Shuffle function need a tweak

[RedPred]

Half the times I use shuffle function on the scenario deck (mostly with Mari or Ilsoari Gandethus) I get the same card on the top (with often more than 5 cards). I m really thinking to avoid using it to save the next exploration paid with blessing/ally..

Edited July 23, 2016 by RedPred

[saxmaam]

Ilsoari is supposed to be put back on the top of the deck

[nsr]

There are 2 decks that can be shuffled:

1) The location deck

 This is shuffled by Merisiel's Evade ability as well as Ilsoari's "shuffle" ability

2) The character's deck

 This can sometimes be shuffled by character powers or by healing.

 

I assume you mean #1.  Yes I have noticed that Evade sometimes just ends up putting the monster at the top of the location deck, but I haven't recorded down the number of times in order to see if that's consistent behavior.

[RedPred]

Yes, case #1. I corrected the first post to make it clearer.

For example I shuffled a monster in a 8 cards deck and I got it again in first or I shuffled door-trap and it was again there in a 6 cards deck.

I don't use the shuffle function so often to have a good statistical data, so I m just asking you guys, if you get similar or better results than mine. That's to determinate if the function really requires a tweak.

[elcoderdude]

My feeling is a lot of the time a card is shuffled into a location deck, the card goes to the top. My impression is this happens at least 50% of the time. But I haven't actually tracked this.

[RedPred]

Ok, the same impression

[RedPred]

Just an other 'lucky' shuffle. The only one monster in Academy shuffled to the top of 10 cards location deck..

[PRuano]

Did a small experiment, and although the sample was small, it points to the fact that you guys are just unlucky.

 

Managed to cycle throw the top card 45 times, and only got 4 repetitions, on a single try.

 

You may wanna try it yourself.

 

The trick is using Merisiel Evade power to get the top card, evade, force reshuffle.

 

To setup, go to a quest and create a party with Merisiel (for the peek and shuffle), Kyra (for the 6 blessings and cure Merisiel), Seela (for 6 blessings and cure), Sajan (for 8 blessings), Lem (for 5 blessings and cure) and Lini (for 4 blessings and Cure)

 

Next put all chars on Town Hall (if available, the best since it allows a new explore with any type of card, not only blessings or allies) or another that allows extra explorations (either through discard or other effect).

 

Merisiel turn will be:

 

- explore, evade, use blessing/ally/location power, repeat until no more cards can be used to explore (be mindfull of always leave enouch cards in hand that you never run lower than 5 with hand reset)

 

For OTHER chars:

 

- give card to Merisiel (preferably blessings or allies that allow extra exploration, or other cards if location allow discard to explore), cure Merisiel if possible (Kyra has Cure spells AND Heal power) and then skip turn.

 

For each Merisiel exploration, make note of card drawn; for better results, keep a table with a number and a card, and just keep track of card number.

 

Do above until blessings deck runs out.

 

Then just check number of sequential cards drawn.

 

Results will vary of course, but doubt very much it will be 50% on average like you claim.

 

Its obviously an empiric and limited method, since we cannot program a large number iterations of draw a card, put back card and reshuffle of the location deck that would give better results, but does gives an idea that the shuffling method isnt as flawed as claimed (not claiming its good either, like i said its an ineffective method and only allows a small sampling for testing, which cannot give a good evaluation of the algorithm, but not as flawed as claimed).

Edited July 24, 2016 by PRuano

[Hannibal_PJV]

I Also have good results considering shuffle. There only Are 10 cards, so From time to time you get same gards to the top and I have Many times had same card in the same location deck that Also may cause some "what is happening" feeling.

[RedPred]

It could be for sure I got the worst unlucky shuffled decks in a row. That's why I started the thread, to determinate statistically through the community if it's a rare casualty or a 'common' condition. I don't know the algorithm they used, but normally you need the 'module' operator to implements the casual numeric extraction. So a bad implementation could get everyone to think to have an unlucky row..

 

Perhaps I also got some prejudice. I played a lot in the last week end and I have seen so many different bugs all together.. Hardly to miss them if your job is develop application too. So in my opinion game is a commercialized beta version. And the rate fix/lifetime is so low, it will still be there for next year. They have first to fix their quality system and modular test procedures. All the durable bugs and instability are just a consequence.

 

It's really a pity, this is actually the best digital rpg board-game (without the bugs) for Android system and pc (through several Android emulators) actually in the store. And I looked for it so long... and multiplayer feature is still so far.. So sad about that

[Ethics Gradient]

I'm going to agree with everyone that shuffle by itself probably isn't broken.  

 

But be careful about narrowing it down to just those two cases.  There are literally dozens of things which initiate a shuffle in this game.   Evading a bane starts a shuffle, undefeated villains start several shuffles, augury starts a shuffle, shuffles galore starting a scenario, etc...

 

Given the sheer number of times the game calls for the shuffler, it seems evident that it behaves most of the time, so it isn't globally broken.  However, that that doesn't rule out a narrow instance where an expected shuffle goes awry because of some other code issue...

[firestormkirby]

I also get the impression of bad shuffles. But the human brain has a bad memory bias, so all the times it works as intended trend to get over shadowed by the few times it is bad. I haven't done any analysis myself so I wouldn't endorse anything being wrong yet. But as mentioned above there may be a line of code somewhere not working for an edge case shuffle. More concrete details are needed for sure.

[Ethics Gradient]

I also get the impression of bad shuffles. But the human brain has a bad memory bias, so all the times it works as intended trend to get over shadowed by the few times it is bad. I haven't done any analysis myself so I wouldn't endorse anything being wrong yet. But as mentioned above there may be a line of code somewhere not working for an edge case shuffle. More concrete details are needed for sure.

EDIT: (previous post was probably wrong)

 

There is a shuffle, and it has a classic CS101 vibe to it.

 

If there are any "issues" in shuffling (rather unlikely), it is a Unity Engine or PRNG thing, not necessarily a fault of the game.  Unless you played the physical game with precision dice and a mechanical shuffler, things are likely more far random in the digital version.

Edited July 27, 2016 by Ethics Gradient