New Strats, Who Dis?

2023-11-05

I've been theorycrafting (and testing) optimizations for about a week now, and am reaping the benefits. I originally started posting runs around 7-8 minutes but later plateaued around 6.5 minutes using my initial strat with a few variations along the way. I just posted a new best at 4m 51s via recent optimizations and I'd say there's still room for improvement + basic practice to get all the aspects down to rote once settling on a strategy (which I haven't done yet).

A summary of the progress so far:

• Level 1: Bypass what was essentially an RNG element which could basically ruin runs. This was successful and also yielded time improvements via a somewhat unintuitive strat involving backtracking via the "hub" levels at the cost of 100 coins. This also happily cut out a large part of my original strat in level 1 which was awkward. Huge win here.
• Level 2: Essentially no change here besides entering the level in a slightly different place via the hub per the above changes.
• Level 3: Flow is very different. First, I don't "unlock" the large run of gold coins/potions at the top anymore because they're no longer needed due to other optimizations. Despite being level 3, this is the 2nd level visited which is hit immediately after getting the blue key in level 1. I beeline to the hub and backtrack to level 2 to get the green key there, before coming back to get the teal key and moving on.
• Level 4 & 5: These are largely unchanged beyond cutting the use of an invincibility potion in 5 (which I no longer have a spare of due to the changes in level 1/3). Makes the runs riskier re: dying before completion but I generally have enough energy and there are opportunities to pick up some extra along the way if needed with very little time loss.
• Level 6: Mostly unchanged except for an easy-win optimization that saves 200 coins which frankly I should have seen/done much earlier on.
• Level 7: Route through the level has changed a lot to eliminate part of the previous strat that involved waiting for a "green chaser" (ghost) to clear out some "cursed dirt" (moss) before I could proceed. Also, I'm picking up some extra coins/health to make up for losses elsewhere.
• Level 8: There have been quite a few changes here to the point that it's not really worth writing them out. Essentially, several optimizations have been made with the primary one being cutting the need for a whopping 400 coins to complete the level by doing a mostly-reliable strategy which involves body-blocking sentries to allow a green chaser to unlock the route instead of having to pay through it. I was aware of this possibility early on but wrote it off as too finicky, however after a fair amount of practice, I've got it to be quite reliable (successful well over 50% of the time I'd say). Cutting the need for those 400 coins really opened a lot of possibilities, some of which I imagine haven't dawned on me yet.

I'm much closer to the razor's edge in terms of breathing room on coins/energy, but I basically knew from the start that's what optimizing this game would come down to. The funny thing is, in my best run (as of writing), I finished with 0 energy. The game actually lets you go to 0 energy without dying, but if you take one more hit then you die. I actually screwed up part of the run and missed a terror potion (which would have saved me taking a good amount of damage), but was still able to barely complete it without dying to post a new best time.

My new (current) strat does seem fairly optimized outside of getting things down to rote. Obviously, there is a theoretical best time for the game. The game is mostly linear in the sense that you're basically just hitting a series of points as fast as possible. The shortest distance between two points is a line, etc. There's the monsters/coins/mechanics that complicate things, but not so much that there is a huge amount of depth to explore (as I see it). This goes back to my earlier thoughts about a 100% run being more interesting than an any% (which I'm doing now) because it would feature much more complexity. That continues to be a long-off consideration though.

Anyway, that's enough on that since I've got another topic to discuss...

New computer, who dis?

My new computer just showed up and I'm very happy with it so far. That's meant that I haven't been too focused on Drakklor this week as I was setting it up, coincidentally doing more real-life stuff than in a typical week, and then getting distracted by playing the new Talos Principle 2 game which is amazing (so far) - and the original is one of my all-time favourite games.

Anyway, back to some on-topic points re: the new computer vs. Drakklor:

• The game seems to emulate the same vs. on my old PC which is great news.
• I've confirmed a quirk which I noticed on my last PC: I've been running the game at 4000 cycles in DOSbox, but really what I've been doing is running it on auto (which seems to equate to 3000 cycles) and then upping to 4000 after the game starts. This seems to behave differently than just telling DOSbox to start it at 4000 cycles, for some reason - that ends up feeling much slower. I don't really care enough to dig into it more at this point since I'm getting a consistent experience, but it seems a little weird.
• Now that I have this new computer, I kind of have to follow through with my plans to stream preceded by all the mise-en-place needed to be able to do so.

That brings us to what is becoming a common section in these posts...

Next Steps

To keep the project advancing:

• Continue the optimize/iterate/practice feedback loop.
• Mise-en-place for streaming.

I actually think that's it. Short list. I'll probably get to it more once I stop being semi-distracted by Talos #2, but I'll try to at least make iterative process along the way.

Ciao!