[v. long post]
I posted this in the How Likely Was Italy to Form? Topic below, but it's right on topic for this one.
Like all questions regarding probability of history, this one requires us to consider whether, say, the (admittedly artificial) event "The Unification of Italy" constitutes the 'default' for probable outcomes.
This is a major philosophical problem.
In RL, there was no Unification of Italy event in history (at least, not in the same way as there was, this morning, the event Hochelaga Wakes Up or Hochelaga Drinks a Cup of Coffee). If history is linear (it may not be), then in RL there were in fact a series of tiny events that coalesced into what we call the Unification of Italy (e.g. Garibaldi Wakes Up, Cavour Drinks a Cup of Coffee). Without those tiny events, who knows? Maybe Garibaldi sleeps late, Cavour doesn't get his caffeine, and the Austrian generals take advantage. No Italy. In a linear history model, it's very hard to take the Unification of Italy (or any event) as a single, discreet event.
However, if we ignore all of that (I often do myself), then it might be possible to consider, abstractly, the likelihood of the event the Unification of Italy occurring in isolation. This requires setting up camp with either the view that:
(a) Italy unified in real life, real life is the default, therefore Italy was more likely to form than not (the Planned History Theory); or
(b) Italy unified in real life, but considering the multitude of possibilities Italian unification in real life was a fluke and should not stand in as the default (the Coincidence Theory); or
(c) Italy never unified in real life, because I live inside my copy of Victoria and it is those things on the outside that are artificial, therefore whatever happens in Victoria is exactly real and true (the Nutcracker or WAD Theory).
Leaving aside possibility (c) (which to many -- including myself -- is the most attractive but least likely), we are left with the Planned History Theory and the Coincidence Theory.
Planned History Theory
Real life is the model for the simulation. Italy unified in real life. Therefore, it should be more likely than not that Italy unifies in the simulation.
Perfectly acceptable theory. Totally ignores the minute complexities of linear history, of course, but that is necessary to keep the Victoria game files down below 8 million terabytes+ in size. In my view, the best option for game play, because I think on the whole player's expectations are that things *like* what happened in RL should happen, with some (occasionally radical) variation. In the particular example of Italian unification, I think it would be acceptable if:
In 50% of games, Italy unified as it did or in a like manner as in RL;
In 25% of games, Italy partly unified or unified in a manner unlike RL; and
In 25% of games, Italy did not unify.
That would be a huge improvement over the present, which is:
In 100% of games, Italy did not unify (correct me, please, if anyone has seen an AI unify Italy in any of the millions of games this community has played).
Coincidence Theory
Real life is not the model for the simulation. Chaos theory and complex mathematics are the model for the simulation. Italy unified in RL because of changing weather patterns in Taipei.
Also a perfectly acceptable theory. When you consider the myriad of possibilities (as we are required to do every time we play!), we realise that ANY event, however discrete, is unlikely. A complex series of events culminating in the event the Unification of Italy is extraordinarily unlikely. Impossible even.
If in 1855 someone told you that within 15 years the fragmented nation states of the Italian peninsula would, with some exception, form one country, you would have (be honest) offered odds of 1000-1, or at least odds to 1 corresponding to the number of states... 14-1? 15-1? It's a longshot. You wouldn't bet on it.
In this model, proponents would like the following outcomes:
7% Italy unifies as it did in RL;
3/100% Italy unifies as it did in RL except for Parma;
1/100% Italy does not unify as it did in RL and Garibaldi's armies turn on the Ottomans... etc.
Probably, in this model, the favourite to win would still be the real life outcome because of our penchant for re-creating history (with often minor differences, occasionally major differences, rarely radical differences).
But note that even under this model:
In 100% of games, Italy did not unify.
... is still garbage.
Middle Ground
The middle ground for any event in the Victoria simulation seems to me to be something like:
-- A little more than half the time, events work out as they did in RL. (55-60%)
-- Often, events work out slightly differently than they did in RL. (20-25%)
-- Occasionally, events work out significantly differently than they did in RL. (10-15%)
-- Rarely, events work out radically differently than they did in RL. (5-10%)
Admittedly, this is a Planned History model. The problem with using a Coincidence theory is that is that the simulation cannot possibly calculate all the possibilities, which means that we end up with artificialities (which thankfully, we can overlook -- I mean, if Italy unifies while Garibaldi is still making candles in Manhattan, who cares?).
So those are my thoughts.
Hoch
