Subject: most prolific streaker

Date: Thu Jan 6 16:34:43 2011
User: mightybass
Message:
I came up with a ranking system to honor the most prolific streaker across all variants. This system doesn’t weight any streaks more heavily than others, nor does it weight any variants more heavily than others. It simply shows who streaks the best across every variant. For every variant that you carry an all-time streak in (current or busted), your rank is your point value for that variant. In variants that you have not played or have not made it into the all-time list, your point value is 1001 (sadly this is not 100% accurate, but unless someone wants to compile a list of every streak in every variant all the way down to 1, this is the next best thing). Then just add all your point values and divide by 70 (total number of variants). A perfect score would be 1 (1st place in every variant) and the worst place would be 1001 (never played any variants/ no all-time streaks). I imagine most players will have a score ranging somewhere between 701-1001. Players that are really dedicated to streaking in every variant will be 401-700. I’d guess that less than 5% of players would fall under 400, as this would be a truly incredible feat. My score is 771.56.

Date: Thu Jan 6 20:15:16 2011
User: Kaos
Message:
What do you do with ties - Go for the highest ordinal? e.g. if 27-39 positions are all by players with the same streak length, do the all get 27?

Date: Thu Jan 6 20:56:39 2011
User: TNmountainman
Message:
What's the point of being the most prolific? And how does one define "best"? For me, it's winning percentage; for others, it's a long streak or fast times. And the thought of saying that just because someone hasn't played a particular variant, that they are at best the 1001st "best" in that variant introduces a vast amount of error. What if someone only plays 5 variants? To be clear, I'm not trying to torpedo the idea; others have been floated before, and I think it's a cool concept. The definition of "best", though, is the elusive thing....

Date: Thu Jan 6 21:01:06 2011
User: Kaos
Message:
mightybass could just edit his post to change the one use of the word "best" to "most prolific" from the title.

Date: Fri Jan 7 00:54:32 2011
User: mightybass
Message:
kaos: for ties, i agree that it does pose a problem. i would say it should be that if positions 1-5 are all tied, then they all get a score of 1 and then the next person after them gets a score of 6. this is how almost all sporting events are ranked and it makes the most sense. tnmountainman: anything that can be done, can be ranked in an infinite amount of ways. you could definitely make a ranking of the best winning percentages. you could make a ranking of the best times. we already have a system to rank the longest streaks for each individual variants. this is just another way of comparing each other that i found compelling and that i was interested in. i never said that this defines who is the "best" player. it is just a way of seeing who has accumulated the most high ranking streaks thoughout ALL the variants. it is like the decathalon. some people may be really great sprinters. and some people may be really great hurdlers. but some people are really consistent across all the events. this is just a way of ranking people who have chosen to play many of the variants and have also ranked in many of the variants. and i disagree with your comment "just because someone hasn't played a particular variant, that they are at best the 1001st "best" in that variant introduces a vast amount of error". one of webster's definitions for the word ranking is "a position in a list based on a record of performance". if you have no record of performance, then you are not on the list. kaos: to your second posting, i would like to point out again that i am in no way saying this ranking defines the best player. it determines the best streaker in ALL the variants. huffy is obviously the best streaker in 10x6. rgk5 is obviously the best streaker in 8x4. my ranking again is more comparable to a decathalon, instead of an individual event. you don't have to be the best in any one individual variant to have the most high ranking streaks across the entire spectrum.

Date: Fri Jan 7 01:44:56 2011
User: roo
Message:
Groucho did a lot of work on variants a few years ago. I can't find the major work of his, but here is another one linked.

Link: variant analysis

Date: Fri Jan 7 03:20:18 2011
User: TNmountainman
Message:
mightybass: "tnmountainman: .....and i disagree with your comment "just because someone hasn't played a particular variant, that they are at best the 1001st "best" in that variant introduces a vast amount of error". one of webster's definitions for the word ranking is "a position in a list based on a record of performance". if you have no record of performance, then you are not on the list." I would counter that by saying that assigning a 1001 "rank" is equivalent to being on "the list". Would it not be more accurate to assign a "rank" as you suggest, but *only* for the variants that particular player plays (with some reasonable minimum of plays per variant, like 50?; 100?), and then divide by the number of variants the player participates in, rather than the full 70? But maybe further define it so that the minimum # of variants played, in order to be evaluated, be something like, oh, I don't know....20? Again, I'm not dismissing your idea, only seeing spots to possibly refine it....

Date: Fri Jan 7 12:10:44 2011
User: Boojum
Message:
I came in at 338.2. For variants where I was tied, I just used whatever number I was assigned by the system. Picking the highest number might be better, but it takes long enough to compile all the numbers as it is. I see the point of assigning 1001 to variants where one is not listed, but there are variants which do not have 1000 people listed. I have no personal stake in the matter, since I'm listed on all variants(I'm also listed on 8x0 and 7x1, but it would be pointless to include them), but a player who skipped a few little-played variants could be affected. Of course, since the listings include different nicks of the same players, the results are necessarily only approximate, if they mean anything at all.

Date: Sat Jan 8 16:29:58 2011
User: mightybass
Message:
TNmountainman: the ranking of 1001 is a benchmark. no variants assign outside of 1000 for all-time streaks, therefore the ranking of 1001 is the baseline for non participation. 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. unfortunately since we do not have data for every single streak in every single variant down to 1, this will have to do. i do not like the idea of only averaging for the variants someone has qualified in because if someone chooses to streak in a new variant, they would actually decrease their ranking until they accomplished a high enough position to match their previous average. Boojum: i agree that figuring out where the tie breaks is tedious, however i do feel it is more accurate. otherwise how does one streak of say 30 compare to another streak of 30. i also agree that for variants that have less than 1000 on the all-time list this introduces more inaccuracies. but back to your previous point...it would take much more time to weed out the variants that don't have 1000 members. i chose 1001 because most variants range down to 1000 on the all-time list and the few that don't probably don't impact things dramatically anyway. i appreciate all the suggestions and i would be interested in seeing how all of them impact ranking. i'm a bit of a statistics nut and enjoy computing data. if it wasnt so time consuming to compile everything, i would definitely run all the countering suggestions and see what impact they make. however as Boojum pointed out, just computing your own score takes a long time. this was just something i thought was interesting and there was no current system in place to track it. i also found it interesting because i am getting close to having a rank in all variants and it has taken me some time to do that. i wanted to see how that accomplishment compared to others who were either close or had already done so. on a final note, very impressive 338.2 Boojum.

