Subject: vavava and 12X0
Date: Wed Oct 3 12:39:54 2018
User: jamesblackburn-lynch
Message:I was just looking to see the state of 12x0 and vavava has all of a sudden leapt into first place. I thought...hey...have they been asleep...but no...they won 55 (and counting) all today. They've played more than 10000 lifetime. I wonder what their previous high score was.
But...this is happening live. They are headed top 5. Will they make it to number 1?
James
Date: Wed Oct 3 12:50:36 2018
User: TNmountainman
Message:Uh...…….jinxed!
Date: Wed Oct 3 12:58:44 2018
User: jamesblackburn-lynch
Message:Well...they got to 56. So does that count as a jinx?
Also, TN, that is you at #1 isn't it?
James
Date: Wed Oct 3 13:03:07 2018
User: TNmountainman
Message:It counts as a jinx because as soon as you mentioned it, he(?) lost. And yeah, that's me at #1. Although I only check in on that list rarely. I like 12x0, but it's a cruel variant. But I don't think my mark will stand forever - although it's been there a long time.
Date: Wed Oct 3 13:16:55 2018
User: jamesblackburn-lynch
Message:Tn,
Well, going by free@last's data those 12x0's are winnable about 15/16s of the time.
So, the probability you can actually win 66 in a row is 1%. It really does seem like someone should get that lucky soon.
James
Date: Wed Oct 3 13:29:58 2018
User: TNmountainman
Message:Yeah, it takes care, as you well know, but it's mostly the odds, if enough good players were to take it on, 'til someone takes it down. I think "diederik" gave it several good goes. And that is NOT meant as a challenge! LOL.
You may not recall, but when I was setting that mark, I sort of gave an update every once in a while, playing games sparingly. I think I "rested" several months 1 short of the mark before I gave it an attempt. I was relatively new to this site at that point, and I don't think completely realized how 'lucky' I was to get that many winnable ones in a row. I'm sure I hadn't played remotely near 100 'streaks'-worth of attempts.
Date: Wed Oct 3 23:11:15 2018
User: TNmountainman
Message:Further info.....
Perhaps I misunderstood your initial comment, james. It appears that today's 56 was a *daily* record, besting his own 53 from Jan. 27, 2014. I think it likely, however, that you were referring to the all-time record since you referenced my effort. Still, a very noteworthy accomplishment.
Date: Thu Oct 4 00:43:47 2018
User: jamesblackburn-lynch
Message:No, you followed the first time. I didn't notice the daily best. Very impressive.
James
Date: Thu Oct 4 02:48:11 2018
User: HopDiriDiriDattiriDittiriDom
Message:James,
How do you calculate the probability you can actually win 66 in a row?
It can be calculated precisely but I tried MonteCarlo simulation again to see what it is approximately and it seem to be about 0.01% (not 1%).
I have coded 1,000,000 experiments and here is a partial list of how many times you see a streak length equal to a number out of 1,000,000 experiments (i.e. streaks):
60: 266 0.00026
61: 240 0.00024
62: 214 0.00021
63: 160 0.00016
64: 136 0.00013
65: 108 0.00010
66: 114 0.00011
67: 83 0.000083
68: 67 0.000067
69: 66 0.000066
70: 47 0.000047
The more the number of experiments the more accurate it gets. For example for 10,000,000 experiments:
60: 2595 0.000260
61: 2137 0.000214
62: 1830 0.000183
63: 1547 0.000155
64: 1395 0.000140
65: 1137 0.000114
66: 980 0.000098
67: 796 0.000080
68: 762 0.000076
69: 585 0.000058
70: 494 0.000049
Date: Thu Oct 4 04:05:00 2018
User: TitanicTony
Message:Hop, I think that what you show above is for random deals, but ours are far from random. We have a fixed set of 32768 deals at 6 levels of "difficulty" (tweaking, levels 5-10). So, the chance of getting 10 winnable deals in a row (at level 5) is much much higher than what is expected for random deals. Also, at each level (5-10) it is known (by free@last) exactly how many games are not winnable (according to his solver), so the probabilites for each of the 6 sets of 32768 deals could be computed exactly.
