Subject: most prolific streaker
Date: Tue Jan 18 14:26:35 2011
User: huffy
Message:Wow this thread has developed quite a life of its own. At this point in time I regularly play 35 of the variants and am quite happy to stop there. I do consider myself to be a streaker and like to see those congratulations you win messages so variants where just getting dealt a winnable game is a major challenge hold little interest (do not take that as any sort of vote to remove unwinnable games, I am opposed to that) . Roo has pointed out to me quite correctly that should I want to win a masters (and I do) that I should be practicing and by that he was meaning getting more exposure to the 10sum variants. Winning a pile of 10*6 games the day before a masters is certainly not the right preparation. So at some point in time I may change that focus, but right now not interested. I have a lot of time restraints in my life and even here in freecell land battles on so many fronts. The most likely change of my focus will be should coffee1 get closer to me in the 10*6 to shift all my focus there, it’s the highest streak on this site and has to date been compiled on the back of once a week efforts, if I can still play this game the only way I will be passed is a crash and lets face it when I play that’s a possibility, I don’t loiter about when I play and speed is a major killer on the freecell highway.
In the past I have done some tables of my own across all the variants but unlike the generous top 1000 restricted it to the top 3. With 3 points for first, 2 for second, 1 for third. Where there were ties points were shared subject to only the above points being available to share and if the average went below 0.5 none of the sharers got anything. Eg 4*4 SirPape gets 3 for being first but with a huge pile of joint efforts for the remaining 3 points no allocation was given. I did add the scores from known multiple identity players like Ed & cxmiv but restricted that to a player can only have 1 score in a variant. In 8*4 RGK gets first was not eligible to get third as well so PhoebeDee gets promoted to third.
I knew I would be high up but was expecting major opposition to come from T1-T3 and Ez-Ed having little exposure to loskazanes he caught me quite my surprise at his/her score.
Those who score 10 or more points
loskazanes 22.5
huffy 18
jane30066 11
T1-T3 11
EZ-Ed (Multiple)11
SirPape 10.2
Deno 10
Jane’s position is also quite remarkable given how long she has been gone from this site. Last seen in 2007 and that is misleading in that she only played a few games and it was quite some time before that when she previously played. I did read that she was in very poor health and suspect that we lost a very good player.
Of the 35 variants I play I keep records of how many all-time player identities I am listed as behind (so am counting draws as names I am behind). The number is 1697. There are 1196 player identities in that 1697 score. Which shows most are ahead of me in just 1 variant. Of the multiples the top 3 are
canman1944 13
TitanicTony 11
Laffman 10
That was a surprise I expected it to be TitanicTony, T1-T3 & kenwa at the top. T1-T3 and kenwa both score 9. For the record Zonny and loskazanes both score 7.
Date: Tue Jan 18 15:28:53 2011
User: TNmountainman
Message:I had no idea how many variants I normally played until just now, when I counted them. Using this as my main nick, I have played at least one game of 34 variants, but of those, 9 were 'exploratory', way back when I first discovered there were others (besides 8x4) on this site, and I had no feel for relative difficulty levels, and have less than 10 plays in each of them. That leaves only 25 variants in which I have played some reasonable number of games. (Plus maybe a couple more with the two obvious alternate nicks I use) For me to invest the copious amount of time to make even a tiny noise in (at least the majority of) 50 more variants is just not feasible for me.
As I've stated on other threads, unlike huffy (no disrespect intended whatsoever), I get little or no thrill from winning a game in which 99.9+% of them are going to be won. I find it much more "fun" to win when winning percentages fall somewhere in the 30-95% range, with a real "sweet spot" in the 65-90% range. Enough sense of accomplishment when one wins, and enough challenge to make that in doubt......
Date: Tue Jan 18 17:05:38 2011
User: mightybass
Message:the reason i didnt want to use multiple nicks was because there is no way to verify that the streak was alive from one nick to the other. i could easily start a nick and run up a 100 streak, then lose a bunch with a 2nd nick, and then win another 100 with a 3rd nick. its just too difficult to confirm and therefore i don't feel multiple nicks should count towards one streak. they don't count in the all-time lists for individual variants, so they shouldn't count for this either. again though, probably something that will raise a large spectrum of opinions.
Date: Tue Jan 18 17:31:04 2011
User: TNmountainman
Message:I don't think anyone was suggesting that multiple nicks count towards one streak. I think what others posters, including myself, would suggest is that streaks in *different* variants, from alternate nicks, could count towards one *person*'s score, especially if it's obvious (and that's a slippery slope, of course). For example, maybe I might rank #300 in variant A as TNmountainman, and rank #400 in variant B. But if by chance I also played variant B as Dr.TNmountainman, and maybe ranked #250, then my 250 ranking in that variant could count towards my 'overall' ranking as TNmountainman. At least I think that's what Boojum, as one, was suggesting in that last post on the first page of this thread. But he is also correct in stating that that would probably be impossible to do in an automated calculation, and would just depend on the word of the participant/calculator, to be calculated 'manually'....
Date: Wed Jan 19 00:12:33 2011
User: EZ-Ed3-LAnseMI
Message:Not many folks play more nicks than I do. It would be too difficult to try to keep tabs on one player with lots of nicks. Thus, if something official comes of this, only a player's best nick should count. In my case, and surprising to me, it seems to be this nick, EZ-Ed3, for me.
I have lots of non-EZ-Eds that are my best in many variants. Probably the two that have done the best are ScaryGaryTheFairy (who recently re-claimed the top spot in 7x5 for my family) and SarahBarracuda.
Unofficially, I will continue to go through all the variants and figure out my score using all my nicks. I will also be exploring variants I have not previously played with this nick, EZ-Ed3. And eventually I will try and replace my better streaks made by other nicks using EZ-Ed3.