Date: Sat Jan 8 21:07:57 2011
User: Kaos
Message:
I see 72 variants listed by Denny (not 70) are you excluding 4x4 and something else? For the 72 listed variants, my average was 414.1 so Boojum I am impressed too. I've got 25 variants I have never tried. If I exclude those, I drop down to 101.9. For a couple of variants, I had to pull in some streaks under either old nicks or alternate nicks.

Date: Sun Jan 9 10:14:04 2011
User: Boojum
Message:
Kaos - I assume we are excluding 8x0 and 7x1, which are both included in the all-time list but have fewer than 10 people listed, since they do not turn up on the play list. I think the reason my number looks impressive is because I have no 1001s to factor in. Your number for the variants you have tried is astounding. mightybass - If I have the time, I'll recalculate my number using the break, for your statistical delectation. It should make a difference in a few variants, for example, 6x2, where I would go from 618 to (tied for) 54.

Date: Sun Jan 9 18:26:42 2011
User: Boojum
Message:
Using break points, I score 281.2. It makes a significant difference on variants whre there are few streaks over 2 or 3. I love this statistic - it makes me seem like a much better player than I really am.

Date: Mon Jan 10 03:55:25 2011
User: TNmountainman
Message:
mightybass - IMO, the "error" caused by assigning a probably wildly "incorrect" "rank" of 1001 in a variant to someone who's never played it seems gigantic compared to your rationale that beginning a new variant would temporarily throw off that person's average. Kaos's post clearly illustrates this. His average on the variants he has played is 101.9, but adding the 25 he hasn't attempted changes it enormously to 414.1. So, one needs to be careful what one is trying to measure. Surely you can see that the disparity between those numbers demonstrate that including all variants, including those not played, skews the "ranking" greatly... Of course if you're the one defining it, then you are at liberty to design it however you wish. (After reading that sentence, it sounds a bit "catty", but I assure you I don't mean it that way....) Maybe I truly don't understand what you're going for. I understand your concept of a "baseline" of 1001, but it seems to me that that would act more like a weight dragging down a player's "truer" ability to streak. But I think some of my confusion stems from, as Kaos theorized, your use of both the words "prolific" and "best", things which of course have very different meanings in the freecell context. Again, just trying to be helpful and play Devil's advocate; not trying to be argumentative....

Date: Mon Jan 10 09:23:25 2011
User: Turpin827
Message:
mighty bass uses the analogy of the decathlon which is apt, but Olympic athletes don't have to divide by the total number of events in the Olympic games. I believe that a lot of players choose to play only a few variants. Perhaps a rating system could divide by a player's favorite five or ten variants?

Date: Mon Jan 10 11:44:55 2011
User: Boojum
Message:
Plainly, mightybass's approach penalizes those who do not play in all variants. Leaving aside the size of the penalty, though, it does not seem to me to be so easy to develop an alternative approach which does not have its own drawbacks. The fact is, I think, streakers have different tastes, and it is hard to compare them. For example, some players like the low-total variants (4x4, 8x1), where only a small percentage of the games are winnable. Winning just one game in those variants automatically puts one high on the alltime list. By my calculation, winning one game each in ten of those variants would give an overall average rank of 67.8. How would we compare that to someone who prefers to play the high-total variants like 10x6, where it is much easier to win any particular game, but it takes hundreds of victories even to get on the alltime list? Whatever its weaknesses, mightybass's approach has the virtue of comparing apples with apples.

Date: Tue Jan 11 00:04:30 2011
User: ElGuapo
Message:
I get 294.01 looking at all 70 variants and using break points. I could possibly get that down a little if I compared rankings on some old nicks, but I think I'd rather stick with the one name for this. I've put in a little time in most variants, but the few you don't play really bite you. More than half of my total points came from 11 variants. Personally I don't see that as a fault in mightybass's idea though. If people are interested enough, they'll focus more energy on those variants that are costing them the most. Of course the biggest hurdle to this taking off is the data gathering. It's slightly easier if you just modify the URL as you go down the list, but still tedious. It'd be cool if Denny could automate this and add a separate ranking list for it.

Date: Tue Jan 11 01:28:43 2011
User: TNmountainman
Message:
ElGuapo: "If people are interested enough, they'll focus more energy on those variants that are costing them the most." TN: True; but that'd be an absolutely huge time investment to play however many thousands of 9x6s (or whatever) to get streaks in those variants, right? (Admittedly, this shows my bias as more of a solver. If indeed, measuring "prolific" is the goal, then I can see the rationale there.) Turpin827: "mighty bass uses the analogy of the decathlon which is apt, but Olympic athletes don't have to divide by the total number of events in the Olympic games." TN: This is a good comment, and the way I view his suggestion, too... Maybe some sort of 'compromise', in which instead of all 70 variants, maybe the top 35 are used. Or 50?

Date: Tue Jan 11 01:43:13 2011
User: ElGuapo
Message:
Well, in whatever the time difference between my time stamps here and above, I was able to make a pretty big dent in the few variants I wasn't ranked in. Score is now 257.03, shaving 2,589 total points off! TN is right about some of those, though. You'd have to be pretty determined to go after the 9x6, 9x5, 10x6, and 10x5.

Date: Tue Jan 11 07:29:27 2011
User: firenze
Message:
Wouldn't the most prolific streaker be the person with the most games played over all the variants?