Date: Thu Oct 4 05:50:38 2018
User: HopDiriDiriDattiriDittiriDom
Message:TT,
My code exactly simulates what happens when you pass the levels. I used the stat given by free@last:
1841 12x0-5.txt
2387 12x0-6.txt
2973 12x0-7.txt
3470 12x0-8.txt
4201 12x0-9.txt
4769 12x0-10.txt
So what my code does is to flag 1841 deals of 32768 for level 5 as an example. Then a random function generates a number between 0 and 32767. If the generated number is one of those 1841 deals your streak ends othwerwise you are assumed to win another game. After wining 10 games your level becomes 6 and so on until you win 50 games. Then you are at level 10. And you continue at level 10 until you hit one of those 4769 unwinnable games.
I am pretty sure my code simulates it correctly and my results are very close to the reality.
For example my MC simulation gives the average streak length as 13.58033150.
And the exact answer (calculated analytically) is 13.58014937978973116349390333
Date: Thu Oct 4 06:49:22 2018
User: TitanicTony
Message:Excellent!! I'm impressed, :)!
Date: Thu Oct 4 09:54:14 2018
User: jamesblackburn-lynch
Message:Hop,
I looked at free's data and estimated that 15/16 games are winnable. That's not as precise as your method since 1) it's an estimate and 2) I didn't account for the change at game 50 when you go to level 6.
But I don't understand how we can be so far apart. I just did (15/16)^66=.0141...
Every time you play you get a random deal. With repeats. So the probability of winning each time is approximately 15/16 (I tried 32768 and 1841 just to see if I got me closer but it doesn't change it much...certainly nothing like the factor of 100 we are differing by). Independent events. So multiply.
But I must be forgetting something...no way we can be that far apart.
James
Date: Thu Oct 4 10:25:39 2018
User: HopDiriDiriDattiriDittiriDom
Message:James,
I think your estimate of 15/16 is not right.
I find a different result if I use your method:
? prod(i=1,5,((32768-[1841,2387,2973,3470,4201,4769][i])/32768.)^10)
%12 = 0.008420964462977733530330695318 // probability to win 50 games in a row
? ((32768-4769)/32768.)^16
%10 = 0.08073862879713230735141486522 // probability to win 16 more games at level 10
? %9*%10 // product of 2 results above
%11 = 0.0006798971238902018322862056367
Not so close to my simulation result but much lower than your estimate.
Date: Thu Oct 4 10:53:38 2018
User: free@last
Message:I get about .1% but please remember my data is only a upper limit on the number of unsolvable games (although it's likely with 10% or so).
Date: Thu Oct 4 12:45:03 2018
User: jamesblackburn-lynch
Message:There it is! I for some reason was thinking you only get to level 6 when you get to game 50! I don't know why I thought that. My 15/16 is for level 5 games which is only for the first 10. I tried again, lazily, using the average of the number of unwinnables Hop pasted above from free's stuff (I went with 3200 unwinnables) and I found my number was, like free's, about .1%.
So still no agreeing with Hop's experimental numbers, but at least a factor of 10 closer.
Let's try it the not lazy way:
1841 12x0-5.txt
2387 12x0-6.txt
2973 12x0-7.txt
3470 12x0-8.txt
4201 12x0-9.txt
4769 12x0-10.txt
((32768-1841)/32768)^10*((32768-2387)/32768)^10*((32768-2973)/32768)^10*((32768-3470)/32768)^10*((32768-4201)/32768)^10*((32768-4769)/32768)^16=0.00067989712
Okay! Now we are close. Still much larger than Hop's number but getting there.
In the end, then, TN, it is less likely than I originally thought for someone to get your high score, but still...it should happen. People play that all the time. vavava needs to try again.