If I ever get all the variants played, I would like to see how close to 100 (mightybass' system) I can get my score. I'm sure I can top 200 and rather suspect I could best 150 given lots and lots of time.
And, huffy, I am very familiar with jane30066. In years gone by, jane, Loskazanes and myself have battled in several variants. I figured the 30066 was a zip code. I looked it up and it was in Georgia. She was a wonderful player.
Date: Wed Jan 19 09:19:29 2011
User: mightybass
Message:TNmountainman: after that explanation i did the typical "slap your forehead, duh" response. i should have interpreted the multiple nicks as that way, and not the way i assumed. it makes perfect sense and i agree that you should be able to count your best score from across all nicks to calculate your personal ranking (as opposed to combining them into one super ranking). i also agree it would be next to impossible to automate this ranking unless a database was created to house everyone's multiple nicks that they use. even then it would still be difficult. i also agree that a win in say 9x2 is more fulfilling than a win in say 13x3 or 10x6, but i still like using those variants in this calculation. a streak in the easy variants is still impressive as it takes not only a large time commitment which takes you away from playing other variants, but you also have to avoid messing up that long streak. all of us have probably played one of the easy variants at one point and made that one stupid move that ruins what should have been a guaranteed win. in my opinions there are certain variants and certain accomplishments that should hold a stronger weight and more bragging rights. however, for simplicities sake i chose to evenly weight every variant.
huffy: another interesting variation on ranking across all the variants. in this thread i have been excited to see everyone's opinions and ideas for how to rank. there are obviously infinite ways that we can rank ourselves, with weighting and amount of variants, etc...but it would be cool to see a couple of the better ideas get automated and that way the ideas could be more widespread than just this forum.
when i first started playing freecell online, i had come up with numerous goals for myself...beat huffy's 10x6 record, be 100% in every variant, have every streak over 100. as i started to play more, i realized that all of these goals were either impossible or would take me the rest of my life. then when breezing through one of the all-time lists one day i came across "top_100_in_all_variants" and i thought to myself, 'i could accomplish that goal. as i play more i realize this too is a highly difficult goal, but i also started to ask myself how one would rank someone who was top 100 in all variants. and 50 some posts later, here we are.
as far as my previous goals go, i obviously have not surpassed huffy in 10x6, i have played 56 of the 70 variants, 34 of those still have 100%, only 16 have streaks over 100, and only 5 are in the top 100. still a long way to go, but that is what keeps me coming back. happy playing to all!
Date: Wed Jan 19 09:43:52 2011
User: EZ-Ed3-LAnseMI
Message:mightybass--Good luck getting a streak over 100 in 4x4.
.wink mightybass
p.s. If your nick has anything to do with bass fishing, that was one of my passions when I lived in Iowa (basically my first 32 years). My career took me out of Iowa for good, but I will always be an Iowan at heart.
Date: Wed Jan 19 15:42:11 2011
User: Gulbis
Message:I compiled my rankings in all 70 variants (excluding 8x0 and 7x1) and I got an average of 263.37.
I guess I'm precisely the target audience for this because I like to play different variants and I've been consciously working to get into the top 100 in as many variants as I can. 24 done, 46 to go!
Date: Wed Jan 19 21:11:00 2011
User: Kaos
Message:I've dropped the 8x0 and 7x1 from my stats and also played some more 12x0, 9x3, and 4x10 to get the overall average down to 382.23 and the average for 48 variants played to 98.625.
Moving up the 9x3 sheet was fun seeing the names of numerous familiar streakers in Denny's list of people you'll pass with one more win. Lots of long-time stalwarts.
My average per columns ala CubicSprock's list:
4: 149.9 - it's an 8.0 if I drop 4x4
5: 11.5 - seems to be a friendly variant for me.
6: 84 - 11.2 except for 6x9, 6x10
7: 424
8: 660.75 - No plays in 8x5 thru 8x8 and using my original nick for the 8x4 streak
9: 332
10: 503 - Using ThunderRoad nick for 10x1.
11: 529
12: 699
13: 832 - Only played 13x0
I'm cutting some corners on 6x9 and 6x10 as I'm using the streak in 6x8 as the length of the 6x9 and 6x10 streak. On one hand, if you've won XXX games in-a-row in 6x8, it means you've also won that many in 6x9 and 6x10 but on the other hand, you're not showing up on Denny's tables that way.
Date: Thu Jan 20 02:50:49 2011
User: TNmountainman
Message:Kaos:
"I'm cutting some corners on 6x9 and 6x10 as I'm using the streak in 6x8 as the length of the 6x9 and 6x10 streak. On one hand, if you've won XXX games in-a-row in 6x8, it means you've also won that many in 6x9 and 6x10 but on the other hand, you're not showing up on Denny's tables that way."
TN:
Interesting point, that. It certainly makes sense..... However, it fails the aforementioned persistence/prolific test that mightybass has stated as one of the "three pillars" of streakology (streakonomics?).
This is a very clear illustration about what is being measured....
Is there a "point" to playing x number of 6x9s and/or 6x10s, when one has had a significant streak in 6x8? Well, yes, according to the working template so far put forth, because there could possibly have been some nasty 6x9 (Is there such a thing?) that could trip one up in that form, but not in 6x8, for some inexplicable reason. That as a secondary argument, but the persistence factor would be the primary objection.
In 6x9s and 6x10s, it's hard to see how that would be such a valid argument, since both must have easily 100% win rates for someone paying attention. However, consider the recent example of CubicSprock, recently winning 3 4x6s in a row (in a mere 14 plays). And yet, I just looked at his 4x7 stats, and he has only won 1 in a row there (in 11 plays). So, in the narrow variants, it's not at all a trivial thing to say winning x number in a row in a CxF (columns by freecells) equates to winning x number in a row in a Cx(F+1) array. (Not that CubicSprock couldn't easily win 3 in a row there, too.) But we also know that "luck" is one of the other three pillars of streakology. So while luck surely plays a part in getting 3+ in a row winnable 4x6s or 4x7s, it's persistence/prolific that gets the job done in 6x9 and 6x10.