Date: Tue Jan 11 07:37:37 2011
User: firenze
Message:
I guess I should amend the above statement to read "most games won" since there have been instances of players using freecell games for other contests not involving winning and playing hundreds of thousands of games.

Date: Tue Jan 11 22:34:33 2011
User: mightybass
Message:
i'm not going to respond to every single comment that has been made. i will pick a few that i found intriguing and then just summarize the rest. not having to average the decathalon scores makes no difference nor would it here if we didnt average them. the order of rank would still stay the same. the reason i chose to average them is because a score of 500 is much easier to wrap one's head around than say a score of 35,000. since my system includes all the variants, everyone's average is still based off the same factor. i would be highly excited to see a ranking of everyone's score based strictly on the number of variants they have played. i think that ranking would be just as meritful as one that incorporates all the variants. the reason i like using all of them is because you dont have to be the best in every variant, or even in just one variant, but consistancy across all the variants is what puts you on the top of this ranking. will it appeal to everyone...of course not. is it meant to...of course not. i would like to apologize for using the words "prolific" and "best" (although i don't recall using best). although i felt these words in my head conveyed what i was trying to get across, on an online post it is sometimes very difficult to get your meaning understood using only words and not being able to explain things directly. my intent was simply to see how i ranked among other players who play a lot of variants and streak well in them. to summarize: everyone seems to want to rank only the variants they play. go ahead. i would love to see that ranking. that is not what i sought to find out though with this ranking. go enter a decathalon. only participate in the events you enjoy or are good at. you won't do very well overall. that doesn't mean you aren't great at the events you entered. and there are rankings for those individual events or for only 2-3 events if that is what you prefer. but this ranking system i envisioned is my "decathalon" of freecell. and by definition of a decathalon you will be scored in all events (entered or not) and if not entered, you will get the lowest score possible. hence my assigning 1001 to non played variants. thank you to everyone who has posted so far. this has been very enjoyable and i hope the discussion continues. the only thing i could hope for in addition would be elguapo's suggestion for automation from denny.

Date: Wed Jan 12 00:31:43 2011
User: Catailong2
Message:
I have to say that I have been inspired by the concept of a system that ranks streakers using a scope extending beyond just single variants. Ideally, the system would need to be a) based on sound logic, b) calculated accurately and c) somehow kept up-to-date. From what I've read so far, there has been some good progress made in establishing the basic framework. Sure, there will undoubtedly be differences of opinion regarding the logic used to develop such a system but that's part of the fun of pioneering a new concept (or one that has yet to be made publicly available). After some consideration, I believe there are two broad choices that make sense (with much of the finer tuning left up for further discussion): 1. A point-reducing system: This approach would assign each ranked nick a score for each variant based on his/her individual rank, then sum and divide by the number of variants. The sum would either be divided by 70 or just by the number of variants in which each nick is ranked. My preference would be 70 for a more global view. I refer to this approach as a point-reducing system because the goal is to have the lowest score possible. 2. A point-amassing system: This approach would assign each ranked nick a score for each variant in which he/she appears based on his/her individual rank. The twist here would be to determine the score by how many ranked players in each particular variant that player outranks. Example 1: Instead of being assigned 250 if you are ranked 250th in the 9x4 as you would in the first system, you would receive 750 points (1,000 players ranked – 250). Example 2: Instead of being assigned 100 if you are ranked 100th in the 4x5 as you would in the first system, you would receive 64 points (164 players ranked – 100). The highest possible point total would be for someone ranked 1st in all 70 variants and would be 62,215 (the total number of all ranked players across all variants minus 70). I refer to this approach as a point-amassing system because the goal is to have the highest score possible.

Date: Wed Jan 12 00:32:46 2011
User: Catailong2
Message:
As far as calculating the totals would work, I feel that this actually poses a much bigger challenge. Sure, each interested player could calculate his or her own score (as some posters have done) or someone could devise a system and post the top results semi-regularly. If there was enough sustained interest and the concept gained some traction, a more permanent solution might be developed. I would push for the semi-regular posting idea (weekly, bi-weekly, or monthly depending on the man-hours required). The simplest solution I can come up with is to import the all-time ranking pages for each variant into an Excel worksheet and use this data to assign scores to each player, sort by the score and then provide a list of the leaders. With this in mind, I believe the most manageable approach would be to divide the 70 variants into 10 levels (e.g., all seven of the 4x_ variants would comprise level 1 while all four of the 13x_ variants would comprise level 10). The data could be used to determine the highest scoring player (or lowest scoring player depending on the approach) for each level as well as for all 70 variants combined. One could even go on to determine the best player for all “_x0” variants (no free cells available) and so on. After playing around with both of the scoring systems mentioned in my previous post and scoring both mightbass (credit definitely due for the inspiring idea) and Zonny (one of the best all-around streakers I’ve witnessed), I came up with the following results: mightybass: System 1: 769.00 average System 2: 15,812 total points Zonny: System 1: 105.33 average System 2: 54,911 total points As far as the ten Excel files for the different levels, I’ve started the project (I’ve completed the files for the 4x_, 11x_, 12x_, and 13x_ variants) and expect that I can come up with something that fits the bill pretty soon. It’s just that setting up the “master” file and then modifying that for all ten levels by tweaking the imported web page addresses for each of the 632 results pages is rather tedious work. I am determined to make it work though, so stay tuned. As a sample of my findings so far, here are the top ten 13x_ streakers: 1. Zonny (3,662 points) 2. canman1944 (3,631 points) 3. huffy (3,514 points) 4. Jene (3,467 points) 5. raederle (3,467 points) 6. TitanicTony (3,396 points) 7. Baggins (3,395 points) 8. Top-100-of-all-variants (3,353 points) 9. Gulbis (3,321 points) 10. jerrkar9 (3,295 points)