Thanks,
James
Date: Fri Oct 5 09:49:09 2018
User: HopDiriDiriDattiriDittiriDom
Message:James,
0.000679897 calculated in my previous post seems to be correct. I have found my error in the simulation code. I didn't count the streaks longer than 66 but all of them has to be included. Now my simulation code outputs very close results:
70: 494 0.04% (0.000360700)
69: 585 0.04% (0.000419200)
68: 762 0.05% (0.000495400)
67: 796 0.06% (0.000575000)
66: 980 0.07% (0.000673000)
65: 1137 0.08% (0.000786700)
64: 1395 0.09% (0.000926200)
63: 1547 0.11% (0.001080900)
62: 1830 0.13% (0.001263900)
61: 2137 0.15% (0.001477600)
60: 2595 0.17% (0.001737100)
If we calculate for 70:
? prod(i=1,5,((32768-[1841,2387,2973,3470,4201,4769][i])/32768.)^10)
%9 = 0.008420964462977733530330695318 // probability to win 50 games in a row
? ((32768-4769)/32768.)^20
%10 = 0.04303796338773185567603679114 // probability to win 20 more games at level 10
? %9*%10 // product of 2 results above
%11 = 0.0003624211602470267433145490821
Date: Fri Oct 5 10:16:23 2018
User: jamesblackburn-lynch
Message:Hop,
Yes, I think your actual calculation (as opposed to the experimental evidence) and mine are the same but you wrote yours in some code (MatLab?) so I wasn't able to figure out what you were doing until I did it myself. For some reason, my final answer has disappeared, but it is the same.
Funny that we were both doing it wrong! But now we have it.
James
Date: Fri Oct 5 10:26:07 2018
User: TNmountainman
Message:Just from anecdotal evidence, it appears perhaps(?) there's something in Denny's algorithm that makes going from climate 9 to climate 10 more 'penalizing' in 12x0. Note that many players (18) have streaks from 50-58, but that 66 is an outlier. One would expect, though, if that were the case, that the same effect would be seen in 11x0, or 10x0, but that doesn't seem to be visible. One *could* make an argument it's there in 13x0, but the evidence is not clear. There's a gap of 20 between #4 at 168, and #3 at 188. (And my understanding is that there's considerable doubt as to the veracity of all of tiorapatea's numbers - and indeed it stands out like a sort thumb in this case.) Or...…...it could just all be "statistics". lol.
Date: Sat Oct 6 03:54:23 2018
User: HopDiriDiriDattiriDittiriDom
Message:James,
That is PARI/GP not MatLab.
It can be made one-liner function like this:
pf(n,v)=my(s=5+if(n<50,n,50)10,d=2.^15);prod(i=5,s-1,(1-v[i-4]/d)^10)*(1-v[s-4]/d)^(n-if(n<50,(s-5)*10,50))
then the probability to have a streak of 66 in 12x0:
? pf(66,[1841,2387,2973,3470,4201,4769])
%110 = 0.0006798971238902018322862056367
similarly the probability for a streak of 500 in 11x1:
? pf(500,[34,70,93,140,186,200])
%116 = 0.05420799013584287729137429548
That's much higher. So it's easier to aim for the top position of 11x1 instead of 12x0.
Date: Sat Oct 6 20:32:40 2018
User: jamesblackburn-lynch
Message:I don't agree that it follows that it is easier. All our work assumes that every winnable game is actually won. But, of course, that's not a reasonable assumption.
To win 500 11x1's would take more than just getting lucky enough to hit 500 winnables in a row. It would take a lot of concentration. And no bad luck. To get 66 winnable 12x0's is less likely but a lot less concentration would be required.
I'm going to go out on a limb and, in writing, predict the 12x0 top spot falls before the 11x1 does. Obviously, based on nothing.
James
Date: Sun Oct 7 10:22:59 2018
User: HopDiriDiriDattiriDittiriDom
Message:Well, easier may not be correct word but it certainly requires more effort to reach the top position in 12x0 than in 11x1. How is that?
