So I should ask...

Is this happening often enough that I should pull the iPad update from the process? It will probably go through the approval process sometime today or tomorrow if I leave it be.

I havent noticed it on the iPad, but then again I never saw the AI do that on the iPhone either....

agfan, would you mind sending me your UDID for your iPad?

Yeah, I don't think this is as wide spread as I'm afraid it is. You really have to have combination of a lot of trumps/doubles and roll a nice random variable to make that bid.

I honestly wouldn't even mind this bid if an agressive player is making it. If his partner is able to pick it up, they are probably going to make their bid.

Happened to me yesterday and my step dad this morning. Regardless I'd rather have the nello option and drawing correction than miss out on that on the iPad for another two weeks to fix this now, but that's just me.

Well, this was fast. iPhone revision from last night is already in review.

The iPad version has been reviewed and is ready for download.

quote:

---

The iPad version has been reviewed and is ready for download.

---

Will update tonight.

Sure thing, ill get it to you this evening when I'm out of classes

No rush...especially now that it's been released.

I expect to start multiplayer testing in another month. Still trying to wrap my brain around a few things.

like what?

(email if you'd rather)

Is there a way to make the nello option have doubles low in their suit as an option and not just a suite of their own?

aka low doubles low, which zone416 taught me to play with Brenhamag0412 and me setting him on the second trick.

[This message has been edited by outlaw08 (edited 8/29/2012 2:36p).]

Philip J Fry - Can you mulitplayer from ipad to iphone? If so, I can send you my iPad info in email and help you beta if you'd like.

Nope. Doubles are treated as high or a suit of their own in this release. Doubles low will come out later. Best I could do in the amount of time I had.

Rebel,

The issues I'm facing right now involve playing anything less than 4 players. For 2-3 players, someone has to be in charge of the AI. The code right now is set up so that the host makes these decisions and then distributes the move to everyone else. There are a couple problems with this.

1: What happens if the host disconnects?
- Simple answer is for someone else to be the host if the current host disconencts. But what happens if everyone else has has the game in the background? Does the game just quit?

2: There is a bit of an issue with lag. If a non-AI/host wins the hand and starts shuffling before everyone is caught up the game hangs.

3: If a player lags while dealing the tiles, it's possible for play to start without him having a full hand. Obviously causes the game to crash.

Those are the types of things I'm dealing with when I just have 2 players. I haven't even tried 3 players yet.

If I understand it right, gamecenter does let you connect two seperate apps together like this, but I haven't tried it yet. Still on baby steps and its a big learning curve.

[This message has been edited by Philip J Fry (edited 8/29/2012 2:42p).]

The best solution is to have all players connect through a remote database that stores the table information... I'm guessing that is out, as it would take infrastructure and bandwith, along with a server side architecture, along with major redesigns of simple operation... maybe this would be best in 3.0 once ESPN finally picks up tournament 42.


It sounds like you are wanting 1 host and 1-3 clients. My big question is, does iOS finally have true background tasking, or is it still the freeze & store the RAM then switch over to the new foreground task?

My first thought is to create a database holding the table info on the host machine, then the clients push update requests. The host updates the DB, then pushes it back out. This sounds pretty packet intensive depending on the DB size. On a disconnect, the hosting title moves to whoever is active with the newest version of the DB...

A second idea, which is also pretty packet intensive would be to just pass the DB from machine to machine allowing each player to make their change... essentially passing the hosting title around the table.

The third idea, reduces the amount of data that must move from machine to machine, but seems a bit delicate... The Host creates a DB and pushes it to each client. During play, the update requests are passed to each client to update their own DB... I can see countless "DB mismatch" errors coming from this, however.

You may have already considered all of these, and likely have a much better method in mind. Is this going to be WiFi only? 3G? Is Bluetooth possible since you need 3 connections to one device using the same service?

ah, I see the problem now that I wrote that...

assuming that each device will play identically, you can have the "current human player" device fill in any missing plays that he gets from AI between the DB leaving the last human player's control. This way, if a player drops, he will have either already played, or a play will be generated using the AI algorithms.

It will be compatible with both 3G and wifi. Haven't given blue tooth a thought at all to be honest.

Your first thought is basically how it's set up currently. The host player will receive a play by a human and then redistribute it to everyone else.

Each device will not play identically though the way it's written now. When faced with the same hand, it may make a different bid or play a different tile.

And the iPhone version is now live. Holy crap, I wish it was this fast every time!

Here's another small one. Fry needs mine point to make his bid and has the high trump but punts. Obviously he'd win regardless but it was a little goofy.

When someone goes 2 marks, is the computer taught to hold count vs holding things that might set the bidder?