Date: Wed Jan 12 00:57:24 2011
User: mightybass
Message:
catailong2: i must say that your tweak on the ranking is quite interesting. i will be most interested in seeing the results you tabulate. i have gone through and calculated everyone's ranking for my system (with a few tweaks of my own) for anyone who is ranked #1 in a single variant and also anyone who posted on this topic. it was extremely tedious and i cannot fathom trying to constantly update it as people continue to play. if this is something you are willing to do, i applaud you. as i mentioned earlier, i also applaud you on adapting the ranking with a point amassing system. i believe the two systems will be very close to each other. especially since the biggest thing i tweaked was to only assign 1001 to the variants with 1000 ranked players. for variants that had less than 1000, i simply added 1 to the lowest ranked player. this should mimic your point amassing rank system (just in reverse so to speak). i also am pleased to see you did not try to lobby for why using all 70 variants was "not your cup of tea", but instead came up with a clever rendition of your own still using all 70 variants.

Date: Wed Jan 12 11:44:23 2011
User: ElGuapo
Message:
I like the simplicity of the average ranking you used, mightybass. It's nice when the score has some intrinsic meaning rather than just being a number for ranking people. With your approach the score is basically just your average rank across all variants (with some necessary tweaks of course). Catailong2, I may be wrong but I don't think those two scores you showed for mightybass jive: mightybass: System 1: 769.00 average System 2: 15,812 total points If the average rank is 769 and there are 70 variants, then I would expect 769 × 70 - 62,215 to approximately equal 15,812. It's very possible I'm missing something.

Date: Wed Jan 12 15:54:26 2011
User: mightybass
Message:
ElGuapo: i don't believe that Catailong2's 2nd adaptation of my ranking correlates in reverse exactly to my original system, and this is why... 1)ties. it makes a difference on if you count the highest place for a tie or if you just take the face position you have inside the tie. since one of the rankings is how many people you are above and the other is how high up you are, ties will make it so the two systems will not jive back and forth with each other. especially if you do not calculate ties the same way from one system to another. even if you do, the ties will still cause some gaps when trying to correlate between them. i have suggested that for my ranking that you take the highest position of the ties and everyone gets that score. for Catailong2's sytems i suggest that you only count yourself above another player on the list if you have a higher streak numerically. if they are ranked lower than you, but have the same streak total, you are not really ahead of them. 2)the other thing i noticed is in my system you score yourself even if you are outside the all-time list and in Catailong2's you only receive a score if you are ahead of someone. this is what causes the other discrepencies when trying to calcualte the two systems back and forth. they are both great systems in my opinion, you just can't correlate them that way. the proper way to interpret your correlation would be to look at the difference your calculation creates and this would be the number of people you are tied with + the number of variants you were outside of the all-time list on.its not very useful information, but it at least explains the confusion (assuming i'm correct. i could just as easily be full of crap)

Date: Wed Jan 12 19:24:06 2011
User: Catailong2
Message:
System 1 would give you 1,001 points for each variant where you are not ranked (points you don't want since you're seeking the lowest possible score): Variant .. Points . . . . Variant .. Points 13 x 3 . . . . 641 . . . . . . 10 x 3 . . . 906 13 x 2 . . . . 418 . . . . . . 10 x 2 . . . 909 13 x 1 . . . . 546 . . . . . . 10 x 1 . . . 1001 13 x 0 . . . . 1001 . . . . . 10 x 0 . . . 1001 12 x 4 . . . . 240 . . . . . . 9 x 7 . . . . 384 12 x 3 . . . . 99 . . . . . . . 9 x 6 . . . . 702 12 x 2 . . . . 326 . . . . . . 9 x 5 . . . . 1001 12 x 1 . . . . 646 . . . . . . 9 x 4 . . . . 986 12 x 0 . . . . 1001 . . . . . 9 x 3 . . . . 752 11 x 5 . . . . 272 . . . . . . 9 x 2 . . . . 1001 11 x 4 . . . . 99 . . . . . . . 9 x 1 . . . . 1001 11 x 3 . . . . 114 . . . . . . 9 x 0 . . . . 1001 11 x 2 . . . . 561 . . . . . . 8 x 8 . . . . 99 11 x 1 . . . . 973 . . . . . . 8 x 7 . . . . 324 11 x 0 . . . . 978 . . . . . . 8 x 6 . . . . 989 10 x 6 . . . . 593 . . . . . . 8 x 5 . . . . 1001 10 x 5 . . . . 630 . . . . . . 8 x 4 . . . . 1001 10 x 4 . . . . 489 . . . . . . 8 x 3 . . . . 1001 . . . . . Variant .. Points . . . . Variant .. Points 8 x 2 . . . . 1001 . . . . . . 6 x 2 . . . . 1001 8 x 1 . . . . 1001 . . . . . . 5 x 10 . . . 653 7 x 9 . . . . 99 . . . . . . . . 5 x 9 . . . . 1001 7 x 8 . . . . 65 . . . . . . . . 5 x 8 . . . . 1001 7 x 7 . . . . 307 . . . . . . . 5 x 7 . . . . 1001 7 x 6 . . . . 944 . . . . . . . 5 x 6 . . . . 1001 7 x 5 . . . . 990 . . . . . . . 5 x 5 . . . . 1001 7 x 4 . . . . 1001 . . . . . . 5 x 4 . . . . 1001 7 x 3 . . . . 1001 . . . . . . 5 x 3 . . . . 1001 7 x 2 . . . . 1001 . . . . . . 4 x 10 . . . 632 6 x 10 . . . 565 . . . . . . . 4 x 9 . . . . 1001 6 x 9 . . . . 395 . . . . . . . 4 x 8 . . . . 1001 6 x 8 . . . . 581 . . . . . . . 4 x 7 . . . . 1001 6 x 7 . . . . 945 . . . . . . . 4 x 6 . . . . 1001 6 x 6 . . . . 945 . . . . . . . 4 x 5 . . . . 1001 6 x 5 . . . . 1001 . . . . . . 4 x 4 . . . . 1001 8 x 4 . . . . 1001 . . . . . . 6 x 4 . . . . 1001 8 x 3 . . . . 1001 . . . . . . 6 x 3 . . . . 1001 Sum = 53,830 Average = 769.00