Hope all that makes some amount of sense.....
Date: Thu Jan 20 03:04:47 2011
User: Gulbis
Message:Yeah, I don't think you should use a streak in one variant for a different variant. For example, my streak in 6x7 is 81 with a 100% winning rate but my streak in 6x8 is only 36 with a 96% winning rate. You can't make assumptions like that. If I have a streak of 500 in 9x3, does it mean I can also use the same streak for 9x4, 9x5, 9x6, and 9x7? I don't think so.
Date: Thu Jan 20 10:39:35 2011
User: mightybass
Message:i agree with both Gulbis & Tnmountainman on using a lower freecell variant to determine success in a higher freecell variant of equal columns. like Gulbis i have some variants where my lower freecell variant is more successful that the higher one. and just like TNmountainman said, it goes to show the importance of luck and time vested / persistence. these were some of the factors i found important and wanted to use for determining ranking and by skipping the time factor, you've unfairly jumped your stats.
TN: i would also like to say two thumbs up on the following usage ... "streakology (streakonomics?)". i highly enjoyed both of these when i read through your post. a bright spot to my otherwise hum-drum day at the office.
with that being said, nice job to both Gulbis & Kaos on their impressive scores. i haven't moved my scores much lately as i am trying to branch out and have started playing in a few tournaments. i'm still not very fast, but i'm getting better and have a new found appreciation and enjoyment in the tournaments from back when i first played in them. hopefully by the end of the month i can have finished the requirements in a few themes and have a rank/rating to be proud of.
Date: Thu Jan 20 16:39:19 2011
User: ElGuapo
Message:I like how you think, Kaos. In practice it would be tough to figure out how to score a 6x8 streak in 6x9, even if there weren't the "persistence" problem to factor in. Wouldn't you need to compare both ranking lists and give everyone else on the 6x8 list the same credit?
Date: Thu Jan 20 16:42:54 2011
User: EZ-Ed9
Message:Would like to comment on the previous four posts. First, I deeply respect Kaos both as a player and a person. Kaos is an extremely honest person and he told us exactly what he did. I took it as Kaos explaining what he would be capable of rather than trying to fudge his score. There's no doubt in my mind that Kaos, if he wanted to, could crack 100 for 70 variants.
Once upon a time I held the record in both 6x9 and 6x8. Now I am fourth and second, respectively. Until the last few days I had only played 6x10 four times. It is one of the variants holding my score back in the approximate 40 variants I have played. I have moved my ranking in 6x10 from 724th to 271st this week. Yet, if I could use my 1413 streak in 6x9 I would be in the top 10 in 6x10. I don't think that would be right. So I play a variant (6x10) that these days rather bores me.
My scores in the 7x's might be more pertinent. I hold the all-time records in both 7x6 and 7x5. In 7x6 it is 1670. I went after the record in 7x7 and flopped. My 7x6 streak is 1670, yet I could only do 1198 in 7x7 (now 6th place). This shows you can not take anything for granted. My mind wandered and I lost a game I should have won in 7x7 -- so I have a substantially higher streak in 7x6 than 7x7. That's the reason we play the games.
One final note, if I could use my 1670 in 7x6 I would be top 10 in both 7x8 and 7x9. Yet, I have not played in either variant. Someday, if I want a good 70-variant score I am going to have to overcome boredom and play both 7x8 and 7x9 lots.
Date: Thu Jan 20 17:17:42 2011
User: mightybass
Message:i hope my post in no way implied that Kaos was being unethical or that he was cheating by using a lower variant to qualify a higher variant. if it did, this was not my intention. it was a good idea, just not very accurate as several users have commented on. again, i apologize if i made it sound like that.
if nothing else, this thread has opened up a very fun conversation and has gotten people playing some variants they normally might not. granted some of these variants may be less enjoyable, but it still has broadened everyone is some way.
Date: Thu Jan 20 23:35:55 2011
User: rws33315
Message:Thanks for the broadening!
Date: Fri Jan 21 15:52:18 2011
User: TNmountainman
Message:Speaking for myself, no, mightybass, I did not perceive any negative connotation concerning Kaos, and to be clear, I certainly intended none, either.......
Now.......let me re-interject another idea. It appears many on this thread are at least "ok" with using all 70 variants for calculating this "score". I opined that I liked groucho's idea better, with a couple of modifications. I don't know if there is any hope of Denny (or anyone else) making these calculations automatically, or not, but if there is, then anyone else want to think like this?:
How about, similar to the wide variety of tournament themes, having maybe 3-4-5 "flavors" of "scores" for streakology? One like mightybass suggested, one like groucho suggested, maybe another like mightybass's idea but not using all 70 variants, etc. The idea would be to maybe tweak the "three pillars of streakology" a little, with the different flavors each emphasizing the three to differing degrees. (Although I think "luck" should always have the least weight.) I know for myself, I have no interest in playing a bunch of 14-, 15-, or 16-sum variants to get a ranking in those. Ed's "overcome boredom" comment above seems to echo that. I think many more players *might* be inclined to try this out if the barrier(s) weren't so high. This is just an idea, maybe too diverse. I'm not trying to steal mightybass's thunder, but if indeed there exists an opportunity to measure these metrics, especially if it would become automated, then now would be the time to have this discussion - if anyone else sees merit in it, that is.......