For 2 mark bids and v1.1 (actually fixed in 1.0.1), the computer is indeed taught to hold on to it's high tiles regardless if they are tiles with count or not. There are several spots where I reference count tiles. If it looks like its trying to play for count, send me a screen shot.

Lucky, he needed 1 point to make his bid...had the high trump and didn't play it? That might be an area I can look into for a later release. I know there is code for trumping in if it means they'd lose the hand otherwise...but that's for plays 2-4. Not sure if I have something for the first play now that I think of it.

[This message has been edited by Philip j fry (edited 8/29/2012 10:47p).]

I usually play where you can only bid nello on a force bid.

And where's my plunge bid!

Ap is getting better every update

Right now, you are the only player that can make that bid so I thought it would be a little tedious to add a completely seperate menu for house rules in this case.

Once I start allowing the computer to make this bid, or once multiplayer comes out, I'll definitely allow you to choose whether or not nello is played on forced/unforced bids...or not at all.

I know its a nuance but it has happened to me multiple times, but as soon as someone draws the 6-6 at the beginning can you just go ahead and end the draw process as well, nothing is going to beat that.

I'll add it to the list.

danke

This just feeds to my addiction of the game. Getting to be a small part of the process of refining it and making it better. Awesome job

I'll bring it up again...

is it even necessary to actually draw to see who randomly goes first when a pseudo-random number generator can pick a number 1-4 and be done with it?

if it is just for aesthetics, can it be optional?

It certainly isn't integral to the program running. It's mainly for aethetics as of now, but really meant to have user interactions during multiplayer instead of just assigning roles.

It would not be difficult to add an option to turn off that function altogether and just make it random to start off with, but it isn't on the top of my priority list at this time.


[This message has been edited by Philip J Fry (edited 8/30/2012 5:23p).]

you guys really complaining about a draw process that takes all of 5 seconds? And, I thought I was nit picky about stuff.

Hey, as both a programmer and a professional 42 player I am all about efficiency.

quote:

---

you guys really complaining about a draw process that takes all of 5 seconds? And, I thought I was nit picky about stuff.

---

This is texags. If we aren't complaining about something, we aren't happy.

Honestly, it would take about 10 lines of code total to skip it altogether. The problem is real-esate. Making it an option won't fit anywhere in the current options scene, so it's going to have to go into the "house rules scene" (which doesn't exist in illustrator yet) or I have to re-arrange the art-work in the current set up.

So like I said, I can do it...it just isn't a priority. At least not for a stand-alone release. Multiplayer and Android is where my focus is right now. This 4 day weekend should be a good time to get some work done.

[This message has been edited by Philip J Fry (edited 8/30/2012 6:04p).]

I can't let this one go...

I bid 31 in 6s

66 65 62 61 53 21 42

I lead to pull trumps

Play in order:
66 64 60 63
65 ?? 50 ?? (indication?! I don't trust the AI to indicate yet)
62 ?? 55 ?? (okay, partner is out of count, and did indicate, he's also out of grease)
53 51 ?? 54 (well, damn. Once again my partner sloughed a double he couldn't cover... Noted)
33 42 41 3? (well there went that plan...)
22 21 ?? 32 (!!! What?! My partner dumped a double for count and STILL HAD POINTS?!)


Should have been such an easy hand to make... I'm trying to pseudo code my thoughts for you.

Partner has no ability to indicate. I'd love to get that kind of code in there, but it isn't in there right now. You are going to feel nothing but frustration if you are trying to read more into the AI than is there.

I'm having a little trouble following the issue here. Do I understand it right that your issue is that he played the 55 for count when he did because you permanently lost control of the board with your 53?

That issue is on the first page. I have not yet attempted to fix it yet :/

[This message has been edited by Philip j fry (edited 8/30/2012 7:49p).]

Kind of.

If the 55 is played for count, it should always be last. The big issue is that my partner had the 32 and decided to play the 55 (a double) for count instead of a lowly 32.

The setting point was when P was forced to play count. Had he been forced to play the 55, we still win it.. The 55 is very rarely a danger to itself. The 64, 32, 41, and 50 almost always are.

This is probably 42 strategy++ and getting into why I abandoned the venture to code the AI five years ago on a windows platform.

Yeah, that's what meant. If he played the 32 and held the 55 for your 5 off, that would have won you the hand.

It's plays like this that I go really want to get right. It'll get there at some point. The next revision (probably just 1 more before multiplayer comes out) will make better decisions during unforced moves. Not sure about indicating though. I'm sure it could be done, but I'd need a pro like you to walk me through all the different indicating moves out there...likely a 3.0 release.

[This message has been edited by Philip j fry (edited 8/30/2012 7:58p).]

Heck. It might be interesting to give different AI players different indicating strategies.

I know a bunch, most logical, some prohibitive, and a few illegal.