Date: Wed Jan 12 19:27:52 2011
User: Catailong2
Message:
System 2 would calculate your total points by subtracting your rank from the number of ranked players in each variant and then summing: Variant . . . . Points . . . . . . . . . . Variant . . . . Points 13 x 3 . . . 1000 - 641 = 359 . . . . . 8 x 3 . . . . . . . n/a 13 x 2 . . . 1000 - 418 = 582 . . . . . 8 x 2 . . . . . . . n/a 13 x 1 . . . 1000 - 546 = 454 . . . . . 8 x 1 . . . . . . . n/a 13 x 0 . . . . . . . . n/a . . . . . . . . . . . 7 x 9 . . . 1000 - 99 = 901 12 x 4 . . . 1000 - 240 = 760 . . . . . 7 x 8 . . . 732 - 65 = 667 12 x 3 . . . 1000 - 99 = 901 . . . . . . 7 x 7 . . . 1000 - 307 = 693 12 x 2 . . . 1000 - 326 = 674 . . . . . 7 x 6 . . . 1000 - 944 = 56 12 x 1 . . . 1000 - 646 = 354 . . . . . 7 x 5 . . . 1000 - 990 = 10 12 x 0 . . . . . . . . n/a . . . . . . . . . . . 7 x 4 . . . . . . . . n/a 11 x 5 . . . 1000 - 272 = 728 . . . . . 7 x 3 . . . . . . . n/a 11 x 4 . . . 1000 - 99 = 901 . . . . . . 7 x 2 . . . . . . . . n/a 11 x 3 . . . 1000 - 114 = 886 . . . . . 6 x 10 . . . 1000 - 565 = 435 11 x 2 . . . 1000 - 561 = 439 . . . . . 6 x 9 . . . 1000 - 395 = 605 11 x 1 . . . 1000 - 973 = 27 . . . . . . 6 x 8 . . . 932 - 581 = 351 11 x 0 . . . 1000 - 978 = 22 . . . . . . 6 x 7 . . . 945 - 945 = 0 10 x 6 . . . 1000 - 593 = 407 . . . . . 6 x 6 . . . 1000 - 945 = 55 10 x 5 . . . 1000 - 630 = 370 . . . . . 6 x 5 . . . . . . . . n/a 10 x 4 . . . 1000 - 489 = 511 . . . . . 6 x 4 . . . . . . . . n/a 10 x 3 . . . 1000 - 906 = 94 . . . . . 6 x 3 . . . . . . . . n/a 10 x 2 . . . 1000 - 909 = 91 . . . . . 6 x 2 . . . . . . . . n/a 10 x 1 . . . . . . . . n/a . . . . . . . . . . 5 x 10 . . . 1000 - 653 = 347 10 x 0 . . . . . . . . n/a . . . . . . . . . . 5 x 9 . . . . . . . . n/a 9 x 7 . . . 1000 - 384 = 616 . . . . . 5 x 8 . . . . . . . . n/a 9 x 6 . . . 1000 - 702 = 298 . . . . . 5 x 7 . . . . . . . . n/a 9 x 5 . . . . . . . . n/a . . . . . . . . . . . 5 x 6 . . . . . . . . n/a 9 x 4 . . . 1000 - 986 = 14 . . . . . . 5 x 5 . . . . . . . . n/a 9 x 3 . . . 1000 - 752 = 248 . . . . . 5 x 4 . . . . . . . . n/a 9 x 2 . . . . . . . . n/a . . . . . . . . . . . 5 x 3 . . . . . . . . n/a 9 x 1 . . . . . . . . n/a . . . . . . . . . . . 4 x 10 . . . 1000 - 632 = 368 9 x 0 . . . . . . . . n/a . . . . . . . . . . . 4 x 9 . . . . . . . . n/a 8 x 8 . . . 1000 - 99 = 901 . . . . . . 4 x 8 . . . . . . . . n/a 8 x 7 . . . 1000 - 324 = 676 . . . . . 4 x 7 . . . . . . . . n/a 8 x 6 . . . 1000 - 989 = 11 . . . . . . 4 x 6 . . . . . . . . n/a 8 x 5 . . . . . . . . n/a . . . . . . . . . . . 4 x 5 . . . . . . . . n/a 8 x 4 . . . . . . . . n/a . . . . . . . . . . . 4 x 4 . . . . . . . . n/a Point total: 15,812

Date: Wed Jan 12 22:42:50 2011
User: BrownsRedSox
Message:
I love these ideas and spent some time getting on the lists in the low played variants. Because I am nowhere near the term "prolific" in my abilities and could still get on the all time lists in approximately half the variants, it leads me to believe we should shorten the list from the original 70. I'm thinking we should at least remove those variants where a streak of '1' qualifies on the all time list. I don't see how a streak of 1 can qualify someone as "prolific". I'm thinking that we should use the top 50 variants.

