Ok so Recently Niftski got 55:52 super fast because NIFTSKI IS A THICC BOI Then today I got 55:48 because HOLY P NIFTSKI IS A GENIUS WITH THAT 8-1 STRAT. At first, I was confused how I could be 4 seconds faster, as my worst run was a 5:21, while his worst run was a 5:15. That's a big deal when the median/average run is a sub 5. So I did the math, and when I counted up all of Niftski's times which do not include the time from when he hit the axe in 8-4 up until he entered the level of 1-1, he executed a 55:16.655.

When I counted up my own times, I found that I got a 55:16.366, a whopping 0.29 seconds faster than Niftski in execution. So where did the other 3.7 seconds come from?

From me consistently resetting faster than Niftski

As this category is approaching high levels, where saving just seconds in this 55 minute speedrun matters, getting 1 more sub 5 than before could make or break a potential WR, how good you are at resetting will make a huge difference It took Niftski on average 3.6 seconds to get from hitting the axe in 8-4 to the beginning of 1-1, while it took me an average of 3.22 seconds (uh idk how it's 3.22 seconds, seems strangely fast to me but idk what I could have done wrong in math) This difference is basically 1 framerule, and the math really shows how consistently saving 1 framerule is going to make a large difference over the course of the run.

I've only talked about 2 people who play on emulator GTAce99 is 3rd place at the current time of 6:26 pm PST April 28th 2020, and from his description, he has a run time and a real time Subtract the real time from the run time, divide that number by 10, and he's spending about 3.75 seconds per run in between hitting the 8-4 axe and entering 1-1, which is even slower than Niftski, and a whopping 0.5 seconds slower than my resets

I think it would be fair to take out the time in between hitting the 8-4 axe and when the reset button is pressed (if you lose a framerule because you pressed start late on title screen then that's an execution error thicc boi). yes YES NIFTSKI

I doubt we're timing 10 reset times and 11 individual run times...

Wow. Such a short answer for a long post.

does it need to be longer? I doubt we're going to remove reset times. they are variable between every run so there's no efficient way to do that

this is a category extension, not a main category. if someone wants to run a leaderboard with reset times removed, they are more than welcome to. But I seriously doubt the mods will want to do this.

It's pretty standard in multiple games/runs to include reset times. Learn to optimize that as well.

length of a post doesn't have to correlate with length of reply, FYI

Also please read Change 2 here:

https://www.speedrun.com/smbce/thread/kof23

We will be dropping millseconds from the times for some categories and implementing thresholds for the others where milliseconds will be used. This leaderboard is not intended to be taken as seriously as the main leaderboard, and as such, does not necessarily warrant framecounting a very high percentage of the runs that are submitted.

I think we can at least discuss it. I get it from the runner's end.

It would be painful to remove the resets from the times admittedly. IF we happened to go through with this, then what would the upper threshold for resets? We would need to implement one, so I'm open to suggestions. If we didn't establish one, then someone is going to do a run, reset and then walk away for 5 minutes and expect us to remove all that time from the run.

if that's the case, i'd propose we only need to framecount if a WR was in doubt, or maybe a top 3 time. I don't think anything beyond that needs a framecount/reset removal, and we just leave the reset time in (also incentivizes not doing long resets). We can change the rules to say that's the case for non-top 3 times or something.

I also agree on upper threshold, sitting for five (or however many long) minutes also can give an unfair advantage, what if I just left my stream on for 11 days and did 1 run per day to feel fresh each time (extreme dumb example, but yeah).

Also, how do we count reset time, are you able to wait 8 seconds on the title screen to fix the RNG?

I know in lost levels we only remove loads times, so title screen time still counts to the run time. That's why I've optimized my resets there for New Game runs. The loads are constant time each run (so you only need to calulate one run worth of loads), and i'm mashing on the title screen.

I think one of the easiest ways to make it more fair for the mods would be to time each individual run, and then add up all those times, vs frame counting all the time spent resetting up Although when I did that it took me like an hour, not sure how big of a deal it is for moderators to retime runs in this method At that point then we could decide whether waiting on the title screen on runs past the 1st run should count as time loss or not

^^^

Okey, I think also the best idea is to just count the time of the runs without the intersections into one run another, but with a max of time Maybe like a máximum of 20 seconds into both runs or something like that

@KilleDragon

There is still no data that suggests you run this game. Why do you disagree with every potential change? And why is it that whenever you are called out you completely disregard the reply to you and continue liking and responding to every forum post? Like Roopert said before, you are free to express your own opinions, but you have no runs of this game on the site and it makes zero sense.

One problem with counting run time is it would be harder to use hammers - since knowing the actual "delay" used would be much more unclear. That leads you to frame count but that adds a range of error, especially when 30fps or 24fps, which would multiply over 11 runs. You could eliminate the blank space. Which would work. But also has an an element of error, especially when multiplied 11 times. Granted we're only talking about a small error.

Either of those methods works, it just adds a lot of complexity, and subtracts accuracy (at least a little)

I get the concern, but I see negatives along with the possible solutions to the proposed problem

By far the simplest thing is just to consider restart efficiency as part of the category, as @Tecate said.

Just my thoughts.

If this isn't too time consuming for the mods (disregard if it is), one suggestion is rounding or flooring reset times to the nearest second - this'll fix some issues, and make things slightly more fair for the runners who reset slightly slower, while preventing them from waiting an extra few seconds for better patterns, or from walking away for 5 minutes like mentioned above. idk if this has issues on its own, i havent thought it that htorughly through yet, but just a dsugesttions

@eddiecatgaming . _ . what the f*ck i say of opinion I say is a good idea to just count the time you are playing and no the time of the reset And is more, im agree with that "potential change" Which is the problem

But okey, i understand i have to record something to have any sense here :)