??????
You are not logged in.
Just looked at this.
In terms of implementation, the mousing over cells works only on cells that are assigned to red or green, so it knows what you're trying to test.
You're asking to mouse over uncolored cells (or click on uncolored cells) before betting after the second pick. But what color should the game assume you're trying to test for those cells?
During pick three, you can only mouse over an uncolored cell after you have placed a slider---thus, the game knows what your mouse over represents.
I suppose that, if a cell is uncolored and unassigned by you yet, mousing over it could assume that you want to know what happens if it becomes green (and offer no way to test what happens if becomes red).
Great stuff there!
Been playing some Hearthstone recently, and thinking about MTG.
Part of the reason those two games have mana curves (MTG implicitly, Hearthstone explicitly and linearly) is to ensure a decisive result. We swing bigger and bigger until one person is knocked out (as opposed to just chipping away).
That makes the games more dramatic and decisive.
The secret key is:
origin ensemble
But demonic ritual also works so as not to frustrate people who solve the anagram.
Yes, that kind of sharing is encouraged!
And if you're communicating with a friend directly (like by email), please feel free to pass the SECRET KEY on to them.
The only thing recorded right now is how much they buy in for and how much they cash out for. I'm tracking whenever money moves in the game, but that's it. This ledger allows us to recalculate a player's balance from scratch using their full money-movement history (and I actually do this, as a sanity check, before letting them withdraw or even buy-in to a game).
The server does log text messages about each move made, but this is just for debugging later (like, if someone gets DENIED BY SERVER I can manually look at these messages and see who made what move at what time to trigger it). Those logs are discarded after a few weeks and wouldn't be useful for reconstructing games automatically.
Storing move lists and chip movement wouldn't be too hard on the server---I'd just have to come up with a format for it---and that would really be the way to do it. Playback on the client would require several weeks of work, though, I imagine.
In The Castle Doctrine, it was a core part of the game thematically (watching security tapes), so it was worth it. It was also necessary to document exploits, because everything was happening on the client.
In Primrose, it was likewise necessary to prove and verify high scores that were submitted (that was the initial motivation, until I discovered that the recordings were really interesting to watch).
Here, where everything happens on the server, there's no security motivation, and there's also not a particularly strong thematic motivation. In fact, the fleeting nature of the games adds to the experience thematically, I think.
As a lone developer, I've got to pick my battles.
I also believe that a third-party tool could do this pretty well, either by parsing the recordedGame files or by acting as a proxy to scrape the protocol directly. Not so easy to have the playback happen directly in the client, but I don't think that's necessary. Maybe web-based playback. Maybe the third-party tool could automatically upload the recordings to the web. By scraping the protocol live, a web-based spectatior mode could also be possible.
claspa, yes, this can be adjusted with a simple server setting change. I could make it rise every two rounds or every three rounds, though I imagine this would be more confusing to keep track of for new players.
Every 2 rounds would be good if I ever implemented the "flipped board" thing (so both players get to play each board from both perspectives to mitigate slight advantages that the random boards might confer on one player). That way, both players get to play the same board from both rotation perspectives with the same ante.
I can confirm that the mystery player is NOT me.
There's a lot of development overhead in supporting client-server game history review. You know, menus in the client, speed control buttons, and all that. It really feels like a "someday" feature, like if this game really takes off.
I'm experimenting with something new.
Instead of antes that are fixed at 3 coins or any other number, antes now start small and rise in each new round.
These settings (starting point and rise rate) can bet tweaked on the server, but they currently are:
--Ante starts at 1 coin
--Ante increased by 1 coin each round
I just played a few games with the new setting, and it adds a very dramatic feeling as the pressure builds for both players.
Furthermore, it adds texture to the game as the strategy balance changes on a curve. In the beginning, betting skill matters the most, and picking skill is less important. But several rounds in, the picking skill becomes more and more important.
I believe this increases the skill depth of the game, because the necessary skill changes across a game. Expert players will learn to manage early vs. late game, for example. Staying at a table longer is very different from hopping tables. And so on.
I've talked to several poker players about the ante, and they ALL say the ante is too big already (at 3, or even at 1), and that this "reduces skill" in the game.
But poker players have a strong interest in their POKER skills, and having those skills translate to this game. The best players in this game so far include quite a few poker players (including the current Elo leader).
When the ante increases, their poker skills (betting skills) become less important.
But, I've come to realize that this isn't the only skill in this game. That's kindof the point.
On one far end, you could have a 100 coin ante and play one round. In that case, the skill of picking would be the only skill on display, winner take all.
On the other far end, you could have a 0 coin ante, and your skill in the first pick wouldn't matter at all. You could just keep folding for free until you finally got a good first-round number by accident, and then employ your betting skill. Rolling dice to pick in the first round would then be totally viable---after enough rounds, you'd have a guarantee of eventually getting a high number on turn 1.
By settling on a fixed ante, we're essentially saying, "this is the right balance between these two skills." By "right" I guess we mean "most interesting." But it seems like having the balance between these skills change over time is actually even more interesting than a fixed skill balance.
Finally, it makes sense for the picking skill to start diminished (because you know nothing about this opponent) and grow over time. A skilled player SHOULD be able to read their opponents picks better over time, and that skill should be rewarded. I feel like I've been in lots of situations where, after several rounds, I could read my opponent's picks like a book, FINALLY, but have them just fold away against my superior picking skill.
CANDYMAN
Drop one of those recordings into playbackGame and then run the game. You can watch your every mouse move.