Date: Thu Jan 13 09:37:28 2011
User: mightybass
Message:
BrownRedSox: while i appreciate your comments and suggestions, perhaps you should read my several previous posts on why i am using all 70 variants. although i would be interested in the results of a ranking that only used a certain number of variants, that was not what i wanted to know when i came up with this system. i feel we have beaten this poor horse enough at this point in time. if someone wants to derive a system of ranking for say 20, 30, or any other number of variants...be my guest. start a new thread and i will follow it and be happy to compare the results. this ranking is for all 70 variants though

Date: Thu Jan 13 19:32:38 2011
User: mightybass
Message:
when i first posted this idea i had guessed in my head a ratio of probably 5%/35%/60% (1-400/401-700/701-1001). by using the highest rank in ties, 1001 is no longer the benchmark for no ranks/no played variants...it is 887.83, with that being said i still feel the percentage breakdowns will stay the same just ranging from (1-350/351-625/626-887.83). i took a small sample size of all the 1st place players in each variant + every participant in this thread and surprisingly it came out 8%/32%/60%. more suprisingly was that someone cracked the sub-100 score. i didnt anticipate seeing this with a total of 70 variants. Zonny came in with an impressive 91.17 (again, using the highest rank in the event of ties). also impressive was TitanicTony with a score of 250.00 (good for 2nd place). what i found most impressive about TitanicTony's score was 1) TitanicTony is ranked 1st in 11x3 and 2) has 16 variants with no rank. i didnt expect anyone who had dedicated the time to get to 1st place in a high streak variant or someone with that many "no ranks" to have that good of a score. kudos!

Date: Fri Jan 14 00:51:25 2011
User: CubicSprock
Message:
average of 338.6 I imagine I could very easily get that under 200, by running off a few wins in those 11x,12x, and 13x variants. My averages by # columns: 4: 133 (need to play and win 1 4x5 and 1 4x6 and 2 4x8's to improve greatly) 5: 47 (brought down by 0 wins in 5x3) 6: 133 (can improve quickly with 6x10 of 798) 7: 66 8: 412 9: 758 10: 355 11: 438 12: 523 13: 771 just haven't put in the time on all the 8+ column, 13+ sum streaks...could pick and chose to get overall under 200 if I feel so inclined. Interesting topic mightybass...might motivate me to revisit some of those variants to get see how quickly that average can rise...and see if I can catch zonny :P

Date: Fri Jan 14 02:04:01 2011
User: CubicSprock
Message:
up to 308...time for sleep. :P a quick 3 in a row in 4x6 and I went from unranked to tied for first :)

Date: Fri Jan 14 03:26:14 2011
User: TNmountainman
Message:
wow. .bow CubicSprock That being said, it just somehow seems "silly"(?) for a system to reward that equally with someone needing to play (and win) 195,610 straight games in 10x6 (huffy's all-time and current streak), with a time expenditure of 2400 hours (60 40-hour weeks). (I suspect 3 straight 4x6s is harder, but that's difficult to quantify.......) And yes, I know it's the *average* that's the thing..... It took CubicSprock just over an hour (at least that's the time between his two postings) to accomplish what huffy spent over a full working year doing..... (Yes, I know that's an oversimplification, but still....)

Date: Fri Jan 14 12:12:55 2011
User: CharitonIowaUSA
Message:
I've found this thread most interesting. Kind of makes me wish I had played a whole lot less than the dozens of nicks I have played over the years. At one time or another I have held the all-time records in about 10-12 different variants. A few of them might still be waiting to be topped. I see where not being in the top 1000 might pose a problem with everyone getting a rating of 1001 for that variant. But what do you do since Denny only goes with the top 1000? This simple solution posed by mightybass is by far the best I have seen to rate streakers. I agree with TNmt that winning percentage is the best way to rate a streaker. Rating where you stand on an all-time list kind of takes winning percentage into account. I know I would have trouble playing easy variants like 13-3, 10x6 and the like. I once decided to play variants that didn't take much concentration. I tried 10x4 and was super bored by the time I had a streak in the 200's. Could never play that variant again. Where is the challenge? To each his own, I guess. I have played enough variants to know that CubicSprock would rank high on an all-time list. Another person who would rank high would be mercury. Trouble with merc is that he plays about as many nicks as me and would have his scores diluted.

Date: Fri Jan 14 12:55:16 2011
User: ElGuapo
Message:
There's no reason huffy can't play a few 4x6s and level the score, right? As to the whole conundrum about a ranking of 1001, if Denny were to step in and automate this process there's no reason he couldn't bump all the ranking lists to 10,000 at the same time. The data is there already.

Date: Fri Jan 14 13:19:11 2011
User: firenze
Message:
It seems that with these "rankings", streaking to a top 10 position in 10x6 carries the same importance as streaking to a top 10 position in 10x1. In other words, game solving skill does not matter. What is being measured is persistance and who has the most time on their hands.

Date: Fri Jan 14 14:00:22 2011
User: TNmountainman
Message:
ElGuapo: "There's no reason huffy can't play a few 4x6s and level the score, right?" TN: Correct. But it's *not* so easy for Cubic Sprock (or anyone who isn't a long way down that road already) to play several 10's of thousands of games in several variants. But I guess that's where "prolific" comes in, and that's fine.... firenze: "In other words, game solving skill does not matter." TN: Correct again. I think that makes my point. And if all one is interested in is "prolific", then that indeed is fine.... I would think, though, that that information is already out there. All one need do is look at a "games played" list....

Date: Fri Jan 14 15:27:16 2011
User: ElGuapo
Message:
I can see this kind of ranking is not going to appeal to everyone. To me it's markedly different from just looking at a "games played" list. There's a reason, for instance, that I don't have a streak of 3 in 4x6 and CS does. My best streak is 2, and I've played 222 games in the variant. CS has played 14. Yes, there are variants on the opposite end of the spectrum where it's all about persistence and having time on your hands. But even then not everyone can do it. Yesterday I blew a streak of over 200 games in 9x4. Of course someone could come up with a way of doing a weighted average based on the perceived significance of each variant (maybe weighted based on popularity?), but I like that this encourages people to branch out. And now that I think about it, weightings based on games played would make 8x4 the only rating that matters. Or you could try a more subjective weighting system, but you'd never find consensus on which variants deserve more consideration. For me, the simple average method works.

