YoshiRulz
Any
Editor, Emulator Coder
Joined: 8/30/2020
Posts: 106
Location: Sydney, Australia
How has no-one mentioned this Flash classic yet? The OG AG TD. Some initial thoughts:
  • The game runs smoothly in Ruffle. (And I really hope lowering the graphics quality isn't optimal, because it looks remarkably good on max. on a modern machine.)
  • The official strategy guide is still up and serves as a good refresher for those of us who haven't played in a while.
  • On the subject of Armor Games, there's a domain-gated feature, the third skill. Have we figured out domain spoofing yet?
  • I'm imagining that a TAS would want to visit each 'battle' only once. In my RTA test, I got halfway to the quotas for the first few battles, so it's not too much of a stretch. Though I don't know whether the game will force you to backtrack at some point since I've never actually beaten it.
  • There doesn't seem to be a speedrunning community for this version. I saw a few runs of later releases on SRC, but the mechanics are probably too different to be applicable. And the optimal strategy for RTA may not be optimal for TAS anyway.
  • You get skill points by levelling up, and levels are just score milestones. It seems you can pass multiple milestones with a single 'battle'.
  • There are no restrictions on respec'ing; with one click you reset every skill and get all your points back.
I contribute to BizHawk as Linux/cross-platform lead, testing and automation lead, and UI designer. This year, I'm experimenting with streaming BizHawk development on Twitch. nope Links to find me elsewhere and to some of my side projects are on my personal site. I will respond on Discord faster than to PMs on this site.
Hey look buddy, I'm an engineer. That means I solve problems. Not problems like "What is software," because that would fall within the purview of your conundrums of philosophy. I solve practical problems. For instance, how am I gonna stop some high-wattage thread-ripping monster of a CPU dead in its tracks? The answer: use code. And if that don't work? Use more code.
YoshiRulz
Any
Editor, Emulator Coder
Joined: 8/30/2020
Posts: 106
Location: Sydney, Australia
  • I see now why I never finished this as a kid. Around 30% of the way through, it becomes quite hard (as a casual player), but then around 40% through you unlock the "Number of high grade starter gems" skill and it becomes almost too easy. The last ~20 battles I was doing other things while playing, and fast-fowarding for most of the battle. It was really laggy.
  • Minor spoiler for final boss: You need a LOT of purple. After I won, my total score was just shy of 268 million.
  • I posited the TAS would probably visit each 'battle' only once, but now I'm doubting whether that's possible; maybe by carefully choosing the play order. I'm mostly concerned about the secret stages. I unlocked the first about as early as I could, but didn't manage to beat it without the high-grade starter gem skill.
  • As I noted above, you can re-spec at any time, but the full wipe gets annoying very quickly. But TAS can click as fast as possible, so I don't suspect it will take much time in a run, and it can be combined with naturally spec'ing to reduce load times.
  • General strategy advice:
    • Your main bottleneck will always be fire rate, that is, socketing 2 gems will result in a higher DPS than combining them first. Nevermind the "...grade 7 or higher" counter on the results screen. It's bait.
    • Pure gems are just better (due to increased fire rate, and also range). Again, I think the dual and triple mastery skills are bait.
    • Gem bombs are rarely worth it as an attack. Both in terms of mana cost, assuming you don't have junk gems and have to conjure one—the banishment cost is usually low—and in terms of being 1-time-use while socketed gems are reusable.
  • One mechanic I forgot to mention in the OP is the 'amulets' awarded for repeating actions a certain number of times in a battle. These are a pain to get casually, and don't award many points, but the TASer should keep track during a battle and, for example, build an extra tower during the last wave to reach the next threshold. The combo one should be free for TAS. For the record, the requirements are:
    • Drop gem bombs on monsters: 10k pts. for 15, 15k pts. for 30, 20k pts. for 75
    • Conjure gems: 15k pts. for 25, 22.5k pts. for 50, 30k pts. for 125
    • Combine gems: 20k pts. for 15, 30k pts. for 30, ??? for 75
    • Create grade 7+ gems: 40k pts. for 2, 60k pts. for 4, 80k pts. for 7
    • Build towers: 20k pts. for 3, 30k pts. for 6, 40k pts. for 9
    • Combo: 15k pts. for 30, 22.5k pts. for 60, ??? for 150
  • I also neglected to mention the 12 domain-gated amulets. I don't know what they are, but given they're under the "Journey Amulets" heading, I'm guessing they'd be for cumulative/lifetime stats, and therefore shouldn't affect strategy. (edit: Looking at the submission guidelines, apparently Ruffle has a --spoof-url flag.)
  • The epic battle amulets give 100k, 150k, etc., and the hidden battle amulets all give 150k.
  • Calling waves early gives the same mana as letting it play out, it's "only" a score bonus.
  • There is no score cost to banishing. I'm not sure whether it breaks your combo.
  • The only mechanics which I know to be RNG-based is the colour/hue of the gems you start with and conjure, and obviously gem damage and the gem specials which are listed as a chance. The number, speed, health/armour/hue, and mana reward of monsters all seem to be fixed. I'm not sure about gem bombs, gems' prioritisation of targets, whether a monster will be affected by a trench, or the sort of "sideways" movements of the monsters. I'd be interested to know whether the time at which you create a gem influences the RNG for that, or whether you get a seed chosen at the start of a battle.