If this is boring/slow to watch, you may want to enable "enableSpeedControlKeys" in the settings folder. Then ^&*() keys control framerate, wtih & being normal, ) being fastest. That can help you speed through some of the more boring parts.
Matches are currently NOT recorded on the server. I have thought about that, because that was an important aspect of some of my other games (Primrose and The Castle Doctrine). I've also thought about spectator modes. My instinct is to keep the game simple, though.
Okay, fixed. If both players are past the deadline when one reconnects, the pot will be split.
Blinds in heads up poker are currently smaller than this game. Big blind is 1/100 of the max stack size, small blind is 1/200 of the max stack size. With the new 3-coin ante in this game, the ante is 1/33 of the max stack size. That is appropriate, though, because there's a core picking skill in this game in the place of the random number generator that drives Poker. Still, antes might not be large enough here.
Increasing blinds at one table, gradually as the table lives on, are something to consider as the main mode for the game. Still, the bigger the blind relative to the stack, the shallower the stack effectively is, meaning betting skill gets reduced as this happens.
For tournaments, this is fine, because remaining players have stacks that have grown from beating other players who are knocked out (so as the blinds go up, the stacks are growing to match them).
So, if we had rising blinds at a normal table, what would we be saying? Yes, games would speed up, but we would also be saying something like, "Betting skill matters at the start of a table, but gradually matters less and less as the table lives on and the blinds grow." Picking skill would dominate at the end. Maybe this makes some sense, because by the end of the table, you should be able to read your opponent's picks much better. Still, it seems like a very delicate thing to balance. It's much easier to balance the game by picking a good ante.
The obvious thing to do would be have the ante increase by one coin per round. Starts at 3, goes to 4, 5, 6, etc. This is linear growth, but the relative growth to the last ante is decreasing non-linearly. I.e., the ante going from 3 to 4 is a much bigger deal than the ante going from 19 to 20. This may also make sense, because what you learn about reading your opponents picks will decrease over time as you settle in with them.
So... why not start at 1 coin, and go up by one coin each round?
This would certainly make the game more dramatic as it continued (as opposed to more and more of a grind).
Yeah, thinking about a way to fix this.
Well, that's a great idea for someday, but right now, I think it would shrink the pool of players too much.
Pilam's issue has been fixed on the server.
Also changed it to return pots to both players if both time out together.
He never drinks the whiskey....
Assuming that it lasts more than 60 seconds, whenever the database comes back, a flush will be triggered on the server, which will notice those games as timed out for both players, and whoever created the game gets kicked first, with the pot going to the other player.
This doesn't make too much sense, though, and I'll fix it to split the pot in that case.
There's also the chance that one player is able to connect again BEFORE the flush, and in that case, the request of that player will time the other player out, and the player that connected first will get the pot.
There's not a good way for the server to differentiate between this occurrence and the real deal where the other player has timed out. You know, if we've heard from A recently, but not from B, and the deadline is up, we assume B timed out, and give the pot to A.
Yeah, Josh, that should be fixed so the game count doesn't go up until you end a game. Thus, that exploit should have been impossible in the test3 tournament (though you could have still been imagining that it was possible by watching the wrong number go up and thinking you were playing someone that you weren't actually playing!)
I've thought about these kinds of "forced" match-ups for the top players.
The problem is that the skills involved in winning such a tournament would be quite different from the skills involved in doing well at the main game. It seems sensible to have "profit" be the way you win a tournament. An important skill in the game is knowing when to leave a table, for example.
There's even a song about that....
Yeah, MySQL was hosed today on the server. Affected The Castle Doctrine too around the same time.
Pilam's issue was a real bug and happened at a different time.
Yeah, I just changed this. Now if there's not at least one coin matched above the ante, there is no tribute taken.
If you bet your side up to 10 and your opponent folds, you get your 7 back, your 3 back, and you win their three.
If you bet your side up to 10, they bet their side up to 4, and then they fold without matching your 10, you get your 7 back, your 3 back, and you win 3 of their 4, and one goes to tribute.
Yeah, there's no requirement to play any games at all, and it is possible to win a prize by doing that if lots of people go negative. That's the nature of this game: net $ is king.
But, people are motivated to play games, because going up 1 step in the ranking multiplies their prize by 1.5
The number of games you play has no bearing on your prize or how ties are split.
I've fixed a few things about the tournament structure.
1. Prizes for ties are now more sensible (players split the total prize sum at their tied levels evenly).
2. The Games Played stat on the leaderboard only updates after you finish a game (so you can't tell who you're playing by watching that stat for upticks).
I've launched a new test tournament with a 10-cent entry fee. It will start in about 3 minutes and run for 9 hours.
Hypothetical prizes can be previewed here:
http://cordialminuet.com/gameServer/ser … players=20
The leaderboard can be seen here:
Well, I'm not sure I *want* an effect in this case in terms of player behavior. Do I want players folding less? What would that mean for the game overall?
I was trying to do something more subjective and immeasurable, which is to rebalance the two skills in the game, putting slightly more focus on the picking skill and slightly less focus on the betting skill.
The complaint from more skilled players was, "When I'm up against a conservative player who folds a lot, the game is a slow, boring grind, one chip at a time, even if I'm out-picking them." Perhaps the increased ante hasn't changed the prevalence of conservative, fold-heavy play, but it has changed the result for the player on the other end, because you can chip away at that player in only 34 rounds now instead of 100 rounds, especially if you're skilled at picking.