Posts created by kimberph

 
 
 posted in Midnight Castle on Jan 16, 18 1:38 PM
I was away for 10 days during this event (earlier, back now) and had made the decision even before I left to not collect gifts in the hopes that collecting them later would help with CC items. So, when I finally started opening my presents around 1/3 I had a lot, at least 60 total items and I'm pretty sure far more than that (I wasn't counting). (And it did get me at least one of the stamps I needed for the CC, so the strategy worked.)
 
 
 posted in Midnight Castle on Dec 18, 17 6:34 PM
I'm sending you an invitation. I play as Brin.
 
 
 posted in Midnight Castle on Jul 22, 17 1:27 PM
I just timed myself for a round. My path wasn't the most efficient since I was completing DQ steps and key quests along the way, as well as trying to remember to hit hints. But even with that it took me 62 minutes to get through all the quests. (I'm maxed out, so I guess that is 93.)
 
 
 posted in Midnight Castle on Jul 11, 17 5:15 PM
Woo hoo. Inspiration!
 
 
 posted in Midnight Castle on Apr 26, 17 11:54 AM
I have yet to see anyone provide any actual data driven evidence that there are any functional "tells" for the pets. Personally I'm not willing to spend the resources to do a few hundred repetitions of different scenarios to have useful amounts of data, and I don't believe anyone else has done this either. I personally seriously doubt most of what I've read about tells. No one is talking about something that works 100% of the time. People are talking about relatively small changes to percentages (e.g., increasing the chance of a gold chest from say 50% to 70%). The sample size to discern this is very large. Most people's samples are on the order of a handful which simply isn't meaningful.
 
 
 posted in Midnight Castle on Apr 24, 17 12:14 PM
TiramaSue wrote:Guessing which hand is no different from flipping a coin and guessing whether it lands heads or tails. There's only two choices and the odds are the same. So no way to make the game more difficult.



Actually the odds are whatever the devs set the odds to. I believe based on several hundred rounds of data that the odds are likely fairly close to even and if there is a bias it is probably in the direction of winning.

 
 
 posted in Midnight Castle on Apr 24, 17 11:20 AM
I posted odds of winning (and why and what the right strategic approach was) in another thread ("Odds for puzzle game ").

Bottom line:
If your goal is diamonds:
If you lose both of the first two rounds, buy exactly one more round.
Otherwise, play until you have won the diamonds and lost one round.
This gives you a 75% chance of winning 5 diamonds, a 12.5% chance of winning net 3 diamonds, and a 12.5% chance of losing 2 diamonds. (And 25% of the time you will win some objects too.)

If your goal is objects, the strategy is trickier, see the other thread and the "Strategy 3" post there.

All the odds calculations assume an even chance of winning any given round. I've got enough data now to be pretty confident that if there is any bias, it is toward winning rather than losing, but it looks pretty close to even.
 
 
 posted in Midnight Castle on Apr 17, 17 6:29 PM
So sorry Sherry. Take care of yourself.
 
 
 posted in Midnight Castle on Apr 12, 17 5:55 PM
With the diamond bounty that is the puzzles, I'm not worried about the Troll. We can afford to accelerate the wait as many times as it takes.

I'm much more worried about the 23 trades with the kids plus the 4 spring items to craft. Combined need to 23*7 + 15 + 3*20 for a total of 236 flowers and sweets. Add to that the DQs that may need more of these same items and I guess it's back to HOS cycling.
 
 
 posted in Sunken Secrets on Mar 17, 17 3:28 PM
Has this game ever had more content added (other than special events)?
 
 
 posted in Midnight Castle on Mar 15, 17 5:56 PM
Before the update I had enough data to be pretty confident that the odds of winning the round were about 57% in the first two rounds. My sample sizes weren't large enough for later rounds to be confident of numbers, but it looked like the odds dropped to about 50% after that.

After the update I definitely don't have enough data yet, but what I have is not trending the same. It looks like the odds of winning have now dropped to 50% for all rounds. In other words, the game is now fair instead of favoring us.

If this holds, then the odds I gave in another thread for various strategies and what results you can expect are accurate.

My bet is that the result is actually predetermined and it doesn't matter which hand you pick. I haven't seen anyone provide any data to contradict this and it is certainly simpler to program.
 
 
 posted in Midnight Castle on Mar 13, 17 4:56 PM
One more pretty interesting change. I won all prizes in my first puzzle. I can't pick any of the same items in my second puzzle. I hope that resets with the puzzle reset.
 
 
 posted in Midnight Castle on Mar 13, 17 4:53 PM
Also recharge is 5 hours now, and diamonds have moved down a level (below the first pick) [edit] depending on the value of the items picked.

These aren't surprising changes. The puzzles were way, way, way too profitable and would have quickly accelerated the game. I don't play much and even I managed to pull in 400+ diamonds (maybe 500, I wasn't keeping track) and probably 80 eggs.