Date: Fri Jan 14 22:11:27 2011
User: mightybass
Message:
it seems to be the big topic as of late is the variants with a high streak count...10x6 in particular. what i think people are overlooking is that you don't have to be first in every variant or even in any variant to place well in this ranking. Zonny is clearly evident of that. huffy's 10x6 record will probably never be beaten, but for this ranking it doesnt have to be. in fact of the 60-some rankings i calculated, 13 were higher than huffy and 6 of those 13 were people who have posted on this thread. when i originally started playing on this site, my goal was to beat huffy's 10x6 streak and i quickly realized at my rate of play each day, it would never happen. so i decided to start playing other variants and realized i probably wouldnt ever be first in those either. that's when i started to ponder, 'what if i just try to get the highest average rank in all variants.' hence the ranking system i came up with and this thread. i think this ranking system does a good job of rating 3 characteristics that any good streaker has / needs. and the person who excels in all 3 of those traits is the person who shop be atop of this list: 1) persistence. all the games in 13x3 or 11x5 or 10x6 are winnable. but it takes someone who is extremely persistent to make it high in the all-time list. and these variants are fun when you just want to vedge out and crank out a large number of wins. 2)skill. the x0, x1, and the lower 4-6x's variants are all skill variants no question. in some of these variants less than 10% of the games are even winnable. and its really exciting when you win because not everyone does. 3)luck. whether its the variants that are 100% winnable or 5% winnable, a good streaker will need luck every now and again. even if 100% of the games can be won after 1000+ played even the best streaker can make a goof and lose all that hard work. that's why i think huffy's 10x6 streak is impressive. not because he dedicated the time or because it's so large, but because he has managed to avoid that late night sleep deprivation or drunken moment of play that ruins a long streak. and in the variants where not every game is winnable, you need some luck to string together winnable games in order to have a chance at a high streak. on a final note, i didnt like the idea of weighting variants because it would have introduced a huge can of worms. no one can agree as it is and that would just have added another confusing element. i could come up with a dozen logical ways without even trying on how to weight the streaks and i can't even agree with myself which one would hold the most merit. during a sales job i held once my boss told me the more complicated the bonus structure the more the employee is probably getting screwed. this comparison is not completely relevant but to me the more complicated the ranking the less weight it carries. happy playing to everyone. i've spent too much time here lately and not enough time playing. i've gotten my rank to 596.67 and looking forward to those elusive wins.

Date: Mon Jan 17 12:53:41 2011
User: CharitonIowaUSA
Message:
I spent some of my playing time the last few days going through variants, trying to figure out which ones I've played. I have a lot more to go through, but it's starting to look like I have not played about half the 70 variants. What I also found was CS with very recent streaks in lots of high-sum variants, especially the wides. It looks like CS is going to pop a 100-range or less total number on us one of these days. Another streaker who should rate well here is Loskazanes. I've battled him for spots near the top in lots of 11 to 13-sum variants over the years. But now I see where he has played in most, if not all, variants. Although he is not top 10, or even top 100, in many of these, he has lots of streaks still alive that rate in the top 300 or 400 range. So these will not totally kill him. I also don't think Loskazanes is currently active here. I agree with firenze that this rating system does not take the problem solvers' skill levels into account. That's too bad. Even though the easier variants carry the same weight as the more challenging, this is still the best rating system for streakers I have seen. Huffy is the real deal. I have seen his work in more challenging variants than 10x6, etc. He can play well in all variants, but seems to prefer the high sums.

Date: Mon Jan 17 13:54:21 2011
User: CubicSprock
Message:
So Chariton...are you Ed? I have been playing some of the wider variants, but they get boring pretty quickly...my overall average rank is up to 230, and I can see playing those enough to get sub 200, but I doubt I'll put in the time to try and get in the top 1000 (let alone top 200-300) in 9x5 or 10x6 because of how many games that would be. Also, the 13x0 is popular and mostly luck, meaning I could play 1000's of games without getting lucky enough to put a big streak together...and because it is popular you have to be very lucky to have a chance to climb high on the all time list...just not my preference for streaking. Overall, I imagine the 8x4 is daunting for most people, but I've managed to put a nice streak together there (still only rank 426). As for other players, I agree loskazanes would do well as I've competed with him in many of the tougher variants. I have to imagine kenwa would rate pretty well at this, he's played pretty much every variant, including a lot of games in the higher sum variants. T1-T3 is also a good streaker, with a lot of variants...missing a few less popular high sum variants, but could quickly move up in those if he was aware of this and wanted to improve.

Date: Mon Jan 17 13:59:24 2011
User: CubicSprock
Message:
and to answer the issue brought up about how quickly I got to 1 in 4x6, making that somehow unequal to 10x6...yes and no. Keep in mind the benefit of being #1 in that, is much less than being #1 in 10x6. Anyone who wanted to compete across all variants could win a single 4x6 which would put them at 38th for the variants, so getting 3 in a row only knocks 37 extra points off my score...so in terms of the average it is not very significant...and pales in comparison to the 1000 point difference in the 10x6. So looking at it that way, huffy put much more time into his 10x6, and appropriately has much more to show for it in terms of average.

Date: Mon Jan 17 16:36:08 2011
User: EZ-Ed3-LAnseMI
Message:
CS--Yes, to your question. Chariton, Iowa, is my hometown. The Chariton, Iowa, nick is about seven-eight years old, so I just figured it was known. I used it because that is what I was playing with in 9x3's when all this popped up. Like you, CS, I find this intriguing. But, all those wide 14 to 16-column variants are a vast journey in boredom. I agree, T1-T3 would rank high. kenwa is a good player. I just didn't realize he had played across the board so much. There are a lot of others who would do well. I ran a quick total for the approximate 35 variants I have found I have played. My average is below 120, but all those un-played variants would kill me. And I have to bring in several of my nicks to accomplish this. Now for the question...do I really want to put in all the time to play all those variants with only one nick? I don't think so. But...