Date: Fri Jan 21 16:42:57 2011
User: mightybass
Message:TN: i don't mind sharing the thunder. and i think most of the ideas for alternate rankings have all been very good. since most of these are difficult to calculate for just one individual (let alone everyone), i have not eagerly supported many of the counter ideas. if they were automated though, i would love to see these ideas come to fruition. i mentioned many posts ago that i am a statiscal nut and love to crunch numbers. but moreso i love someone else to crunch numbers and to them be able to check out their hard work. personally i say the more rankings the better. as much as i love freecell and all its flavors, it does start to get tot he point where you need to "overcome boredom". i think if we had some of these rankings in place, it would give people new things to strive for and keep that boredom from sinking in. if we do start to look at alternate ways of ranking and weighting systems, i highly agree that luck should play the least heavy role. i think skill should weight the most, with persistence somewhere in between. 50/35/15% would be a fair breakdown in my opinion. you'll have an endless discussion though on what defines skill, what defines persistence, and what defines luck (luck being the easiest to define). i would define luck as a combination of speed and your overall win % compared to the overall win % of the variant.if 70% of games are winable, then the most skillful player is the person who is closest to 70%. in the eevent of ties, speed is the tiebreaker. for persistence some combination of at least streak length and who knows what else. all of these discussions are moot though unless someone decides to automate the rankings. otherwise its just a fun discussion.
Date: Tue Jan 25 01:24:12 2011
User: Or_to_take_arms_against
Message:Goals-wise, this is the most prolific streaker I've ever seen:
http://www.spike.com/video/streaker-scores-goal/2728987
Date: Tue Jan 25 01:46:17 2011
User: !_--FAST-ISHAM--_
Message:Now thats a Golden Goal.
Date: Tue Jan 25 12:10:52 2011
User: mightybass
Message:not really what this thread has been about, but awesome none the less.
Date: Tue Jan 25 23:59:19 2011
User: Kaos
Message:EZ-Ed wrote: "Would like to comment on the previous four posts. First, I deeply respect Kaos both as a player and a person. Kaos is an extremely honest person and he told us exactly what he did. I took it as Kaos explaining what he would be capable of rather than trying to fudge his score. There's no doubt in my mind that Kaos, if he wanted to, could crack 100 for 70 variants."
Well, nice to see Ed call me extremely honest but I'll have to admit to being ethically challenged at times - especially if mrbuck has passed the Bushmill's my way.
At to top 100 in 70 variants, cracking the first page of 8x4 is now a tall order. When I first played here (with my original, pre-tournament nick), I managed to get to #50 or so on the 8x4 list but that streak is now on the bottom of page 2. Getting onto the first page of 13x0 stats is also hard. My original nick is on the 4th page there. I've also just slipped onto the second page of 9x5 and at what might be considered my fast but not blazing speed, I'd likely need 100 hours of flawless play to get into the top 100.
10x6 is even worse. 175 hours at an 0:45 per game pace would get you a streak of 14,000 and perhaps a first page entry by the time you got there. Earning the title of Most Prolific Streaker would require serious addiction.
===========================
ElGuapo wrote: "I like how you think, Kaos. In practice it would be tough to figure out how to score a 6x8 streak in 6x9, even if there weren't the "persistence" problem to factor in. Wouldn't you need to compare both ranking lists and give everyone else on the 6x8 list the same credit?"
On one hand this is true, but on the flip side, if you look at the 8x4 top 100, there are easily seven variants of RGK in there and who knows how many other duplicates. To counter-act all the duplicates, I'd advocate just plugging it in and seeing where it would fit.
That said, I've gone back and looked at my spreadsheet and can't see where I actually used a long streak total from a lower freecell group when making my calculations.
Date: Sat Jan 29 21:18:48 2011
User: Dr.Bombay
Message:Thought of another way to score Most Prolific Streaker. Simply divide your longest streak in any variant by the length of the overall longest streak.
So, anyone at the top of the list in any variant would get 1.00 and others would drop from there down to 0.00 if you've never played. The highest possible score would be 72.0 but realistically, anything above 10.0 would like be pretty impressive. My guess from reading above is that loskazanes (with a number of 1.0's) or possibly huffy would score highest but who knows.
This method of scoring has some nice advantages:
1) It is very stable - your score doesn't change as others pass you in variants you've let slide for a while. The only way your score drops is if someone posts a new all-time best.
2) Don't need to worry about how to handle ties.
3) It rewards really impressive leading streaks.
4) It rewards streaks that are near the longest.
5) You score points even if you don't show up on the top 1000.
6) Your score isn't affected by the same player clogging the pages with multiple nicks.
For example, Ez-Ed was recently aced out on the 6x8 list by loskazanes 1064 to 1063. With this scoring method, he hardly suffers as 1063/1064 is 0.999.
Got a Lucky 13 streak of 49? That's not going to crack the top 1000 any more but with this method, it would still be worth almost one quarter point (49/197 = 0.2487).
Date: Sat Jan 29 23:06:03 2011
User: mightybass
Message:the only thing i don't like about that method is my score is only 2.357. it definitely doesn't make me feel all warm and fuzzy inside. but as you mentioned, anything over 10 would be impressive. so i guess i'm a little over 1/5th impressive and almost 1/4th impressive.
only an unrelated note, i finally won a tournament. 1 victory is hardly bragging rights considering the resumes of some of the other posters on this thread, but it was still exciting when i saw my name at the top.
Date: Sat Jan 29 23:43:07 2011
User: TNmountainman
Message:Really good idea, there, Dr.Bombay. I think that's much better than anything else heretofore.
Sorry, mightybass. The Dr's idea includes to some extent the three pillars of streakology, and doesn't penalize one too severely for not playing all, or a large majority of variants. I think it's pretty fair...
Date: Sun Jan 30 00:32:24 2011
User: !_--FAST-ISHAM--_
Message:Dr.Bombay. Thats good. If youre #1 on all variants,youd be at 72 but divide them by 72 and get an overall score if you want which makes 1.0 and looks just like tourneys.
