I am not sure Nintendo has to justify themself about adjusting stock or not. If UK was severly under supplied, I would expect other country to be just fine without too much stock overall. I had a different theory not mentioned here :

If they had announced 20 or 20.5, what would have been then the FY2021 forecast ?...This would have not allowed them a lot of margin, they would have announced more ?... Or a bit less ? admitting that FY2021 would be the decline year of the Switch ? Starting from 19,5M, they can then announce a safe 20M to show that they still forecast progression for the switch. Capitalism is just about "always more, more and MORE..."

Once Nintendo announces a weaker forecast on year N compared to year N-1, the stock could collapse quickly.

2000 games includes digital only titles which are not included in Nintendo's figures.

You keep acting like those numbers are hard to achieve when they are actually too low compared to what the switch will sell.

On those two points:

1. Only games that have either both a retail and digital version or are exclusively available at retail contribute to the software total, so it's not anywhere close to 2,000 titles. Nintendo's financials list the number of games per region; Japan has 412, Americas 543 and Europe 549. The majority of those title counts overlaps because most games are available worldwide, so the number of unique titles is probably ~600. Third parties are unsurprisingly not going to see big sales when they refrain from investing; a large number of their games are ports of old titles. Nintendo has stated that five third party titles have shipped 1m+ units during the nine months of the current fiscal year and people are trying to figure out which ones. That's the kind of exercise that makes you realize how few big third party games have been released for Switch, because it's hard to think of 2019 releases that have had a realistic chance to sell 1m+.

2.  With Sword/Shield going the expansion route instead of a third version or something similar to it, it's safe to say that it will have better legs than most Pokémon games. 25m lifetime is certainly in the cards.

No I just try to bring the view as (very small) investor. The thing with the forecast is not to predict the reality but to impress investors to attract more cash. If they would announce 20 today, it would make them in a delicate position to announce a risky 21. No one can predict what impact will have the PS5 and series X, the only one similar example of under-powered console fighting against bruteforce consoles is the Wii against the 360 and PS3, and we have seen the collapse of the sales. I see people believing into a "counter offensive of power" with the Switch Pro...This is ridiculous, at best this new version will just reduce a little bit the huge gap with the base XBOX1...I think they can't hold on then after winter 2020-2021. So they won't sell 23 and then 25 M on the 2 following fiscal years.

2000 games includes digital only titles which are not included in Nintendo's figures.

Your theory is 100% nonsensical and here's why:

It doesn't matter what Nintendo forecasts because hardware shipments will ultimately judged by actual performance. For example, if Switch managed 20.5m this fiscal year, it wouldn't matter at all if Nintendo forecasted 19.5m, 20.0m or 20.5m, because the actual performance of 20.5m would be the measuring stick going forward in any case.

Secondly, what the stock market is really looking at is the bottom line of a company, so it's about expected profits/losses. Even a company that makes a loss of $1 billion can see their stock rise when their forecast was a loss of $1.5 billion, because they performed above expectations. Now as an example, if Nintendo finished this fiscal year with 19.5m and then forecasted 19.0m for the following fiscal year, that's not necessary going to hurt the stock; because what is important is their profit forecast and that might very well be higher despite lower hardware shipments. That's because software is more profitable and software sales continue to grow even when the hardware has already reached a plateau. In short, you are too obsessed with hardware shipments when it comes to stock price.