On net, probably a really good idea to make these changes.
Wonder if they messed with the win/lose ratio too. Guess I'll have to start over collecting stats on this.
 
 
 posted in Sunken Secrets on Mar 8, 17 3:16 PM
I've got a few left. I can confirm that door, window, and chimney all now only need 1 magmatic stone (no amber) and 1000 magic for the last upgrade and still upgrade magic storage by 20 each.
 
 
 posted in Midnight Castle on Mar 3, 17 1:33 PM
The puzzle is a really efficient and effective way to get diamonds. If you play to optimize your diamond rewards you will win 5 diamonds at no cost 75% of the time and net +3 diamonds another 12.5% of the time.

Edit: should have mentioned that in the final 12.5% of the time you lose 2 diamonds.
 
 
 posted in Midnight Castle on Feb 27, 17 5:49 PM
Annelliegram wrote:Well.....I still cannot say the game is "fair".......and I had already defeated all other players, with only the dragon left for checking.

...

I certainly would not totally disagree with Biker_Bear ......could be that a future update will alter the percentages ....I sure hope so!


The question was not whether MC is fair, but whether the puzzle pick-a-hand game is fair (i.e., a 50/50 shot of being correct on any individual pick), and my experience hasn't led me to believe it is not nor has anyone provided any data to support anything other than this.

Assuming this is indeed true, the odds that I posted for various strategies on the first page of this thread are correct. Whether people choose to use that data is of course up to them.
 
 
 posted in Midnight Castle on Feb 26, 17 1:49 PM
Annelliegram wrote:Regarding the side discussion about dice table......I kept track of the final 100 needed to beat the Dragon......one tick mark for each win on the right....one tick mark for each loss on the left.

I threw away the paper of tick marks, but I remember that I won 47% of the time.....which really surprised me because I would have sworn that I was losing about 2/3 of the time.

So, my record keeping revealed (for me) that the "house" does indeed have an advantage, but not as much as I had thought.


47/53 is well within an expected division for fair dice throws with a sample size in this range. So, now we have rough data at least and it supports that the game is indeed fair.

And your statement that you thought you were losing much more often than this is exactly my point.

Thanks for sharing this!
 
 
 posted in Midnight Castle on Feb 26, 17 11:12 AM
Biker_Bear wrote:Not anecdotal. I indeed ran two sets of each as a trial. Further more I have done the same for the dice games when attempting to win avatars. The odds were indeed skewed. Since winning all the avatars, the odds have returned to fair.


OK, please share your data. A sample size of 2 is meaningless, so I'm unsure what you mean by "ran two sets of each". Until you share data, it is anecdotal.
 
 
 posted in Midnight Castle on Feb 26, 17 10:28 AM
Biker_Bear wrote:The odds for the game as has been calculated is based on a "fair" game. Fair meaning as in the flip of a coin. However, much like the dice table , this game is rigged. I have used the left 10 consecutive times with no wins. I have done the same with right and the same result. I smell a rat.


I disagree with your assessment for both this game and the dice game. My observations indicate that both are entirely fair. I haven't seen anyone post any actual data that would indicate otherwise. If you have some, please post your actual data.

Human nature is strongly biased toward recollection of loss rather than win in things like this. Anecdotal recollection is therefore fundamentally unreliable.
 
 
 posted in Midnight Castle on Feb 25, 17 1:37 PM
(continued)

Strategy 3: maximize objects with minimal diamond loss
This one is definitely the trickiest because now a relative worth calculation has to be made. You have to decide how many diamonds are worth risking for a 50/50 shot at winning the object. This is a personal decision. I've written it up with the assessment that I'm willing to risk a 50/50 shot between losing 5 diamonds and winning 1 object and I'll take risks up to that much, but no more.
This results in 12 possible outcomes.
1 gets 3 objects and 5 diamonds at a probability of 6.25%
1 gets 2 objects and 5 diamonds at a probability of 6.25%
3 get 1 object and 5 diamonds at a probability of 37.5%
2 get 1 object and net 0 diamonds at a probability of 12.5%
1 gets 1 object and 3 diamonds at a probability of 6.25%
2 get 0 objects and a loss of 5 diamonds at a probability of 12.5%
2 get 0 objects and a loss of 2 diamonds at a probability of 18.75%

1 WWWW 6.25%
2 WWWL 6.25%
3 WWL 12.5%
4 WLW 12.5%
5 LWW 12.5%
6 WLLbW 6.25%
7 WLLbL 6.25%
8 LWLbW 6.25%
9 LWLbL 6.25%
10 LLbWW 6.25%
11 LLbWL 6.25%
12 LLbL 12.5%
 
 
 
 
 
 
 
 
Go to:   
 
 
 
KEY:    Unread    Read    Sticky    Locked    Announcement    Moved    Moderator