The difference between a 20:47 and a 19:59 is precisely the difference between someone's 5K and 10K race performance. So if you're a 17 minute 5K runner imagine trying to run a 34 minute 10K. Or if you're a 15 minute 5K-er try running a sub-30 10K. Because that is precisely what we're talkinga bout in terms of race equivalencies.
So for non-runners a 20:47 might sound really close for anyone who actually understands running it basically means he needs to run twice as far as his current 5K time before he could have a chance at breaking 20 for the 5K.
And that is if he really could run a 20:47 right now which I don't think he can. Talk is cheap so the whole argument is moot.
As for the other comments on improvements and newbie gains. I know it's anecdotal and personal but I did go from running 24:04 in January of 2010 as a new runner who just started running to running 19:05 by May and running a 37:11 10K in October (17:52 equivalent and 2:51:35 marathon equivalent) My marathon time at that time was a 3:05 so I was 15 minutes off my marathon equivalent from my 10K. I didn't break 2:50 in the marathon for 2 more years and by then I was running 16:40 for the 5K and my marathon was still 5 minutes off on the equivalency table (2:44 vs. 2:39).
So yes, newbie gains can be huge and yes, marathon times take years to catch up to the shorter distance equivalencies and require a ton of hard training and miles to build the required endurance.
Mike Rossi has already got his newbie gains when he went from a 30 minute 5K runner to a 22 minute 5K runner. For him to drop another 2 minutes is like me trying to get to sub-16. Never going to happen. And I'd be surprised if he could come within 10 minutes of the marathon equivalency of his 5K time ever.
So in a sense I'd say the 5K offer is the most generous one and the most attainable for him but it's still pretty much impossible.