HAL... WHAT A GREAT REPORTER...

The headline for game 1 vs. Giants tonight was:

Mel Ott hits 2 HRs in 14-7 loss to Stars XCVIII...

Yes... I know... we were the visitors...

BUT REALLY...

Josh Gibson had not 1, not 2, not 3, not 4... but 5 HRs and 11 RBIs...

WTF...

https://365.strat-o-matic.com/game/464130/236

Yeah, that's pretty pathetic. 5 HR's is a super rare event!

There's often a strange divergence between the outside headline and then the lead sentence in the actual boxscore. The lead (inside) sentence is usually always more accurate highlight-wise. Go figure, Hal!

Hack what were you a sportswriter or something?

These headlines are surely a product of rudimentary AI on the part of SOM, right? They're still in the Stone Age in this regard, I think. Even if it's not AI, it's not working too well. Sure, some of the game comments are cool, funny, if not cliched. But it's all we have.

Artificial....sure.
Intelligent....hmmmm

I used that Gibson card when it was first released…… it is 100% one of those elite home run cards with the kiner , Mathews , Ruth , Bonds type results.


https://365.strat-o-matic.com/player/80 ... 292/463574

I enjoy trying to predict what some of the quotes will be:

-"like a cool drink of water"
-"worked on my swing last night"
-"wish he played for our team"

HAL the Headline Writer seems obsessed with guys hitting 2 HR in a loss.That will be the headline almost every time it happens.

Then the stories seem to follow the standard format of a Baseball game report.

There's always a designated hero for the winning team--in this case, finally, Josh with his five HR. And the standard sports cliches are trotted out, including a clincher by the losing or winning manager or sometimes the winning pitcher.

One thing I'd like to see in HAL's reporting is a smidgen of playoff awarness. If a team wins game 1 of the semis, or even game 7 of the finals, it's always just reported like a regular game. That ought to be really easy to fix, if SOM ever wanted to get around to doing it.

That's funny stuff.
Yeah it seems to me there is a data file with canned phrases for different sections of the synopsis and a selection algorithm randomly inserts phrases into the recap template.

At least, that is how I would have programmed it 20 years ago.

These days, AI or NLP would be the way to go but I doubt that is being done. I mean...I think I know it.