? pf(487,[34,70,93,140,186,200])
%4 = 0.05869867498217967436320104432 // probability for a streak of 487 in 11x1
? 1/%4
%5 = 17.03615967998579036235232691 // that's about 1/17
? pf(67,[1841,2387,2973,3470,4201,4769])
%6 = 0.0005809460318543017914484091681 // probability for a streak of 67 in 12x0
? 1/%6
%7 = 1721.330287441906030964818808 // that's about 1/1721
? streaklen([34,70,93,140,186,200])
%8 = 186.9764261099795094312181890 // average streak length for 11x1
? streaklen([1841,2387,2973,3470,4201,4769])
%9 = 14.58014937978973116349390333 // average streak length for 12x0
? 1721*%9
%10 = 25092.43708261812733237300762 // requires about this many games in 12x0
? 17*%8
%11 = 3178.599243869651660330709213 // requires about this many games in 11x1
Date: Mon Oct 8 13:10:41 2018
User: jamesblackburn-lynch
Message:Makes me wonder...how come you don't play anymore, TN?
James
Date: Mon Oct 8 13:43:40 2018
User: TNmountainman
Message:Sigh......well, I do - a little. I've been playing a little bit of 10x1 and 12x0, but I guess "life circumstances/exigencies" is as good an answer as any. It's complicated. Playing the SSC and WWC were exhausting at times. I wish I had been able to experience the tournaments when the ratings were still extant, but oh well......
Also, and importantly, as I've said before, when I learned of the existence of solvers, some considerable amount of the joy of this site left for me. Then when I discovered some were blatantly cheating, and therefore the records were not sacred nor reflective of relative skill, an even more considerable degree of enjoyment was killed off. Very sad. Quite analogous to how steroids damaged baseball, which point I've made before. Imagine if a cheater at some point knocks off 67 or more straight 12x0 games. It just ruins the whole thing.
But I really like the community of people on this site, and am thankful I've met a few in person, so that part is still nice; and clearly there is a good reservoir of intellectually-gifted sorts here, and so I get considerable mental stimulation from that, which is perhaps the strongest draw for me now.
Date: Wed Oct 10 23:46:31 2018
User: jamesblackburn-lynch
Message:Well, I wish you would come back. I hate to see the cheaters above your name. But they wouldn't be if you played.
Date: Sun Apr 21 16:58:04 2019
User: jamesblackburn-lynch
Message:Called it!!
"I'm going to go out on a limb and, in writing, predict the 12x0 top spot falls before the 11x1 does. Obviously, based on nothing.
James"
James
P.S. Woohoo!
Date: Sun Apr 21 18:40:16 2019
User: TitanicTony
Message:CONGRATULATIONS, James!! I'm very impressed!!
Btw, I'm pretty sure there are no "cheaters" ahead of TN in 12x0!!
I'm happy to be tied for 3rd, on 58!
Date: Sun Apr 21 19:06:33 2019
User: ix
Message:i think you somehow used the knowledge that the game has 32,767 brothers and was dealt on a sunday to your advantage. i wish there was a way to know how long a record stood, like start and stop dates, that would be fun. congrats, dude, that is amazing.
Date: Sun Apr 21 20:59:39 2019
User: ix
Message:i guess you'd only need the date the record was set, duh
Date: Mon Apr 22 03:09:50 2019
User: TNmountainman
Message:Congratulations, james. As ix said, that is amazing - dude. And yes, you called it - altho it must be pointed out that it wasn't an 'experiment' with a control, as (to the best of my knowledge) you weren't also mounting an attack on the 11x1 record. Or maybe you were and just didn't broadcast it. (Upon further perusal.......I see you've chalked up 406 there, well within striking distance, so I guess maybe I'm wrong about that.) That said, had this been a bet of some sort, I'd have taken the view that the 11x1 would fall first.
I also see big time gaps before and after that particular game (13/17). And I *know*, with that much at stake, you didn't just blithely blitz thru it in 2:35. Or again, maybe you did. I'm just imagining you, terrifyingly hitting "Play", and then stopping it, and staring at the tableau for a while, just to make double- and triple-sure.