Date: Sun Jan 30 00:36:39 2011
User: deadwing
Message:70 variants (not counting 7x1 and 8x0 - though they would help my score) I have a score of 10.7414
2.7405 in 4 columns
1.6509 in 5 columns
1.4989 in 6 columns
1.1221 in 7 columns - have better on the davidxz300 alt due to .8000 in 7x2
.6912 in 8 columns
.9563 in 9 columns
.6246 in 10 columns
.4853 in 11 columns
.3285 in 12 columns
.3731 in 13 columns
I would like to sort by freecells, but it is not as easy to as columns due to the way I entered the numbers in the spreadsheet.
neat idea
footnote add 1 for 7x1 and 1 for 8x0 - both are lead by streaks of 1
Date: Sun Jan 30 01:07:48 2011
User: CubicSprock
Message:congrats mightybass on your first win!! I'm sure it won't be your last, looks like you did pretty good in the U.S. Open!!
as for the alternate method, sure there is some validity, but I still like the other method better...this method ends up weighting all the lower sum variants far heavier than the upper variants (for almost all people), b/c of the ridiculous sums...not to mention streaks will have even less meaning, like in 8x4 a streak of 500 will earn you .025 points, whereas if that streak was in 8x5, it'd be worth almost 10 times as much (.20). Also, the luck factor in variants like 4x4, 4x5, etc become huge, (as they do to a degree if you don't give ties all the highest rank), b/c very few people get lucky enough to get 2 in a row, those couple that do get a full .5 points (huge in this system), whereas in the original that luck, will get you a few spots (1 in 4x4, 8 in 4x5)...which is more accurate.
In all I think the variants are so diverse, no system can be perfect...the relative to #1 isn't bad for the medium tough variants, (10/11 sums, and higher sum 4-5 col. games). Anyway, even though it isn't my preferred variant...I calculated my score, which came to 16.5
Date: Sun Jan 30 01:25:43 2011
User: conflagration
Message:Kaos, is that you there with a 18.2? hmmmm....1st, huh? lol
Date: Tue Feb 1 23:57:53 2011
User: Kaos
Message:CubicSprock wrote: "like in 8x4 a streak of 500 will earn you .025 points, whereas if that streak was in 8x5, it'd be worth almost 10 times as much (.20)."
The same thing is almost as true with the other scoring method. In 8x5, a streak of 500 is good for 70th place while in 8x4 it is down at 269. For a streak of 300, in 8x5 it is good for 136th while in 8x4 it is only good for 708th. So, I can't see where that's a strong argument against this scoring method since the same argument applies with the others. I could see where you may want to multiply your score in 8x4 by some large factor to account for the popularity of 8x4. Realistically, I'd say any scoring method that gets loskazanes, huffy, and rgk* at the top would pass the eye test. I'd expect kenwa, Laffman, CS, myself and likely others to be in the next wave.
=====================
conflagration wrote: "Kaos, is that you there with a 18.2? hmmmm....1st, huh? lol"
I've calcuated my score thru 7-column variants with the Dr.Bombay method as just over 10.0 but can't see how I would get to 18.2 (or even the CS score of 16.5) with the remainder. I do see the loskazanes has a lot of 1.0's....
Date: Wed Feb 2 17:18:22 2011
User: CubicSprock
Message:I agree with you to a degree Kaos, yes there is a bias regardless, but it is significantly less by comparing with everyone...70 versus 269 is a factor of 4, looking at it in the 'worst' way, in another way though, it is 930 versus 731 in terms of top 1000 beat. I guess the thing is, the initial method puts importance on all variants, comparing you to everyone in the variant, while this method is top heavy and the only 2 people that matter are you and #1. Now, the initial method would be ideal (in my opinion), if it was a golf style score, where your rank is how many people are above you, going beyond the top 1000. In all, both methods have some validity, I think they just measure different things. The problems occur b/c people aren't catering their streaking to maximize/minimize these scores. If Denny put prestige behind these rankings, huffy, for example, could get respectable scores in the harder variants without much time spent and move up these ranking considerably...probably close to or ahead of zonny. I like the how many people are ahead of you method, because it rewards luck in the extreme variants less...take 4x4 for example, sirPape has a streak of 3, nobody else has more than 1...so in the how many people are above you, the score is only 1 for everyone else, 0 for sirPape (or 1 and 2 using ranking). In your method, the scores are 1 versus 0.33, an extreme points reward for something attributable primarily to be extremely lucky.
Date: Wed Feb 2 22:38:52 2011
User: conflagration
Message:Kaos: 18.209
13-column total: 0.39
12-column total: 0.50
10-column total: 0.99
10-column total: 0.90
9-column total: 1.25
8-column total: 1.53
7-column total: 1.60
6-column total: 3.46
5-column total: 4.22
4-column total: 3.37
Grand total: 18.209
CubicSprock: 16.52
Zonny: 15.81
kenwa: 15.36
Catailong2: 14.39
ElGuapo: 13.33
canman1944: 12.50
Top-100-of-all-variants: 12.49
Laffman: 12.25
huffy: 11.70
canman1944: 12.50
Laffman: 12.25
TitanicTony: 11.34
Repperzed: 9.68
raederle: 7.82
greggory: 7.56
Date: Wed Feb 9 14:24:27 2011
User: .EZ-Ed
Message:Been figuring out what I have done in the past in relation to this thread. Have also been playing variants I never played before.
Before this thread I had played in 38 variants and not played in 32. I still have five variants to play, and that kills you according to mightybass' system. Roughly one-third of my points come from those five variants.
