Yeah, wish the game's goals weren't so unclear and buggy. I'm lucky it worked for me.

I finally got the Freeze Frame achievement after my 12th playthrough thanks to you and following this guide exactly (and getting really lucky). But I failed the 234 trophy by about 10 screenwraps. I’m still not done with this game. 😞 Edit: I did it! 14 playthroughs in all. What a waste of a weekend.

Maybe? I did have computer assistance for the route, but in 12 years it's certainly possible somebody got lucky enough with powerups to beat it.

I was 28 points away from this.

Wish I still had my memory card. Out of curiosity. Damn good run, and most of all, thank you for that nostalgia hit. Such a criminally underrated title.

It's been a very long time, but I don't think I ever played it online.

Wow, that's a strange request. :) It might be on some backup CD somewhere, but I don't know. I don't actually remember making videos of SMB scores - just my Master Levels run.

@@SnapDragon128 If you wouldn't mind trying to find it for me I'd really appreciate that, along with any other old SMB videos you might still have from back then. Sorry for the delayed reply. I'm sure I replied to you earlier but it seems as though my comment never went through.

Much older than Tourist. :)

@@SnapDragon128 Nice man and I have to say I was trying to solve Polygonal Puzzle (ICPC WF 2016) and I couldn't find any heuristic to make the O(N^6) get AC. So I found your code and I was inspired by the "i/j hint" heuristic and it worked lol. Pretty insane problem btw lol. Thanks.

@@furo.v Heh, nice! Glad my code could help. I'm getting pretty old and crusty these days, but I'll keep posting my WF solutions for as long as I'm a judge.

Yes, I'm SnapDragon.

It's not really in a state that other people could use. But I did repurpose it to help some of the top players of the more-recent PS3 Amplitude game (like TUSMBOX). Since the powerups are fixed in that game, you can compute the single most optimal route.

@@SnapDragon128 oh ye. i have a spreadsheet that explains where each multiplier is, the best route known in the song, and the top 10 players in the game

Crazy to see someone else remembers this game

If you look closely, I actually have at most 102 nodes in my graph. In fact, I was actually originally intending to use y*X+x as one set of nodes and 100+y*X+x as the other, but I buggily forgot to do that. However, it doesn't matter because the two sets of (x,y) are disjoint. I wouldn't know the current contest "meta", but through my entire career I have only ever used the simple DFS-based Ford-Fulkerson (which is so easy to code I don't even have it precoded) and Bellman Ford for min-cost max-flow. In my experience, flow problems do not tend to have extremely tight time limits. However, it wouldn't HURT to have a more sophisticated flow algorithm precoded just in case you ever need it. And yes, the Dijkstra-based flow was simply wrong, because of negative edges. I'm so out of practice I didn't even remember that until after the screencast ended. Sorry! :)

Thanks for taking the time to answer my questions. I didn't read your code correctly the first time. Indeed only 102 nodes are used which works since the graph is bipartite. I'm indeed under the impression that for flow the time limits are rarely a problem, and that it's hard to predict which flow algorithm will be faster in practice. Good to know that it has worked fine for you so far without using the more exotic flow algorithms. Good luck in round 3!

They really are. It's so unfair! ;)

watch ecernwala. It will be like keyboard ASMR

Hah, yeah, it did turn out to be quite a short solution. :) If I hadn't gotten so close just with simple heuristics, I'm sure I'd have needed to rethink things and stumble towards the "correct" DP way.

Ah, I guess I didn't explain it while coding. nh, nm are a simulated "guess" at the number of times the hour/minute hands have completely circled around, so I know exactly how far they have traveled, and can subtract / divide by 11 to get a possible exact time.

Thank you for the feedback! I clearly need the practice anyway, so I'll try to at least do round 1A soon. There's room enough on youtube for both errichto and me. :) (Also, good luck on your round 1C!)

Glad to hear it! :)

Hmm, let me clarify. Since the "hard" problems on a Qualification Round are in some sense wasted (nobody gains anything by solving them), they typically have some other reason for being there: - They might be a tutorial/warning about something that'll show up later (BigIntegers, say, or interactive problems). Median Sort was in this category. - They might be interesting ideas that the writers liked, but for some reason they wouldn't make for good problems in a real competition. At the ICPC Finals my "Towers of Powers" problem ended up like this - we all thought it was neat, but arguably competitors would have no way to know what a valid solution would require, so we put it on a practice round instead. However, I wouldn't think Cheating Detection would be in this category; it seems like a fine contest problem, albeit a bit mathy and requiring some intuition (which I failed at).

Good idea! I added the timestamps to this video. The "tutorials" will unfortunately have to be brief for the rest of the contests since there will be time pressure. :)

@@SnapDragon128 Of course. That goes without saying :). In fact, you don't even have to explain during the contest. Would it be possible to explain the solutions after the contest is over, like the second half of the video being dedicated to explaining the code, your approach and the solution ? That would be a win-win! :)

Sure, if you can get it to work. My personal contest library is at github.com/SnapDragon64/ContestLibrary and the file you want is process.cc.