.

Knowing all of this, why train to run at all?

Now, although relying on one's own intuition is perhaps flawed, it is possible to measure how good of a marathon runner you are, and your progress towards being one, by inviting people you know to observe you while you run. After all, it is generally regarded that large groups of people have an inherent ability to parse how good singular people are as marathon runners, but arranging such a poll is difficult work. Furthermore, they will only be willing to watch you for a few minutes to an hour of your marathons, giving them a flawed snapshot. Additionally, your running behavior while being observed by others may be substantially higher than when alone, which may not matter if you simply want to prove you are a great marathon runner when watched by others, but perhaps less so in proving you are a great marathon runner in aggregate.

If you reject the polling approach, how do you ensure you are not a John Lipid, who illogically thinks they are moral despite that not being the case? Since the only data we can really gain about morality comes from ourselves and the reactions of other humans, isn't polling the only way we can ever hope to access this information?

Edit: 

A couple of references which inspired this post:

Sorting Pebbles into Correct Heaps, by Eliezer Yudkowsky: https://www.lesswrong.com/posts/mMBTPTjRbsrqbSkZE/sorting-pebbles-into-correct-heaps

That Alien Message, by Eliezer Yudkowsky:

https://www.lesswrong.com/posts/5wMcKNAwB6X4mp9og/that-alien-message