The variants I had not played before this all started are those low-sum variants with low top scores (in other words the variants relying on luck) plus the high-sum variants where you just keep grinding out massive streaks. I played mostly 11, 12 and 13-sum variants, plus quite a few 14 and 10-sum variants. I did play some of the higher-sum variants in the narrows but not the wides.
Right now my score, using Dr.Bombay's system, is 16.74. I am using all my nicks to reach the scores posted here. I basically agree with CS that mightybass' system is preferable because the Bombay system rewards luck too much. But I do find the Bombay system intriquing.
In the mightybass system I have been using two sets of scores. The first sums up the 38 variants I had played prior to this thread. The second my score includes all 70 variants. As a streak reaches a rank that pretty much averages those original 38, I include that variant in first total. Now I am using 43 variants for that score.
My scores have been buoyed by 30 streaks in the top 100 of their respective variants, 14 of those in the top 10.
My mightybass score for my top 43 variants is 71.4. My score for all 70 variants (five of which I have a zilch ranking because I have not played them) is 232.1 Obviously, I have a lot of work to do in high- and low-sum variants.
I have a streak of 1237 in 8x4, so I am well entrenched in the top 100 there. But I get very little in the 8x4 according to the Bombay system. My best columns are the 7-x's. Following are where I currently rank in the eight 7x variants.
7x9 -- 73rd (just started playing this one)
7x8 -- 62nd (also just started playing)
7x7 -- sixth
7x6 -- first (I'm all over the top 100 list)
7x5 -- first (also all over the top 100 list)
7x4 -- 104th (have played this one very little)
7x3 -- eighth (Kaos talked me into playing this one)
7x2 -- 351st (also played very little)
My total 7x score is 36.4 without the 7x2 and 75.8 with it. My 7x Bombay score is 3.7.
I enjoy all posts in this thread. Good luck to you all.
Date: Wed Feb 9 14:54:25 2011
User: mightybass
Message:i haven't had much time to comment on the last 10-15 posts, but i have been reading all of them and do agree with EZ-Ed that this has been an enjoyable thread. i've started playing in more of the tournaments because i still feel this is the best system for ranking players (i'm still not very fast and haven't won hardly any, but it is fun). i still like my system because it tracks something the tournament rankings don't. i also enjoyed Catailong2's & Dr.Bombay's renditions. if i get sometime this weekend i may make some additional posts/comments, but if not, i hope you see my name at the top of some lists out there!
Date: Thu Feb 10 06:49:16 2011
User: eightnotramp
Message:Cattail (sry didn't go back for correct spelling)
You did the TOP 50 List - and Denny's not on it?
Anyone figure Denny's score?
Amazing thread - lovin' it.... no, just read the thread - haven't scored myself.
Date: Tue Feb 15 18:26:29 2011
User: mightybass
Message:drinking + freecell ≠ success. that ends today's math lesson
Date: Tue Feb 15 21:25:06 2011
User: Kaos
Message:conflagration wrote: "Kaos, is that you there with a 18.2? hmmmm....1st, huh? lol"
Sitting here with egg on the face over this one. I'd protest I had no idea but with a sometimes wayward moral compass, what good would it do?
Anyway, looks like my score is 18.2 or even 18.3 with a few streaks using alternate nicks. However, loskazanes comes out with 20.12 without any streak at all in 8x4 so would have to guess he/she has had at least one other nick in the past.
One adjustment I could see with the Dr.Bombay method is to divide the 8x4 streak by 1000 rather than by rgk's 19,793. That would get EZ-Ed a 1.237 in 8x4 and, off course, it would give rgk a 19.79 on 8x4 alone. Seems about right.
Date: Tue Feb 22 14:01:22 2011
User: EZ-Ed3-LAnseMI
Message:One aspect of playing variants you have never played before is you come across lots of nicks you never imagined exited. I just came across one of my instant favorites for the first time -- markypeapod! I would like to meet the person who thought that one up.
Date: Mon Apr 4 08:40:23 2011
User: AreWeThereYet
Message:Man, this thread died.
Date: Mon Apr 4 21:34:02 2011
User: Catailong2
Message:System 1 Update:
# . Nick . . . . . . Average . . . . . Ranked Variants
1. Zonny . . . . 105.47 . . . . . 69
2. CubicSprock . . . . 173.27 . . . . . 69
3. firenze . . . . 182.06 . . . . . 68
4. canman1944 . . . . 193.03 . . . . . 68
5. Top-100-of-all-variants . . . . 198.21 . . . . . 66
6. Catailong2 . . . . 214.1 . . . . . 67
7. greggory . . . . 247.53 . . . . . 65
8. kenwa . . . . 261.29 . . . . . 65
9. ElGuapo . . . . 279.9 . . . . . 65
10. Repperzed . . . . 288.57 . . . . . 62
System 2 Update:
# . Nick . . . . . . Pts Total
1. Zonny . . . . 54,881
2. CubicSprock . . . . 50,191
3. firenze . . . . 49,577
4. Top-100-of-all-variants . . . . 49,279
5. canman1944 . . . . 48,809
6. Catailong2 . . . . 47,335
7. Gulbis . . . . 45,406
8. Repperzed . . . . 45,044
9. greggory . . . . 44,997
10. TitanicTony . . . . 44,272
Date: Tue Apr 5 09:44:34 2011
User: Gulbis
Message:I seem to have forgotten how the second system works but I just calculated my score in the first system and it's 224.84 so I think I should be on your list, hmm. Also surprising to find out there are 12 low-sum variants I have never played, or at least never won a game in. If I work on that I might get under 200.
Date: Tue Apr 5 10:13:01 2011
User: firenze
Message:How does the first system work?
Date: Tue Apr 5 17:13:18 2011
User: ElGuapo
Message:The first system is just your average rank in the all-time streak rankings across variants. The two tweaks we make are that you get credit for the highest rank you are currently tied with, and if you're not on the all-time list you use 1,000 as your rank.