Date: Mon Jan 17 20:49:05 2011
User: mightybass
Message:
since i've already done the leg work and several of you have speculated on how well some players would fair, here are the top 15 results i have calculated (these results are for anyone who has placed 1st in at least a single variant or have posted to this thread. i calculated these results about a week ago, so anyone who posted after that won't be in my list...sorry, i'm burnt out and not doing any recalculations). since it has been hinted at in some people's posts, i thought i would mention i do not combine nicks for people even if they are known to be the same person: Zonny 91.17 TitanicTony 250.00 ElGuapo 252.43 Boojum 281.71 Catailong2 321.36 firenze 351.09 Lawrence55Dee 357.97 jedirod 368.21 loskazanes 390.54 Kaos 400.09 T1-T3 413.27 SirPape 415.53 huffy 430.50 chariton: nice to meet a fellow iowan. i'm from waterloo and currently live in ames. i would disagree with you and firenze though that this ranking doesn't account for a problem solver's skill level. while it doesn't weight any variants more so than others, which means there is no emphasis on skill solving, all the variants are being rated in this so all types of play are accounted for. some is luck, some is skill, and some is persistence. you need all these traits to rank high across all the variants. i feel all these skills are important and that is why i didnt weight the ranking in anyway.

Date: Mon Jan 17 22:28:42 2011
User: Catailong2
Message:
Here's the list of the top 50 players using the system that credits you with points for each player you outrank (System 2). There is currently a max score of 62,306. I'll try to automate the other system as well although the results should be quite similar. 1. Zonny . . . (54,722 points) 2. Top-100-of-all-variants . . . (49,225 points) 3. canman1944 . . . (48,731 points) 4. CubicSprock . . . (45,198 points) 5. Repperzed . . . (45,055 points) 6. greggory . . . (45,034 points) 7. TitanicTony . . . (44,128 points) 8. kenwa . . . (43,878 points) 9. Gulbis . . . (42,769 points) 10. ElGuapo . . . (42,617 points) 11. raederle . . . (42,358 points) 12. arthur_fl . . . (40,827 points) 13. Laffman . . . (40,058 points) 14. Catailong2 . . . (39,731 points) 15. dirkie . . . (39,547 points) 16. Boojum . . . (39,177 points) 17. edv . . . (38,766 points) 18. mickyiw . . . (37,604 points) 19. SDBaroon . . . (35,885 points) 20. firenze . . . (35,822 points) 21. etelka . . . (34,804 points) 22. Tomcut . . . (34,768 points) 23. SilvioManuel . . . (34,699 points) 24. loskazanes . . . (34,193 points) 25. willem44 . . . (33,497 points) 26. babydaddy . . . (33,247 points) 27. jsj41 . . . (33,196 points) 28. mikmaq . . . (33,146 points) 29. jedirod . . . (33,143 points) 30. Lawrence55Dee . . . (32,883 points) 31. huffy . . . (32,835 points) 32. GlennT . . . (32,648 points) 33. RacingGenk . . . (32,559 points) 34. Kaos . . . (32,399 points) 35. T1-T3 . . . (31,937 points) 36. EMU . . . (31,866 points) 37. 417sec8 . . . (31,716 points) 38. deadwing . . . (31,196 points) 39. malr . . . (31,080 points) 40. SirPape . . . (31,052 points) 41. bjgesq . . . (30,743 points) 42. mercury . . . (30,597 points) 43. lenaspinoza . . . (30,506 points) 44. diederik . . . (30,358 points) 45. Mastermind . . . (30,112 points) 46. Darkosi . . . (29,543 points) 47. Pianomoods . . . (29,536 points) 48. wasjun . . . (29,379 points) 49. wfl . . . (29,280 points) 50. sillama . . . (29,051 points)

Date: Tue Jan 18 00:16:48 2011
User: mightybass
Message:
thanks for the long awaited update Catailong2. if you compare the common members from your list to mine, we can see that the ranks are very similar indeed from your system 2 to mine (which is essentially your system 1). i'm curious, what method did you use to analyze ties? i don't know if you ever commented on what approach you were going to take. also, you originally calculated a score for me of 15,812 on this system. i've played a decent amount since then and was wondering what my score has grown to?

Date: Tue Jan 18 01:34:11 2011
User: Catailong2
Message:
Yeah, it turned out to be a somewhat tedious project but the payoff was worth it. Now that the Excel files have been created, I should be able to go through and update the standings in less than 15 minutes. Most of the work has been set up to be automated so I just need to refresh each of the 10 files, copy & paste the results to a master sheet and voila!, there are the new standings. I'm interested to see how much the scores change over time. I've also been toying with some ideas on how to rate players on their solving abilities (via average time & win %). I liked what I read in Groucho's post (linked to in roo's post above). Regarding ties, I felt that the fairest way to treat them would be to just plain ignore them. If there's a variant where someone could jump a few hundred spots by completing a streak of just one more game, then that should provide the incentive to work on improving in that variant. As an example: If you've never played a 6x3 game before and then you go and win one, you get no points since you're still unranked on the 6x3 all-time list. Now if you win two games in a row, you suddenly become ranked 274th on the all-time list, picking up 726 points (since you passed that many ranked players who had only won one game). I feel that the benefit of this approach is that more players are likely to strive to win two games rather than settling for just one, thus increasing the quality of the results. Regarding your current score: System 1: 682.86 average (ranked in 48/70 variants) System 2: 20,712 total points

Date: Tue Jan 18 02:04:21 2011
User: TNmountainman
Message:
Just as a suggestion, I thought groucho's idea was pretty good, but requiring playing 1000 games in a variant, or for there to be at least 50 players to "count" a variant are both too restrictive. But I think he was using "current" info, not all-time. But I'm not sure about that.

Date: Tue Jan 18 12:10:56 2011
User: Boojum
Message:
For players who have used several nicks, I can see that it would not be feasible for mightybass or Catailong to take that into account. But there is no reason why a player cannot do the calculation him or herself and post the results. Speaking for myself, I'd take their word for it that they were using their own nicks.