  • Not all gem colours are created equal. From my experience, I'd rank them like this:
    • Steal mana (orange): Exponential growth and all that.
    • Splash (red): How it works isn't explained beyond "there's a radius stat, number go up = better," but while I'm guessing there's a fall-off and it's not dealing 100% to nearby monsters, it's still a huge effective damage multiplier considering most monsters aren't alone.
    • Reduce armor (purple): I think, at least early-game, these are required in the levels which have them. The boss monsters are impenetrable otherwise.
    • Chain hit (lime): A worse version of splash. 40% to chain is approximately equivalent to +66.7% average damage.
    • Poison (green): The poison does stack, but isn't indefinite, so it's only an effective boost to damage (and range).
    • Slow (blue): Special is barely noticable at lower grades, so you'd be better off doing a bit more damage with that hit instead of applying slowness. Especially since you're all but required to use trenches anyway.
    • Triple damage (yellow): Sounds good, but by the numbers the special is quite useless. For example, the grade-5 gem's 18% to do 3x damage is equivalent to a paltry +36% average damage.
    • Shock (cyan): Good in conjunction with trenches for boss monsters, otherwise just a worse version of slow.
  • Breakdown of gem specials for grades 1–7:
    • Poison (queued damage): 5, 7, 10, 13, 18, 25, 35, 48
    • Triple damage (chance): 5%, 7%, 10%, 13%, 18%, 25%, 35%
    • Splash (radius): 7, 10, 14, 20, 28, 39, 54 (that's about 1.5 tiles in radius)
    • Slow (duration and chance): 2 s @ 10%, 2.5 s @ 14%, 3.2 s @ 19%, 4.1 s @ 26%, 5.2 s @ 36%, 6.6 s @ 50%, 8.4 s @ 55%
    • Chain hit (chance): 16%, 22%, 30%, 40%, then the same for grades 5 through 7
    • Steal mana (multiplier): 1x, 1.4x, 1.9x, 2.6x, 3.6x, 5x, 6.9x (nice)
    • Reduce armor (chance): 10%, 13%, 18%, 24%, 31%, 42%, 55%
    • Shock (chance to shock for 1 s): 7%, 9%, 13%, 14%, then the same for grades 5 through 7
  • Gem cost breakdown (with max. cost discount):
    • Grade 1: 21
    • Grade 2: 156, or 135+2*21=177
    • Grade 3: 426, or 135+2*156=447, or 3*135+4*21=489
    • Grade 4: 966, or 135+2*426=987, or 3*135+4*156=1029, or 7*135+8*21=1113
    • Grade 5: 2046, or 135+2*966=2067, or 3*135+4*426=2109, or 7*135+8*156=2193
    • Grade 6: 4206, or 135+2*2046=4227, or 3*135+4*966=4269, or 7*135+8*426=4353
    • Grade 7: 135+2*4206=8547, or 3*135+4*2046=8589, or 7*135+8*966=8673
    • ...So the pattern is, assuming you're starting from scratch, crafting from the grade below costs an extra 21 mana (= 1x grade 1), crafting from 2 grades below costs 63 extra (= 3x grade 1), crafting from 3 grades below costs 147 extra (= 7x grade 1), et cetera. Mathematically pleasing.
  • Towers' costs (with max. cost discount) are 48n+20: 20, 68, 116, 164, 212, 260, 308, 356, 404, 452, 500, 548, 596, 644?, 692?, 740?, 788?, 836?, 884?, 932?, 980?, etc.
  • Water trenches' costs (with max. cost discount) are 57n+20: 20, 77, 134, 191, 248, 305, 362, 419, 476, 533, 590, 647, 704, 761?, 818?, 875?, 932?, 989?, etc.
  • Not useful, just interesting: Some boss monsters are named.
  • Neither useful nor interesting: Orange and lime gems of grade 6+ look a bit like oranges and limes, and yellow ones look a bit like lemons.
  • And finally, since it's not mentioned in my OP (though I did link to the game), I'm talking about Gemcraft a.k.a. Gemcraft Chapter One: The Forgotten, SHA256 of .swf: 2557C15F7800A8526D3A8394BE9FA981553FF4CB0E120412FCB57FC061BF41C9.
I contribute to BizHawk as Linux/cross-platform lead, testing and automation lead, and UI designer. This year, I'm experimenting with streaming BizHawk development on Twitch. nope Links to find me elsewhere and to some of my side projects are on my personal site. I will respond on Discord faster than to PMs on this site.
Hey look buddy, I'm an engineer. That means I solve problems. Not problems like "What is software," because that would fall within the purview of your conundrums of philosophy. I solve practical problems. For instance, how am I gonna stop some high-wattage thread-ripping monster of a CPU dead in its tracks? The answer: use code. And if that don't work? Use more code.
Skilled player (1022)
Joined: 1/9/2011
Posts: 231
Nothing useful, but I would love to see a Tas of this game, too