Or at least that's my recollection. It's all spelled out in this thread.
Date: Tue Apr 5 20:38:47 2011
User: Catailong2
Message:Gulbis, I’ll do my best to explain how your two figures were derived when I ran the update yesterday. Hopefully, if there are any other interested parties out there who want to score themselves, this lengthy explanation will be of some use to them as well. System 1 is calculated by taking a total of all of your rankings in each variant and then adding the unranked penalty (1,001 x number of unranked variants). Finally, you divide this subtotal by 70. For you, this means that your current score stands at 321.33.
[ 9,480 + (1,001 x 13) ] / 70 = 22,493 / 70 = 321.33
Breakdown:
Variant. . . . . Rank. . . . . . . Streak
13 x 3. . . . . 303. . . . . . . 878
13 x 2. . . . . 62. . . . . . . 2061
13 x 1. . . . . 21. . . . . . . 3317
13 x 0. . . . . 264. . . . . . . 19/90
12 x 4. . . . . 187. . . . . . . 248
12 x 3. . . . . 93. . . . . . . 276
12 x 2. . . . . 91. . . . . . . 447
12 x 1. . . . . 22. . . . . . . 1047
12 x 0. . . . . 67. . . . . . . 0/39
11 x 5. . . . . 182. . . . . . . 335
11 x 4. . . . . 92. . . . . . . 446
11 x 3. . . . . 131. . . . . . . 235
11 x 2. . . . . 37. . . . . . . 630
11 x 1. . . . . 10. . . . . . . 95/192
11 x 0. . . . . 68. . . . . . . 0/15
10 x 6. . . . . 543. . . . . . . 2367
10 x 5. . . . . 343. . . . . . . 444
10 x 4. . . . . 227. . . . . . . 307
10 x 3. . . . . 152. . . . . . . 274
10 x 2. . . . . 171. . . . . . . 14/116
10 x 1. . . . . 6. . . . . . . 0/61
10 x 0. . . . . 3. . . . . . . 0/9
9 x 7. . . . . 327. . . . . . . 185
9 x 6. . . . . 454. . . . . . . 244
9 x 5. . . . . 560. . . . . . . 41/1382
9 x 4. . . . . 315. . . . . . . 125
9 x 3. . . . . 13. . . . . . . 515
9 x 2. . . . . 166. . . . . . . 3/36
9 x 1. . . . . 29. . . . . . . 0/14
9 x 0. . . . . n.r.. . . . . . . 0
8 x 8. . . . . 232. . . . . . . 145
8 x 7. . . . . 161. . . . . . . 113
8 x 6. . . . . 115. . . . . . . 207/507
8 x 5. . . . . 314. . . . . . . 157
8 x 4. . . . . n.r.. . . . . . . 70/137
8 x 3. . . . . 105. . . . . . . 0/43
8 x 2. . . . . 133. . . . . . . 0/11
8 x 1. . . . . 460. . . . . . . 0/1
7 x 9. . . . . 105. . . . . . . 143
7 x 8. . . . . 78. . . . . . . 123
7 x 7. . . . . 70. . . . . . . 133
7 x 6. . . . . 37. . . . . . . 242
7 x 5. . . . . 152. . . . . . . 8/45
7 x 4. . . . . 148. . . . . . . 0/29
7 x 3. . . . . 262. . . . . . . 3/8
7 x 2. . . . . n.r.. . . . . . . 0/1
6 x 10. . . . 25. . . . . . . 554
6 x 9. . . . . 49. . . . . . . 159
6 x 8. . . . . 123. . . . . . . 16/36
6 x 7. . . . . 15 . . . . . . . 85
6 x 6. . . . . 213. . . . . . . 0/22
6 x 5. . . . . 53. . . . . . . 3/18
6 x 4. . . . . n.r.. . . . . . . 0/1
6 x 3. . . . . n.r.. . . . . . . n/p
6 x 2. . . . . n.r.. . . . . . . n/p
5 x 10. . . . 13. . . . . . . 0/170
5 x 9. . . . . 46. . . . . . . 0/48
5 x 8. . . . . 64. . . . . . . 0/23
5 x 7. . . . . 220. . . . . . . 0/5
5 x 6. . . . . n.r.. . . . . . . n/p
5 x 5. . . . . n.r.. . . . . . . n/p
5 x 4. . . . . n.r.. . . . . . . n/p
5 x 3. . . . . n.r.. . . . . . . n/p
4 x 10. . . . 145. . . . . . . 0/10
4 x 9. . . . . 105. . . . . . . 0/5
4 x 8. . . . . 646. . . . . . . 0/1
4 x 7. . . . . 452. . . . . . . 0/1
4 x 6. . . . . n.r.. . . . . . . 0
4 x 5. . . . . n.r.. . . . . . . n/p
4 x 4. . . . . n.r.. . . . . . . 0
*** Ranked Total: 9,480
*** Unranked Penalty: 13,013 ... (13 x 1,001 = 13,013)
*** Grand Total: 22,493
*** Average Score: 321.33 ... (22,493/70 = 321.33)
You can see how your score is severely hurt by the fact that you remain unranked in 13 variants. IMHO, a player's streaking habits should not truly be considered "prolific" unless he or she dabbles in nearly all (if not all) of the variants. In System 1, the punishment of 1,001 points for remaining unranked in any given variant is by design a powerful way to weed out the players who tend to focus on just the low-sum or high-sum variants. Remember, the system was designed to assess who the most "prolific" players are. If you were ranked in the top 10 while only being ranked in 57 of the 70 variants, I'd consider System 1 to be faulty.
