Showing posts with label Castle Crisis. Show all posts
Showing posts with label Castle Crisis. Show all posts

Tuesday, February 14, 2017

An Archdemon Highscore

Castle Crisis is on, and as usual the iOS server did the first three subgames on the third day (and not long into it, either).  Fortunately for me, I am usually up at midnight when the tries-per-day is reset for Castle Crisis.  So I always do well on the lucky imp, and I can always get the full set of crest rewards from the Avaricifer.

Now it's archdemon time.  I used to get great scores against the one that could be beaten with a custom Atlanticore and Death Knight.  Then they stopped sending any archdemon except the latest one.  (I spend a lot of shards getting that DK's skill to level 9!)

The current archdemon is very hard to deal with, since it is immune to stunning, scatter, and silence.  It is not immune to slowing, but slowing it doesn't help much since it will switch targets from time to time (not sure how this works).  This will mean that it will target your PD, who is always close, and then, unless you can tank the damage, you lose.

If you can tank the damage with PD anyway, you might as well forget about slow locking and tank the damage.

The only way I am aware of to deal with it is a combination of massive hitpoints, lots of healing, and protection for your heroes.  Unfortunately for me, I have no Ghoulem.  So I am lacking both the massive healing and the group protection that he supplies.   I can get in the 10 million range just by brute force: lots of revive crests, Tre, and luck.  That's about a much as I thought I will ever get unless they change the archdemon, or I roll Ghoulem.

Today another thread about the archdemon came up on reddit, and after reading shlam16's comments I decided to have another shot at it.  One interesting thing I learned is that this archdemon has a high crit rate, which accounts for a lot of his instakills.  Makes sense.  Valentina can protect somewhat against crits; otherwise it's a vulnerability that you cannot fully address.  You mitigate it by having revive crests or talent on every hero, and Treantar to revive them a third time.

I don't have Ghoulem or Valentina, so I used the core team suggested by shlam, but putting in Medusa and Cupid.  Medusa is great for getting rid of the eye-catchers, which is another thing Ghoulem does if you have him.  Cupid amps damage and adds energy, like Valentina.

For those curious, my lineup is: PD/DD/Druid/Cupid/Medusa/Tre.  The first four are double evoed; Medusa and Tre are single evo.  I crested them all with Revive of various levels (4/4/4/3/1), except Tre who has Revive 5 for his talent.  PD has Berserk 8 and the Blitz Scroll 5, so he amps up fast.

My first run didn't work very well.  I didn't exceed the 17.3 million I got yesterday.  I was close, though -- 17.1m -- so I felt positive about the second run.  Obviously there is a lot of luck involved with the archdemon; you have to try a strategy a few times to have a decent sample of what it's capable of.  (Indeed part of what is hard about archdemon is that you only get two tries per day, so it's hard to experiment.)

The second run, things just went much better.  Here it is after two minutes...
Beating on the archdemon.
You can see PD, Druid, and Cupid have all been revived by Tre.  This means they are on their final life.  No permadeaths yet though.  Eventually, though, the archdemon killed PD one more time:
PD down for the last time.
And after that it goes downhill rapidly.  With PD gone, his stacks unwind over 8 seconds and then the eyecatchers can't be killed fast by Medusa, and they amp up the demon.  Meanwhile, Druid stops his heals every 1.4 seconds, and Tre's shield is no longer providing nearly fulltime coverage.
The end is nigh.
Then a rapid end, with about 26 seconds left.

183m is currently good enough for 382nd place, which would give 1200 fame if archdemon ended today.  My old highscore, 17m, was around 4100th place with a reward of 700 fame.  500 fame is worth working for!

To improve my reward, I would need to place above 200th, which currently requires at least 326m damage.  That seems quite unlikely.  Even if I can get luckier and go the full 3 minutes, it would only increase my score by around 30%, to around 237m.

One easy thing I could do to improve my score is beef up my Druid's skill, currently at level 7.  This won't get more than about 5-10% though.  Not enough to matter, rewardwise.

Other than that, I can't think of any cheap way to get more.

Friday, October 14, 2016

Maxxing Damage on an Archdemon

The last post raised an interesting question: which hero is better against an Archdemon (damage cap: 500000)?  Succubus, or Phantom King?

Succubus procs every 7 hits, getting the full 500000 damage in her proc.  PK procs every hit, with proc damage similar to his normal hit damage.  Thus, the damage multiplier is going to determine the difference.

When PD is at max speed and being hit by the Archdemon, he procs roughly every 1.2 seconds.  He'll proc a little more often, actually, because of energy that Cupid adds.  But let's ignore that.  PD's proc lasts 8 seconds.  So, we'll have a sustained stack of 6 or 7 procs.

Let's assume 6 procs for now.  My PD has max skill, so +55% bonus per proc.  So we compute the net multiplier for PD as 1.55^6, which is 13.86.

(Incidentally, note that the speedup with 6 procs is also 13.86.  This is how we know that all heroes, even Atlanticore, are hitting every 200 ms.)

Druid also amps team damage, though less.  He procs every 1.4 seconds, with procs lasting 5 seconds.  That's 3 or 4 procs.  My Druid has skill level 7, so he gets +29% per proc.  Assuming 3, we get a net multiplier of 1.29^3, which is 2.14.

