No, this was the early game and I was very focused on the recording. My plan is to finish my current game (my best so far) then do multiple runs up to h:96 for early game data.
So, my first 768 game ever has become an epic 1536 game! Currently good for high score #16 on Gamecenter, including the 5 (at least) hackers at the top. I have more screenshots of points both before and after getting 1536 but I don't want to clog the thread. PM me if you're interested. Corner strategy was used to great effect here, and I feel like I had a simpatico with the RNG. There were a couple times I got unwanted bonus white cards (one a 6, one a 96) that I thought would sink me, but they turned out ok. The RNG turned on me in the end with a slew of 1s I couldn't recover from. Here's a tip: don't get buck fever. When you have an opportunity to increase your high card, you don't have to take the shot immediately. See what other cleanup you can do on the board before making the big move.
Not sure if the developers are checking this thread and pretty sure this has already been mentioned, but every other time I open Threes there's a long white loading screen and no sound. Reliably, when I quit--close--reopen the game, the loading time is significantly shorter and there is sound. Just thought I'd report. Playing with the latest os on a 4s phone.
Firstly, congratulations! Secondly, we have a recording of that epic game (sequence #8) The weirdness in the beginning is still confounding me but a lot of the patterns we have seen thus far seem to hold even with a high card of 768.
I'm still mostly falling apart once I'm forced to abandon the corner strategy. I can get a 384 pretty easily, but I end up with a 96, 2x48 (separated), 3x24, 3x12 and my board is choaked and I die. 11,000 is my best score.
big big big congrats czech! (and also y2kmp3!) that's how my first (and only) 1536 came about as well. It was my first 768 and i just kept going and had a similar feel with the high card spawns. I consciously bet on the high card spawning in the corner when i was nearly down and out, betting it'd be a 48, and it was! it spawned right next to another 48 and i cascaded into another 384!!! after playing for long enough, you start to get a sense that "you're due" for a high card. that sense was tingling and i trusted it. i think it's a big part of crafting your own luck in threes. anyway, couldn't be happier you guys are seein' ol' terrence. if you have a twitter, be sure to tweet it! then let @aeiowu or @ashervo know of your twitter handle in a separate tweet so we can be sure to see it.
When I first played the game I wasn't very good at it, at all. My scores were pretty low, especially my very first game. So now this has led me to chase the challenge of lowest score ever. I think there should be a leaderboard for that, it's tricky and a totally different way to approach the game. Sorry if was already suggested, didn't have time to read the entire thread, getting pretty large now. BTW, awesome job on the game devs!
aeiowu, Thank you for the kudos! It means a lot coming from the game's creator! Notwithstanding GC, do you know how many players in the world have made it to the elusive 1536 tile? Anyone reached even higher? I am referring to legit trials, of course.
Was about to give up on this in disgust with myself today after getting steadily worse, but I seem to have hit a purple patch and I've had a couple of 768s today. Now I'm determined to go for the 1536. I think you guys are crazy with all your logging and trying to figure out the size of this apocryphal draw pile, but I respect you for trying and I'm glad you're enjoying it!
For your enjoyment the "Whalend" 3072 tile card! I believe this is the first time that such card is seen in a genuine (legit) game. Here are the screenshots showing the critical step that made this run possible and the tile card itself: I was compelled to try for the 3027 tile card after seeing a few players here making it to the 1536 tile card. Some notes of observation: 1. Like my earlier run of the 1536 tile card, this was only possible because I received a 192 tile card at a critical moment in the game. Had I received a card of a different value, I am not sure if the game would survive. 2. I do not believe keeping the "chain" of tile cards (in my case, 192, 384, 768, 1536) to be assembled horizontally or vertically is a good strategy. Instead, a better strategy is to keep these tiles connected DIAGONALLY. This is because such arrangement allows two degrees (instead of just one degree) of freedom for each pair of tiles to assemble incrementally. 3. I am convinced that it is necessary to "leverage" the bonus "high number" tile card to build your chain. The grid is just too small for you to build up the entire chain from scratch yourself. Remember that, in order to assemble the 3072 tile card, you will need to have one of each of the 192, 384, 768, 1536 tile cards at a minimum. There are only 16 spaces in the gird, 4 of which are taken up by these tiles, leaving you with room for only 12 tiles to build your chain. 4. One aspect that I find less enjoyable is the significant amount of chance you need to be in your favor to make progress at this high level. It is very easy to lock yourself in a position in which high number tile cards are staggered between low number tile cards. If this happens, it is very difficult to get out of the jam. A couple of features I like to see: 1. An undo button. I understand that this feature can be abused to "peek" at the next card. One solution to guard against abuse is to limit its use to once per "card stack" or "time period". This may then be seen as a "reward" to a player to "rewind" a mistake. Right now, it is just too punishing. 2. A replay button or a log of your gameplay. Since the tile allocations are randomized to an extent, no two games will ever be alike. I think it is a good idea to give the player a way to "save" their good game to be played back like a time lapsed movie. Since the game is so difficult at this high level, I think many gamers will be curious in seeing a successful run from beginning to end. Once again, I am tempted to "retire" this current game Hate to see it end.
