I think the real reason is that you can not defeat a cavalry archer in a pre-powder age with anything else except of another cavalry archer. Another example of wunderwaffe of some ancient age is Alexander's sarissa which is also undefetable by anything except of even longer spear.
Fortifications, armored cavalry help as well. There are only so many fortresses you can ignore before your supply lines are cut. And only so many sieges you can upkeep at the same time, as a army focusing on strategic mobility at that, before you get out maneuvered by a smaller, bit more mobile, enemy force.
That being said, the Mongols failed to build a lasting empire in Europe / West of the Ural. They pretty much did build a lasting one between the Ural and, sometimes including, mainland China.
In the very large set of "this was particularly nasty" things that make up the early middle ages, this one really stands out. I have little doubt that this siege and subsequent sacking effectively set back humanity as a whole - and for the sole and simple reason that the caliph sent a snooty reply back to the Mongols.
The Mongols failed because the Khan died and the army had to retrain and after that the empire fractured. After that the 4 major part all deal with many local issues and also fight each other.
I really have to read up on the Mongols, and central asian history in general. That subject is terribly undercoverd in Western history. For somewhat understandable reasons, but still.
As an earlier reply said, fortresses and heavy armor did prove effective against the fast, lightly armored Mongolian archers.
The effectiveness of the Alexander's sarissa is as much a result of training and tactics as the weapon itself; I think it took years of training to turn a soldier into an effective sarissa user. Thus I personally would not consider it a wunderwaffe. A crossbow, to me, is a better example as it enabled yesterday's peasant to kill a trained soldier from a distance, without many years of training needed to master archery. My 2c.
> I think it took years of training to turn a soldier into an effective sarissa user.
AFAIK sarissa is a no-brainer without even ability of operating this weapon with one man. That allowed Alexander to take a lot of captured Persian men with no training under realy trained Greeks standing somewhere in more safe places of the formation. Miyamoto Musashi considered a spear as a king of weapons. But I am not a historian and all I told here is just an opinion, not a fact.
I agree that the feedback loop wouldn't make the models smarter than us, but training data would. I can see how with more computing power we will be able to feed it the whole of YouTube, Spotify and perhaps also millions of physics experiments.
With this data, perhaps a model will be able to understand the hidden patterns of the world (e.g. physics equations governing reality) better than us, don't you think?
If you have sufficiently many examples of this relationship (that is, input/output pairs) you can train a machine learning model with this data, and you will get (if successful) an explicit mathematical function which captures the relationship
No, only one. Think about it this way - if the warden randomly generates their numbers, each prisoner has a 1/100 chance of guessing it right. So best case scenario is exactly one gets it right, which is actually possible.
They just need one prisoner to guess correctly - then they all win.
Regarding the base case scenario - if each prisoner guesses randomly, each has a 1/100 chance of guessing correctly, and hence they have an expected number of exactly one correct guess. So no deterministic approach can give 2 correct guesses, since no prisoner actually has any real information about their number.