Finally, Cupid is always active, adding +45%, or 1.45.  He doesn't stack but his cooldown and proc effect are the same.

So, we find a net team multiplier of 43.1.  This means that any hit that is at least 11589 will end up as a maximum hit.  The multiplier can go considerably higher, if both PD and Druid are on the higher number of procs they can do.

Returning to the original question, I find my Succubus hits for 10828.  (She is evo 2.)  So, she is doing near max hits each hit.  Thus her total damage per 1.4 seconds is 40,000,000.

Now we compute PK's damage and compare... but wait just a second!

Succubus does 40m damage?  That makes no sense!  If she gets 40m damage per 1.4 seconds, over a full three minutes she should get 5 billion damage.  And yet the #1 player in the world got about 1.1 billion for his whole team.

Obviously something about the whole calculation above is off.  Way off.

It's obvious in retrospect.  The Archdemon dodges.  You can see it happening in any video of the Archdemon (check youtube).   For the damage reduction implied by the amount I actually scored, he must dodge a lot.  Yet I don't see that in videos... do you?  It's a mystery.

Well, I went to reddit and asked about it.  More than one guy there told me that the dodge rate is fairly low, but in addition there is some sort of stone-skin like damage reduction.  After looking at videos some more, I agree.  The dodging rate you can see is low.  You just don't see a lot of MISS notifications compared to hits.

So, my best guess at this point is that the Archdemon dodges about 10%, with a mysterious damage reduction of about 90% (!).

Returning to the original question... well, let's un-ask that.  We know that practically all hits that we're doing are not in fact capped, and indeed nowhere near capped.  (The only hits that might hit the cap are Deadly Strike hits and (maybe?) critical hits.)  So, amping up damage more is the name of the game.  There are two heroes that can do that: Siren and Beast Tamer.  Beast Tamer adds another whole hero to the lineup, and so is worth roughly +16% just for that.  At low level, somewhat less... call it +10%.  Then on top of that, at skill 7 he adds +22% damage for 9 seconds out of 12.  Siren, also at skill 7, adds +23% damage all the time.  So obviously Beast Tamer is the one to go with.

Hmm.  I already have my Siren at level 120, whereas I still have Beast Tamer as a card because I have never found any use for him.  Ah well... I need the might.
Come out, beasty boy
What's his talent?

Baby Beast Tamer 
Deadly strike!  It's not useless in this application.  I'll put on Slowdown crests once I level him up.

Tuesday, September 27, 2016

Avaricifer Donation Strategy

I've been looking at the Avaricifer rewards now that IGG has scaled back the maximum amounts you can contribute.  The limits are are now 2m gold/mana per day.  Obviously at this point contributing maximum gold and mana per day is a no-brainer.  But what about merits and honor badges?  Is it worthwhile to contribute them?

Even though we get ranked, the rewards for Avaricifer are not based on relative standings.  Rather, they are based only on the total amount you give.  Here is the reward table:

Note that you get all of the rewards of lower levels of donation at a higher level.  So, if you give the maximum donation (180000+ "donation points"), you get:
  2x mythic crest bag, 2x rare crest box, 2x crest chest, 2x mythic crest chest
You can see this on the leaderboard.  All the top guys get this reward.

Currently it is the second day of the contest, and the Avaricifer is 37% done.  So, it appears that it will take either four or five days to complete the contest.

So far I have given just gold and mana, twice each, scoring 44000 donation points.  Thus, we can compute easily that at this rate, I will end up at either 88000 or 110000 points.  So I'll get the two rare crest boxes, no problem.  I will get one crest chest, and I may or may not get the other.

Mythic crest chests give level 3 mythic crests; obviously, that is a nice reward.  But a crest chest (not rare or mythic) gives a level 3 (basic) crest; and that is not much of a reward.  The only basic crest I would use is a slowdown crest; all others I would forge in the attempt to get something useful.  Forging rarely produces a desirable crest.

To get one of the mythic crest chests, I would have to donate an additional 25000 or 47000 donation points.  To get both, I would have contribute an additional 70000 or 92000.

One more part of the puzzle is the cost to donate.  Again, I've maxxed out gold and mana and assumed I will max it out for all two/three remaining days.  So the only things I can contribute are shards and honor badges.  One can contribute up to 400 shards/day, earning 20000 donation points for max donations.  For honor badges, the limit is 20000 hb/day earning 20000 donation points.

So, we have the possibility here to convert either shards or honor badges into mythic level 3 crests.  To get one additional level 3 mythic crest will require 500/940 shards, or 25000/47000 honor badges.  Getting two additional level 3 mythic crests costs 1400/1840 shards, or 70000/92000 honor badges.

Given the uncertainty in the event timing, what seems wisest to me at this point is to aim for the 135000 level in four days.  Then if there is a fifth day, go for the maximum.  I can still give shards today, so I have up to three days worth I can give.  That's plenty to get to 940.  If there is an fifth day of the contest, then I can contribute 22000 from gold/mana, and make up the remaining 23000 using mostly shards and some honor badges.