So, as you will see if you click the previous reply, I modified my probability seven times before reaching 97% on September 10. But what's most interesting is that I was at 42% on September 3. An awful lot changed in one week!

So I thought I'd write a retrospective: why did I end up on the wrong side of 50%?

The biggest factors were the fog of war, and my utter lack of Ukrainian and Russian language skills. These factors mean I tend to have to rely on second-hand, third-hand and fourth-hand information. But I became frustrated with the bias of most commentators, who were clearly a bit too optimistic for Ukraine (though it turns out that much of that optimism was warranted).

In my search for balanced reporting, the best source I was able to find was, strangely, an anonymous YouTuber called "War In Ukraine", This person did sometimes post misinformation, e.g. he thought Russia had destroyed a third of M777 artillery, which was a Russian fabrication: a Russian source reported that a Ukrainian minister had said a third of M777 artillery was lost in "battles" with Russia, but in reality what happened was that a third of the M777s had to be sent for repair due to overheating or overuse. They were breaking because they were popular! So, he's not a perfect source, but certainly a more balanced source than, for example, Trent Telenko who, as I recall, predicted Russia would run out of steam at the end of May. Trent also, perhaps correctly, said that Ukraine has access to software and even hardware that could let it do rapid and accurate counterbattery fire — but then he assumed Ukraine has the capability and is widely using it, which is two steps too far. Plus he shocked me by mistaking a CGI video for a real video. I immediately realized it was fake when I watched it. No military expert should have been fooled, but apparently Trent is only an expert in military hardware rather than military tactics, "how CGI looks", etc.

History Legends on YouTube has also proven to be a balanced source, but it took me months to decide whether he was a good source, and while he is a pretty good source, he seems no better than me at peering through the fog of war and only passable when it comes to predicting the future.

These sources understood that Ukraine could win but were not particularly optimistic for Ukraine. Notably, War In Ukraine (WUI) often suggested that although Ukrainian troops are highly motivated, mid-level commanders and top politicians are not very competent, and that among top politicians, there might be some secretly pro-Kremlin people. I, hearing this, reasoned that if it were true, typical pro-Ukrainian sources wouldn't be talking about it and pro-Kremlin sources largely wouldn't be trustworthy, and so I didn't have any source I could trust to counter WIU's assessment. Similarly, WIU cited some "bad" Ukraine government policies that were hurting Ukraine's economy, such as a 70% rail tarriff. I was able to verify that the tarriff existed, but I couldn't assess whether the policy was good or bad because no one was talking about it in English except one business group complaining that the policy hurt them.

You'll notice I say "pro-Kremlin" and not "pro-Russian" as most people do. I have always been pro-Russian but anti-Kremlin, and since February 24 I am more pro-Ukrainian than pro-Russian.

So, even though I knew that there were many factors working in Ukraine's favor, such as:

- Russian soldiers have lower morale,

- Russian military is corrupt and so uses resources inefficiently

- Russian commanders usually care little about their soldiers

- Ukrainian soldiers care dramatically more than Russian soldiers about keeping Ukrainian land

- Western aid gives Ukraine technological advantages in ATGMs, accurate artillery and intelligence-gathering

My thinking was that while these advantages should be enough to win the war eventually, Russia had some big advantages that could be sufficient to let them hold the territory they already have until next May:

1. Russia has more than three times the population of Ukraine, and Putin has been willing to send almost all available military forces to Ukraine, leaving Russia's borders undefended. Thus, Putin could achieve manpower parity with Ukraine by mobilizing one-third the fraction of the population that Ukraine mobilized. I did NOT think Putin would actually do this, and he didn't, but...

2. Attacking is widely considered much harder than defending; a common rule of thumb is that an attacking force needs a 3:1 advantage in numbers to win. Maybe 2:1 would be enough for Ukraine to attack, but then maybe Putin could successfully defend by mobilizing soldiers at only one-fifth the rate Ukraine did. Since Putin was pulling out all the stops short of general mobilization, it seemed possible that he already did just that. (Note that Russia got around the 3:1 rule using their "moon landscape" strategy of utterly destroying towns to kill Ukrainian forces and destroy their cover; and note that Ukraine cannot use this strategy.)

3. Russia's military production infrastructure is intact and mobilized, while Ukraine's is largely destroyed and they ran out of 152mm ammo for their Soviet artillery. Sanctions limit Russia's production of advanced weapons, but they apparently have stockpiles of millions of shells from the Soviet era, and thousands of artillery pieces. Thus, as long as Putin is willing to spend Russia's entire conventional weapon supply on this one war (which he is, I think), Russia retains a strong numerical weapons advantage (though luckily the logistics infrastructure on their weapons is physically inferior to NATO weapon logistics, which wastes some of Russia's limited manpower). Ukraine overcame their more limited weapons supply by getting superior ATGMs, more accurate artillery and more intelligence, but the big risk was that the West would get bored and reduce support for Ukraine due to a lack of public attention and continuing memes from tankies about how Russia would surely win in the end. Ukraine was therefore likely to have their crucial weapon supply reduced if they didn't achieve a significant victory, which, until September, they hadn't managed to do.

And I was concerned about Ukraine not having counterattacked until the end of August and interpreted it as a sign of weakness, especially because I was unable to find information about how many new soldiers Ukraine had trained. I was concerned about the Kherson offensive only using 30,000 Ukrainian soldiers (as I heard Russia moved 30,000 of their own soldiers to the area) and I was worried that, like Russia, Ukraine had no more to spare. The new Kharkiv offensive explains what happened: Ukraine did had more soldiers available but allocated them to a secret offensive.

The collapse of Russian forces at Kharkiv conveyed a lot of information all at once: it proved that there actually are competent Ukrainian commanders; it showed that the Kremlin hollowed out the defenses in that area to help defend Kherson, so they have a severe manpower shortage. And it showed that Ukraine had enough spare military hardware and manpower and to do a blitzkrieg.

And then, of course, because the Russians collapsed so quickly, they delivered what Ukraine needs most: more weapons. Whatever artillery Ukraine captured will be pretty useless, but they needed more tanks/AFVs, and now they have more.

The thing is, while obviously I ended up a bit too pessimistic about Ukraine's chances, I'm not sure what I should have done differently to avoid that. The only way I could really be sure Russia had a severe manpower shortage — and that Ukraine didn't — was an event like the collapse near Kharkiv.

But I do think I lost my nerve too much. I should've seen back in August that while Ukraine was moving slowly, the wind was still clearly blowing in their direction. It was reasonable to think in August that Russia still had enough troops for the time being, but not that they would still have enough next March.

Anyway, now the emperor is naked, so to speak. Europe can now see what I've known all along: that Ukraine is likely to win, and maybe even win quickly. And that greatly reduces the chance that their support for Ukraine will waver.

Given the recent news of some politicians in Russia publicly calling for Putin to be removed from office, it seems that Putin is in an even weaker position than I thought in my September 9 update.

Putin could've taken his offramp in March and moved all forces near Donbass. A concentrated force like that might have been able to hold out a long time. But once again he overextended his demoralized and corrupt military, and so he is at risk of losing the entire Donbas — and with it, his job as president of Russia. But Putin cannot leave power because he would be at risk of being investigated — his crimes even before the 2022 invasion should be enough to put him in jail for life. He may be able to work around this problem, e.g. by choosing a successor who aided his crimes and will therefore protect him from investigation.

Still, Putin seems to see himself as some kind of embodiment of Russia, and doesn't have friends so much as allies. So he will do his best to hold onto power, and I give an 80% chance he will still be in power in May of next year.