Page 1 of 1

Nerdy scientist Caliph's Tomb post :-)

Posted: 25 Sep 2013 01:13 am
by Wingsrising
So during EpiCon, Caliph's Tomb came to my attention. I like Pyramid, but hate how hard it is to win, and was pleased to discover that the two rightmost Caliph's Tomb layouts were pretty easy to win and pay 10k sP per win for up to 10 wins a day (provided you only send wins).

I started wondering which of the two right-hand layouts was easier to win. Wondering this sort of thing is always dangerous.

About a month and 300 games of Caliph's Tomb later (150 on each of the two right-hand layouts) I've reached the conclusion that I cannot reject the null hypothesis that the probability of winning the two layouts is the same.

The top right layout had a 70% win rate, the bottom right had a 75.33% win rate. The chi-squared test I ran produced a p of .13, meaning that if the two layouts had the same chances to win, there's a 13% chance I'd get a difference as big as the 5% difference I found. (That may sound like a low probability, but the usual criteria for concluding that the two things are actually different is p=.05.)

I'm torn between thinking, "Yay, no more recording wins and losses" and thinking, "Hmm, I wonder if they would still fall short of significance if I took it up to 500 games?"

Yes, I am a nerd (and a scientist) -- why do you ask?

Incidentally, if you like this style of solitaire -- which I obviously do! -- it's not the worst way to make sP in the world (provided you use one of the two easy right-hand layouts). I made 2,180,000 sP over the course of this experiment. Not staggering, but not bad given that I also had fun playing.

Re: Nerdy scientist Caliph's Tomb post :-)

Posted: 25 Sep 2013 01:37 am
by Ierosbats
Unless they've lowered the payout very recently I think you actually get 20k per win.

This game is definitely my favorite thing to have come from Epicon this year. Subeta has a habit of making games that are 1) not fun and 2) not worth your time in terms of payout (most games have both of these problems) but this is definitely an exception. Fair payout for the time it takes, and decently engaging. The new Tiles game is almost good too (I love mahjong, but always hated the clunky Neopets version) but I can't get over how stupid it is that you get 10k per win regardless of difficulty, and that shuffles/hints are unlimited. Maybe they've changed that since Epicon (I've only played a little) but I doubt it.

Re: Nerdy scientist Caliph's Tomb post :-)

Posted: 25 Sep 2013 01:56 am
by Wingsrising
:headdesk: You're right, it is 20k per win. Which is even better, and means I made 4,360,000 over the course of the experiment. I think I was mixing it up with the 10 sent games allowed per day.

I find that Tubular Tiles takes too long to play to be worth the payout, but maybe I'm just not that good at Mahjong.

Re: Nerdy scientist Caliph's Tomb post :-)

Posted: 25 Sep 2013 06:29 pm
by AngharadTy
I ran a test over a few months when they revamped the Crypts game--it was the only way to bring myself to actually do it regularly. I found out that, no matter how many exits you can click on in a room, it's still a 50% chance to proceed every time. However, my data set for some of the innermost rooms was, understandably, small. I think it was when I got through and got a 50k item instead of one of the new prizes that I gave up on that project. Still, running data on Subeta? I support that entirely!

Re: Nerdy scientist Caliph's Tomb post :-)

Posted: 25 Sep 2013 09:04 pm
by Wingsrising
That's interesting information about Crypts.

I've wondered about the probabilities associated with the Ice Fields, myself, but not enough to actually track them.

Re: Nerdy scientist Caliph's Tomb post :-)

Posted: 14 Oct 2013 01:17 am
by shaelyn76
I easily complete all 10 games I am allowed of Caliph's Tomb each day within a half an hour. It is a quick and easy 200k since I only submit the score when I clear all cards...which as you noted, is pretty easy on either of the right hand side choices of layout. I was a seriously happy camper when I saw this addition to the Subeta games room since (like others have said both here and onsite) the vast majority of the other games just aren't worth the time input or effort for the payout. And I'm right there with you Ierosbats on the whole Mahjong thing. Why do I still only get 10k no matter what difficulty level I clear? Doesn't make any sense. I have totally been abusing the multiple reshuffle option though since I tend to be a bit lazy about the way I play this version. Since I know I can just reshuffle to my hearts content...I tend to play it in between doing other stuff just so I can get the wee bit of extra sP from it each day. (trying to stockpile some sP for X-mas).

Re: Nerdy scientist Caliph's Tomb post :-)

Posted: 17 Dec 2013 01:03 am
by Wingsrising
Thread necromancy!

I did indeed decide to take my data collection up to 500 games. Actually, to 600 games. Actually, to 602 games since somehow I went 2 games over.

Now my win rate is 75.4% for the top right, 76.1% for the bottom right, p=.79 meaning that if the two games actually have the same win rate, there's a 79% chance I'd get a difference in win rates at least as big as the one I found. And I made 9,120,000 sP over the course of the experiment.

In short, it appears there's no difference between the two right-hand layouts. Which is sort of surprising to me because it seems like that inverse pyramid in the middle of the top right layout would be harder to clean out, but statistically I guess it works out.

I think I am now declaring myself done with Caliph's Tomb data collection (although not with Caliph's Tomb). Either that or I need to start collecting data on the average number of moves it takes to win with the two right hand layouts. :-P

Re: Nerdy scientist Caliph's Tomb post :-)

Posted: 17 Dec 2013 07:54 am
by AngharadTy
Nice! That's intriguing overall. I also like knowing the amount of sP you made for that effort--pretty decent for a thing you might also (gasp) enjoy playing. ^_^

Now, I expect it might be a little hard to break the habit of opening Excel (or whatever) each time you play a game, hehe.

Re: Nerdy scientist Caliph's Tomb post :-)

Posted: 14 Apr 2014 11:49 pm
by Wingsrising
So I kind of got out of Caliph's Tomb for a while (in favor of questing, which is much more lucrative but much less fun) but I got back into it recently and decided to try to collect information on how many cards were left in the deck at the end of the game following a win, to try to capture a sense I had that maybe the bottom-right deck tended to yield faster wins. (Higher number of cards left = faster win.)

I collected data from 100 games of each (as well as adding to them to my win/loss count, because why not?) with the result that there were an average of 6.34 cards left at the end of the game for the top-right layout, 7.02 for the bottom right layout, p=.28 (so not a statistically significant difference). So no evidence for the hypothesis that either layout finishes earlier than the other. Now, this doesn't capture the situation where you only have one card left and are just flipping through hoping for the right card to get rid of it (which doesn't take long) but I don't know how to accommodate that one. I suppose I could time games, but that seems like it would take a lot of the fun out of it.

As far as percentage wins go, I'm standing firmly at 76.3% for the top right and 76.7% for the bottom right, so no difference there.

Bottom line, still seems like it totally doesn't matter what layout you play between the two on the right.