As to ix's question about how long records stand, I just went back to the thread in which I tried to jinx myself when in quest of the previous record of 58 (see link from Oct. 8, 2018 above). Looks like I tied the old record of 58 on Oct. 6, 2006; then waited 111 days, and got #59 on Jan. 25, 2007. So we can say this record stood for over 12 years. Looks like I didn't post when I won #66, nor lost game #67, but it was likely within days, or more likely weeks, of that date. I know I did *not* start playing quickly again, but didn't wait huge chunks of time, either, after #59.
NOW...................I'd love to hear some quasi-theories on why 'our' method works. Seems as tho no rational mechanism would be plausible or possible, and yet I think this begs the question of non-coincidence. Just something to ponder......
Date: Mon Apr 22 04:45:13 2019
User: HopDiriDiriDattiriDittiriDom
Message:I guess this dealer not perfectly random. There are some days it's serving some "easy" deals more and some other days "hard" ones more.
James is probably having a feeling of it and giving it a try. I also try it. I have various observations to smell it.
As I said in other thread I use 8x4 to get a feeling of whether the dealer is going through a period of "easiness".
Someone else had said somewhere that he looks at daily scores table to see if it is an "easy" day today.
I also try to get a feeling of it by checking the tournament scores. Check for example the "threat" or "speulenker" scores. Somedays people reach the winning score in a very short time but some other days it takes more time to win them.
Date: Mon Apr 22 05:24:31 2019
User: TitanicTony
Message:We should keep in mind that the "dealer" isn't really dealing, but selecting a game (at random) from a group of 32768 that were dealt a long time ago (1996?). There, of course, being 864 such groups -- one for each varient (72) and difficulty level (12).
Date: Mon Apr 22 14:26:35 2019
User: jamesblackburn-lynch
Message:I don't believe in the theory of "easy" days. For this, with a heavy heart, I present evidence. The day I won that last 12x0 (I really did just win it that quickly...that was Easter Sunday so I was at a family thing then had a few minutes before giving a review session for Calculus II and thought I'd start the game knowing I could save it and come back to it if it was scary, but it was very easy and even with the panic it was no problem), I celebrated then played a few more games and then played my daily 11x1.
As TN noticed, I had just tied for #2 all time. I was really hoping to have a good shot at being #1 all time there too. But...despite how "easy" the dealer was treating me...I hit an unwinnable 11x1. So sad.
Particularly annoying is that I am now tied for #2 in 11-1 and for #8 in 10x2. Literally tied. Both times I had finally reached someone and lost on the game in which I would have passed them.
James
Date: Mon Apr 22 14:40:04 2019
User: The_Interpreter
Message:Anything other than #1 is also-ran.
Date: Mon Apr 22 15:12:12 2019
User: Ellie2
Message:In many sports (tennis is my favorite), you can be a multi-millionaire as an "also-ran"!
Btw, imo, you are never an "also-ran" if you enjoy what you are doing, and are happy with the result, 🙂!
Date: Mon Apr 22 17:27:51 2019
User: The_Interpreter
Message:People still watch tennis?
Date: Tue Apr 23 02:08:17 2019
User: TNmountainman
Message:Sigh. She didn't say anything about watching. In fact, she used the word "doing". Seems like a very incorrect interpretation, unless I'm missing something. I would even go so far as to say your immediately-preceding comment is also an incorrect interpretation, altho one could make a case for that, I guess. For example, if Rafa or Roger loses in the final of one of the majors, I really don't think it'd be remotely accurate to call him an "also-ran". Technically, by some definitions..........*may-be*. In common parlance.........I think not.
And yes, people do still watch tennis, altho certainly not nearly as much as in the two glory periods. Kinda like the NBA, the NFL playoffs, March Madness, the W.S., etc..........it's more concentrated in the main, climactic battles. Sorta like the 1% in the financial world - altho not exactly, of course. (Sorry if that's ever-so-slightly stepping on your toes with that as a mild-and-rough 'interpretation'.)
