Fire Pro picks March Madness... for the last time?!?

My personal favorite selection is Kane standing in for Tennessee which I didn't even notice until I was midway through simming the whole thing.

First things first, the first day of the tournament was highlighted by only the 9th time a 15 seed has ever defeated a 2 seed. Virtually no one predicted it... except Fire Pro World. As I've stated before the computers that landed man on the moon were likely less complex than the programs that run your average Fire Pro game (I recommend not actually checking that) so it should be a simple thing to predict a simple basketball tournament.

That said aside from that one standout the first round has been pretty rough for good ol' FPW so far, we are currently in the 7.6 percentile which is... bad. The biggest hit was Georgetown getting smoked as in the most pure Cinderella run I've seen the game toss my way in ages Mikey Whipwreck somehow found a way to keep winning matches he was clearly losing, defeating Kawada (a legit 1 seed) and Cody on the way to the Final Four. It was so darn appropriate it put a huge smile on my face.

Beyond that other pre-Final 4 standouts was 1 seed Inoki and 2 seed Vader demolishing their way through their region to resume their rivalry, this time Vader once again reigning supreme. The only 100% rated match in the tourney was in the first round and of course was the Lawrence Taylor match, who followed up that insanity by getting a 91% rating out of ultimate Warrior in round 2 in a losing effort.

The Final Four ended up being Vader vs Whipwreck and Hansen vs Braun. Mikey after his unlikely run lost by pinfall in under two minutes which is something I'm not sure I've seen before in a non-critical match. This lead to Vader vs hansen in the finals representing Iowa vs Arkansas and sadly no eyes were dislodged from anyone's skull. Also Big Van Vader won, which is notable as that exact same edit won my March Madness simulation last year. That's a twelve match tournament winning streak which is quite frankly absurd. Sentator I'm not saying your vader edit is overpowered, but it is definitely a terror.

So yeah, fun times and remember you heard here first that iowa is the next national champ.

That is in fact something people say, you successfully did a sports! 

So I'm sure I've said this before, but while this is a ridiculous exercise there is a decent idea behind it IMO. With no real basketball knowledge (heck, even with basketball knowledge) picking a bracket like this is an exercise in weighted probabilities. If for example there are four games in the opening round where a team has a 75% chance of winning, on average one of them will lose. While that is reasonable seeming there are also a number of games where a team is judged to have a 90% chance of winning, but while that means one out of every ten of them will result in a massive upset no one will ever pick the 10% team even if they know they should at some point.

Hence it basically becomes an issue of not just figuring out the rough odds, but trying to figure which unlikely choices to choose. I use what is considered the most reliable advanced statistic to see the relative strength of each team, but finding something that can represent that range while also reliably spouting out unlikely results at about the right frequency is a bit trickier. There are likely websites that let you type in pure numbers and get results like that, but that's boring. What I learned ages ago, and what I'm sure many here would agree with, is that Fire Pro has been doing just this for probably decades now. Better still at a low setting criticals are a great tool for simulating those lightning strike unforeseeable events that can dramatically changes things in an instant (ex. a star player suffering an injury).

The main issue is that the only thing I can really balance based off of is edit points, which is a flawed system: logic and how frequently edits go for pins or subs make a dramatic difference in edit strength and there's no real way to represent that numerically. Still it seems close enough that it could work, and I think it generally does work. This sim gave me a final four of a 2 seed, two 3 seeds and a 12 seed which given how the tourney has unfolded in real life (a 1 and a 2 seed have made it so far, an 11 and 12 seed were a game away) reads as reasonable if light on 1 seeds. The issue is, and that is probably unsolvable, is that even if it accurately figures out the relative odds and strengths in each game it is still picking the one out of four 25% likely teams (or one out of ten 10% likely teams) at more or less random. Or to put it another way, even if the game figures out the proper likelihood of a tier three team making a deep run into the tourney it is still picking that tier three team at random. 

This is a long way of saying that I ran out of teams in play today, have 320 total points with no more coming in which currently puts me in the 5.8 percentile which will only go lower and lower as the tourney wraps up. This can only be viewed as a bad result even if it spit out something that appeared at least initially reasonable. If this is the last one of these I run I think I at least have gotten to the point where I've taken it as far as it can reasonably go, and as far as that is it is still subject to a very high amount of randomness that will only give me a good result once every several years if not more.

But hey, I got to watch Vader powerbomb a lot of fools, bear witness to workrate god Lawrence Taylor, and got to watch Undertaker wrestle a blowup doll. If that ain't a moral victory I don't know what one is.