As a bit of a dramatic yet hopefully effective example, let's say you are ranked 1st in 57 variants but you are unranked in the remaining 13. Your total would simply be (57 x 1) + (13 x 1,001) = 13,070. This would result in an average score of 13,070/70 = 186.71. Now let's say that a lesser-skilled skilled but more prolific player comes along and achieves an average rank of 150 across all 70 variants. Although you are ranked 1st on 57 of the variants where he is somewhere around 150th, he would still outscore you under system 1.
System 2 is a simpler system in that its score is somewhat easier to derive. It simply asks, “How many people do you outrank in each variant?” You then total those numbers up across all 70 variants. For example in the 4x_ series, you would score:
4 x 10 -- 145th (1,000 - 145 = 855 pts)
4 x 9 --- 105th (1,000 - 105 = 895 pts)
4 x 8 --- 646th (722 - 646 = 76 pts)
4 x 7 --- 452nd (464 - 452 = 12 pts)
4 x 6 --- n.r.
4 x 5 --- n.r.
4 x 4 --- n.r.
Total: 1,838 (47.8% of the 3,845 points possible in the 4x_ series)
In all 10 “levels,” your scores currently stand at:
4x_ variants: 1,838
5x_ variants: 3,413
6x_ variants: 5,259
7x_ variants: 5,900
8x_ variants: 4,952
9x_ variants: 5,136
10x_ variants: 5,555
11x_ variants: 5,480
12x_ variants: 4,540
13x_ variants: 3,350
*** Total Points: 45,423
Under System 2, there are 62,249 points available at this time. In other words, if you were 1st in all 70 variants, that would be your point total. To gauge your progress toward that end, you can take your score of 45,423 and divide it by 62,249 (72.97%). By comparison, Zonny—who is currently in first under both scoring systems—is at 88.25%.
I never really tried to track ties because they appeared to be too much work and I felt that the net result would be quite similar (some would benefit while some would hurt).
Cheers!
Date: Wed Apr 6 06:44:53 2011
User: malr
Message:13x3 273 13 variants 457.25
13x2 185
13x1 445
13x0 926
12x4 306 12 variants 306
12x3 177
12x2 253
12x1 460
12x0 334
11x5 703 11 variants 378.33
11x4 423
11x3 329
11x2 385
11x1 214
11x0 216
I'm going to do this slightly differently... but by the end, I should have the number needed, by taking the average of all the averages. So far, my overall average for the above 3 variants is 380.53
h
Date: Wed Apr 6 07:43:27 2011
User: malr
Message:10x6 1001 10 variants 608.42
10x5 1001
10x4 520
10x3 1001
10x2 325
10x1 276
10x0 135
hmm - got some catching up to do with these...
Date: Wed Apr 6 10:56:25 2011
User: Gulbis
Message:Catailong, thanks for the long explanation. However, I don't remember reading that 1001 should always be the punishment for no ranking. I thought that if there are, let's say, 200 people ranked in 6x4 then my ranking would be 201, not 1001. I thought that was the proposed idea and honestly it makes sense to me. 1001 is only for those variants were more than 1000 people are ranked. In fact, I just reread the first part of this thread and found this:
Quote (mightybass):
as i mentioned in my original post, the ranking for non participation should be different for each variant...it should be one + the last rank for a streak of 1.
So I stand by my score of 224 for now (unless everyone else agrees on the 1001 penalty).
Date: Wed Apr 6 12:20:23 2011
User: firenze
Message:This looks like it could get too complicated using either system.
Gulbis makes sense with using the last official rank plus 1 if you haven't played a variant.
It would seem fairer if ties were ranked using the lowest number, otherwise those who got there first have an advantage over late comers and anybody who may want to play that variant more than once since they would fall to the bottom of those tied unless they beat their previous best.
Look at the variant 4x9. If you streak only 1 game, you will not show in the all-time rankings. However, technically, you are tied with the person in the number 438 position and I think that should be your ranking. Thus, to truly rank someone in this situation, you would have to go to their personal scores to determine if they did win at least one game.
I am not sure any of this type of ranking is feasible unless Denny got involved with some sort of programming.
And, I am not sure we have really determined exactly what we are trying to rank.
Date: Wed Apr 6 19:21:56 2011
User: Catailong2
Message:Sorry about that Gulbis. Upon review of mightybass's post, you are absolutely correct. It appears that the ranking methodology I employed under his proposed system is invalid. It shouldn't be a problem to tweak the Excel formulas to account for my oversight; I just need to add an "if" statement testing whether a player is unranked and then add one to the total number of players ranked in that variant.
I still can't wrap my mind around an easy way to automate the system to account for ties. It seems that an ideal scoring system would assess a player's best streak in each of the 70 variants, check to see where any ties exist and then assign a rank in each of these instances equal to the highest tied player in each respective variant. For instance, in 8x1 you should be ranked 150th rather than 460th since your best streak of 1 is tied with all of the players ranked between 150 and 472. I agree with firenze that there really shouldn't be an "early bird" reward for having set an older streak.
The variants with the most significant impact in terms of ties could probably be narrowed down to those where the best streak is less than 100 games (there appear to be 26 of these). It appears that all 16 of the 10-sum or less variants, most of the variants with a low number of columns and some of the other more challenging variants are those I have in mind.
The best example that I found is the 7x2 variant. If you go and win 2 games and the system doesn't account for ties, you wouldn't even be ranked among the top 1,000. If we account for ties, your rank jumps to 141. The difference of 840 points (1,001-141) would have a 12 point impact on your average (840/70 = 12.3). With this example in mind, I definitely see why ties should be accounted for. The question that remains is how to track ties in an automated fashion. I'll let that one percolate for a while and perhaps a viable solution will make itself available.
I do have to admit though that figuring out the fairest way to assess one's rank (in terms of being the most "prolific") is half the fun. Back to work :)
Post follow-up