Objectives
- Genre: Board
Difficulty
"Normal Game" was selected here. Unlike the "Hard Game", the AI in the "Normal Game" has a chance of not blocking a win in 1. This TAS plays the fastest game of Connect Four theoretically possible, both in terms of turn number and checker drop time, winning with just a simple four-checker sequence vertically.
Given that there is no faster game of Connect Four possible, any TAS of the "Easy Game" is therefore pointless, especially when considering that merely selecting the "Easy Game" loses more time to menuing.
RNG
If I didn't miss anything, I exhaustively tested every possible alternate RNG scenario that could have possibly led to a different, faster outcome than this TAS. This included:
- Testing different start frames of the game.
- Testing different start frames for each of the computer's turns.
- The computer is OK with stacking three checkers one column left of the middle, but will not make a stack in the column right of the middle, so testing this was a dead end.
Comparison to Existing TAS
Embedded above is an existing TAS of "Normal Game" by k80may from 8.5 years ago. It is 570 frames according to RTA timing (screen going black after selecting "One Player" to when the Connect Four™ text appears after winning). This submission is 574 frames using that same timing. That's 4 frames slower, so what gives? Well notice how k80may's video footage starts when "Connect Four" is selected from the menu, not from power on. We therefore have no idea when power on actually happened relative to this TAS, which I feel is very important to consider relative to the "submission must match or beat all known records" clause in the Movie Rules. I mean, k80may could have waited an hour on that title screen to get godly RNG for all we know, which would clearly NOT be faster relative to the TAS timing rules of this site. And as I mentioned earlier, I exhaustively tested everything to try and get this submission to match k80may's RTA time while still maximizing TAS time without luck. Therefore, while this submission does not beat all known records relative to RTA timing, it almost certainly beats all known records relative to TAS timing; thus I believe this submission is acceptable for the site.
nymx: Claiming for judging.
nymx: Ok...I had to bring this situation to our judges and was briefed on the rules that "used to be". Today, we have a different view on this. Basically, we can separate "Normal" and "Hard" into "Fastest Completion" and "Full Completion". This TAS would basically qualify for "Fastest Completion"...meaning that among the 3 difficulty selections, "Normal" beats the games as fast as possible.
I tested the "Easy" selection and you are completely correct. It takes way much longer and automatically default to "Normal", seeing that the fastest completion of the game (in terms of moves) wouldn't be any different. Additionally, I would like to add that the "other" TAS mentioned here isn't something that I am concerned with, because we have no way to scientifically verify (in this case a movie file) the speed of it. Since it wasn't published, you get the first stab at a publication.
Accepting as "Fastest Completion" for publication.
despoa: Processing...
nymx: setting to delayed, to examine against Tompa's work
nymx: What is so odd about this improvement, is the shift that was done to eliminate time. Not because routing made it faster, but how it changed RNG to get the CPU to make its move faster...thus, eliminating the time for "Black" to make its next move..demonstrated by Tompa. So this is not an obvious situation, as the original inputs by Winslinator.
I would have thought a new submission by Tompa would have been a better decision, but permission has been given for the addition of co-author. I'm not in agreement with this, but I will honor the situation put forward.
Publisher...please make a note about https://tasvideos.org/UserFiles/Info/638356713422978526, being an alternate version for fastest in an emulation setting, under BizHawk.
Accepting.
Publisher may take this on now.
nymx: Rearranged authorship as mentioned by Winslinator.
despoa: Processing...