Yes the "Art of Programming" Don Knuth.
 It was a simple experiment at a conference, first half the console sat there, play either game, mid conference announced that we were tracking hours played on each game and compare second half ratio to first half ratio.
But this one takes the cake as the bugs at the end really made for an incredible story.
Filed under 'things I wish I hadn't searched for.' WTF.
You can check out some speedruns of the game from Monopoli to see a pretty impressive number of glitches. Here's one from AGDQ 2014: http://www.youtube.com/watch?v=po9HS2lIZRw .
But if people really wanted to cheat, they would standby or mod anyway.
It reminds me of being a kid and trying to build a city in SimCity without roads.
(related HN discussion: https://news.ycombinator.com/item?id=1352864)
I thought you were supposed to replace all roads with rails? :)
Well, 262 % 256 = 6. So if you have very old scoring code that expects the score to fit in an unsigned byte, it's going to think the score is 6.
So that explains the 6 in the ones column. The 6 in the tens column, well, that I can't explain.
It's a bit ironic that addition is kinda the quintessential thing computers get wrong. Admittedly, not much of an issue these days - but the comment was amusing :)
I once made a similar mistake. Someone submitted a video to my film festival and it was interlaced the wrong way. I swapped the fields and didn't do a full preview. Half way through the submitted video had swapped field order and I'd not noticed.
After the screening I explained the error and apologized. She hadn't even noticed. Some people just can't see.
Another example - a single field glitch.
me : what's that glitch?
person : what glitch?
me (touching the screen next to the glicth) : here
person : what glitch?
I just couldn't get them to see it.
They just didn't expect anyone to play against the Dallas Cowboys. Rookie mistake.
And the Innovative Title of the Year award goes to...
Any given play generally takes 10-20 seconds, and you're going to be running, bare minimum, 2 plays per 8 points(1 for touchdown, 1 for 2 point conversion). So that's about 25 seconds per score.
If the teams are remotely well matched(i.e. aren't a Breaking Madden team), they'll take at least 2-3 minutes per possession, since you're most likely going to have 4 downs, which have a play clock of 40 seconds(most teams use around 30/play), so the "average" shortest possession is something like 2 minutes.
The point of all this? The Breaking Madden games are really an outlier that you'd have to specifically code for.
Rather than coding special exemptions or rules, it seems that using 8 bit storage would have to be a deliberate choice on any modern platform.
The real question is, what the hell does a Madden developer do for the rest of the time that dealing with data type changes is not worth it?
Who wouldn't just use "int" for all of this logic? It's so simple that you could do it on autopilot. You'd have to specifically try to limit to 8-bits.
While it was annoying at times to be so worried about bits and cycles, it was also a very rewarding type of puzzle, figuring out how to squeeze just a little more performance out of the limited system we had to work with.
The answer from EA's point of view is that people buy it every year, so of course they'll make a new one.
The explanation from the user's point of view is that much of the magic of the game comes from it reflecting the teams, players, and plays of the actual NFL. When some quarterback gets traded to another team, you want to get next year's Madden so he's on the right team in the game too.
This is one of the really cool things about the game. There are millions of people who have detailed knowledge of different teams and players and what their strengths and weaknesses are. Madden lets you take that otherwise pointless barroom trivia and use it to actually be better at the game.
Reminds me of the best one: https://www.youtube.com/watch?v=4Zs4yxkRI6U
Mathematical abomination, indeed.