Spoilers ahead! Spoiler Okay, so the 4/4/4 stack shows problems in the sample sequences which is why I have taken out the line breaks in the formatted output since they were occasionally misleading. The one thing I am absolutely certain of is this: if there is an undrawn stack then it contains an equal number of 1s and 2s. That's the whole point of using such an algorithm in the first place: to keep a critical balance without which the game wouldn't work. To this end, I have analyzed the sequences in regards to their balance between 1s and 2s in the following way: 1. Initialize a balance-variable with 0 (perfect balance). 2. Go through a sequence card by card. If you see a 1, increment the balance-variable by 1, if you see a 2, decrement it by 1. I then gathered all these balance-values and looked for their maximum value (regardless of sign) and their average over an entire sequence. This is what I found: Balance of blue/red cards in sequence #1: max: 4 | avg: 0.373913043478 Balance of blue/red cards in sequence #2: max: 3 | avg: -0.115044247788 Balance of blue/red cards in sequence #3: max: 4 | avg: 0.837301587302 Balance of blue/red cards in sequence #4: max: 4 | avg: -0.0899653979239 Balance of blue/red cards in sequence #5: max: 4 | avg: -1.0 Balance of blue/red cards in sequence #6: max: 4 | avg: -0.0786516853933 Balance of blue/red cards in sequence #7: max: 3 | avg: -0.427983539095 Balance of blue/red cards in sequence #8: max: 5 | avg: -1.63351498638 Balance of blue/red cards in sequence #9: max: 3 | avg: 0.280373831776 Balance of blue/red cards in sequence #10: max: 3 | avg: -0.392996108949 Extending this idea further, I went ahead and inserted possible stack-cutting points into the formatted sequences ('||' in green) where the balance-variable is 0. I have now developed two new hypothesis: 1. The number of blue/red cards in the undrawn stack might not be fixed at 4. It could occasionally be 3 or 5. 2. Similarly, it could very well be possible that 3 is also a possible bonus-card. This would allow the following stacks (blue/red/white): (3/3/3), (3/3/4), (4/4/4), (4/4/5), (5/5/5), (5/5/6) and very rarely (3/3/5), (4/4/6) or (5/5/7) if there are 2 bonus cards. I would encourage you to look through the sequences for supporting or contradicting examples and share them with the team.
Congratulations! I'm afraid that at this point you'll have tro go for the 6144 tile and THEN you can retire
Thanks for the congratulations, and congratulations right back to y2kmp3 on his 3072! Spoiler No (4/4/3) or (5/5/3)?
Spoiler I'm currently working under the assumption that there are at least as many 3s in a stack as 1s and 2s. But I am more and more prepared to give that notion up as well as I look more closely at the sequences. On a different, but not unrelated topic: this feels kind of like a detective story at this point. Running down leads, looking for evidence, eliminating suspects. In non-spoilery news: I got a Sixlock! I am now Prince of the Sixes ;-)
Good job with the sixlock! I tried going for a superlow score yesterday by actively avoiding building, then when that stopped working, I just started swiping in one direction until I couldn't go in that direction anymore, then changing directions, going clockwise. I ended up with only my 4th lowest recorded score. That was embarrassing.
Thanks again, aeiowu, for the kudos for reaching the 3072 tile card. I, too, was very surprised that I made it this far. Always nice to hear from the game's creator. Is the next tile card 6144 the LAST tile card in the game? What happens to the game when you reach this tile card? Does the game end? Further, what happens if you merge two 6144 tile cards (theoretically speaking)?