Date: Tue Apr 23 10:48:09 2019
User: The_Interpreter
Message:She may be doing instead of watching, but is she making multi-millions doing?
Also-ran is a kinder way of saying LOSER.
Date: Tue Apr 23 10:55:21 2019
User: outskirts
Message:Happy millionaire tennis player sounds good to me.
Date: Tue Apr 23 13:19:34 2019
User: TNmountainman
Message:If she's "happy with the result", then "loser" also almost certainly would not apply. Maybe she's on the satellite tour, only making tens of thousands; or maybe she's (perish the thought) playing just for fun; maybe even not keeping score; maybe even giving lessons. She certainly doesn't come across as a loser to me.
*Not to mention*..........she may be talking about *others* who are also-rans. She could be Serena Williams (very odd choice to choose "Ellie2" in that case, but it's possible) and talking about her vanquished foes saying "it's ok that they finish second to me" - they're still being amply rewarded.
Date: Tue Apr 23 13:34:09 2019
User: The_Inquisitor
Message:Are you what they refer to as a politically correct, tree-hugging, nobody is a loser, everybody gets a trophy, let me control your life with entitlements and your mind with my way of thinking, kind of guy?
Date: Tue Apr 23 13:56:52 2019
User: Ellie2
Message:Good grief, is this all my fault? I play tennis for fun. And, it is a good way to keep fit. And, I'm not a millionaire; I pay nearly $500/year for my club membership. The club has 4 indoor courts, so winter is definitely included.
And, of course, I am not Serena Williams.
Date: Tue Apr 23 14:04:08 2019
User: TNmountainman
Message:No, I'm not - altho tree-hugger, yes. Btw, you left out A LOT of hyphens. And also btw.........I was a second-place finisher in a couple of tennis tournaments that I'm more 'proud' of than some first places in tournaments with weaker fields. So yes, altho I was technically a "loser", I won some challenging matches along the way, and won more than I lost.
Good to see you, rather than The_Interpreter handling this aspect of things here now. I think he was missing the bigger point, and interpreters shouldn't be in that situation. But I suggest you inquire more of Ellie2, to get her take on it.
But kinda like outskirts infers................the bigger tournaments have a 128-draw, not counting the qualifier rounds. So maybe 150+ participants, in general. Only gonna be one winner. 149+ "losers" (at some point in the tournament, even tho they may win many matches before losing). But probably the top 10 or so are gonna walk home (or fly home) with $100K or (much, in many cases) more. Do that a few times, and happiness may soon follow. Maybe you should try that. Inquisiting must get old, and certainly can't be much joy there. As if.........
Date: Tue Apr 23 15:01:13 2019
User: The_Interpreter
Message:The only reason losers walk home with $100k is because even worse losers pay to watch them.
Date: Tue Apr 23 15:07:42 2019
User: TNmountainman
Message:Now at least *that's* an interpretation, valid or not.
Date: Tue Apr 23 20:18:32 2019
User: The_Interpreter
Message:His launch angle was too high.
Date: Thu Jun 6 12:16:50 2019
User: HopDiriDiriDattiriDittiriDom
Message:As I said before I felt the dealer was going through some "easines" period and gave some shots to 4x8 and got 3 easy ones. I could go on but that was enough for today:
6/6 11:59 am 8x4 17115-10 Streak 2:54 Won
6/6 11:55 am 4x8 17781-5 Streak 1:42 Won
6/6 11:54 am 4x8 4135-5 Streak 1:10 Won
6/6 11:52 am 8x4 10571-10 Streak 1:12 Won
6/6 11:50 am 4x8 886-5 Streak 4:38 Won
6/6 11:42 am 8x4 6959-10 Streak 1:06 Won
6/6 11:41 am 8x4 4323-10 Streak 1:56 Won
6/6 11:39 am 8x4 15463-10 Streak 1:19 Won
That was enough to set HWM (if it runs today) for 12-sum competition.
Post follow-up