;
How does one amass an increase of a 67.09 ELO score in a single day? If you look at "Purohit's" score for today, Sept 17th, he/she did just that. When you look at their recent play, it surely doesn't look plausible at all. Thoughts?
If he'd won all the winnables, that would have really been something.
I asked this (on another deal) in another thread, a few weeks(?) ago, to no response. I'm *not* a chaser of elo, but one can't help but see if after each game. I just lost this one, and is the ~6.5 pts. loss appropriate for a 1/19 climate 6 12x0? Just asking out of curiosity. Seems like a rather extreme penalty....
| 12x0 30017-6 | |
| Elapsed: | 16:51 |
| Game stats: | 19/1/5.3% |
| Elo: | 1977.048 ▼-6.445 |
ElGuapo would be the one to answer but he’s on extended sabbatical. Anyone else?
User: TheGodfather
Message:
How does one amass an increase of a 67.09 ELO score in a single day?
It's no biggie to do that. I might do that after losing 100 or more points in a day in 10x1 level 12. Uberman is correct.
To the TN question. That's the way it goes. It doesn't really discriminate the size of penalty by how difficult the game is perceived to be, if the game has been won. At least that's what I've observed.
TN,
Well the baseline for 12x0 level 6’s is 1310 which isn’t so crazy since most of the winnable ones are fairly easy. Then that game has been lost 18 times. Assuming, most dramatic case, that the losers of that game have been very highly rated players, the game would gain close to 8 ELO points for each loss, so 8*18=144. Add that to the baseline and we get an ELO of 1454 which is actually far less than the game’s actual ELO of 1737.
Since your ELO was so much higher you were “supposed” to beat it. You would have lost almost a full 8 (which is what happens to all the top people when they lose a game nearly every time) had it been rated 1454. Fortunately, on the original setting of the ELO the game must have already been lost a bunch so it was set higher than the baseline.
Still, it reminds us that probably only the 8x4 games have a proper ELO because 19 plays on a very hard game are not enough to get that ELO where it should be. May Freecell.net last a hundred years because then it will be fairly accurate.
I believe ElGuapo crafted things to try to make Elo converge more rapidly even for less frequently played games. But I'd need to have him talk about it. I was all spun up on it several years ago when we did it but I gotta admit it's over the event horizon for me now.
Thanks for the responses. And yes, ElGuapo put in a fudge factor that added (or subtracted) some incremental bump from *each* game in a variant category, so that probably accounts for the greater variance. No need to point it out to me, but that 1310 # (seems low to me) must have been 'published' at least a couple of years ago, and I'd have no idea where to find it now. Some of you amaze me at what you 'save'; I guess it's a product of how devoted one is to such esoterica?
https://www.freecell.net/f/c/elo_about.html
It’s right on the site
Thanks. It now shows 1319 - so I guess my loss caused it to jump up. LOL.
Ought be tad higher than current rating (one man's opinion):
| 2/13 | 23:44 | 7x4 8750-8 | Streak | 9:45 | Won | 52.8% | 1612 | 19 | 36 |
And for saving stuff: I have a notebook buried somewhere that I kept track of my first won winnables long before some person made a link...I was proud to solve in one try where dozens before had failed, sort of...too hard to do now, seems to me, though.
10x2 24765-12...a bit underrated (had traps, I believe).
As I've noted, I'm not an ELO chaser. But one can't help seeing it after every game. So here's another one that caused me to almost gasp. Does this seem right for a 1/27 game? I went to look and its rating is 1728. That just seems, intuitively, to be awfully low for a 1/27 game.:
| Game: | 11x1 4403-8 |
| Elapsed: | 17:28 |
| Game stats: | 27/1/3.7% |
| Elo: | 1974.58 ▼-5.932 |
Definitely underrated. Well…unless the other 25 losses were all by players with very low ELOs.
But the cool part is the elapsed time is the same as its ELO! That’s got to give you some solace.
Solace? Hmmm. No, not quite that. Bemusement? Indeed. I hadn't noticed, so thanks for that observation. And not that it matters, but that elapsed time is not 'solid'. I started the game last night, and was too fatigued to carry it out when I could tell it would be difficult. But had no idea *that* difficult. But I'd say I spent at least 10-12 minutes on it before giving up today.
Recall, though, that replays count in the game stats but do not count in elo. So if that 1/27 game represents a lot of a person stubbornly trying it over and over, the elo won’t change much.
Looks like an underrated deal to me Tn. Looks very challenging. But, that said, when I was looking at game stats for 11x1-8 I saw this ine:
| 17579 | 30 | 1 | 3.33 | 1407.735 |
Be happy you did not lose this deal.
Yes, I realized that joey, and imagined that was likely the case. But I can never trust myself to remember how the count works. Has that not changed a couple of times?
I see it's now 1/32. And the one you list is *really* crazy, Dr.B.
For a week or two last August he had it so that replays (and tournament games, same thing) didn’t count in the game stats as they don’t count for your streak stats or elo. The community complained loudly. Anyway it was too late to fix anything really, ‘counted’ games and ‘don’t-count’ games have been co-mingled in the game stats for many years.
That's sorta what I had gathered. I didn't try to keep up with that minutia.
6x7 309-12...properly rated...if a person tried to play this as analog, the movements back and forth and back and forth of series 7 and greater again again...utterly slavish tedium.
When you’re playing a variant like 6 x 7–12, it’s really not about the ratings of the individual games. Yes, quite a few of them are severely overrated. But then, a fair number of them are real bears to beat (or I don’t beat them) and yet you get the same score. At my level they’re usually about .9. (Meaning I get .9 for winning and lose 7.1 for losing.) Which means I ought to win about seven or eight of them for every one I lose. And that turns about turns out not to be too far from the truth.
I disappoint myself too often though...but challenge is itself of some reward.
Try this one Brewski if you get a chance. I found it way under-rated and wanted a bunch more elo points than what I got. But I’d be interested if you agree.
| Mode: | Streak |
| Game: | 6x7 20033-12 |
| Game stats: | 1/1/100% |
| Player Elo: | 1923 ▲ 0.875 |
| Elapsed: | 06:52 |
| Streak: | 16 |
| Today: | 5 |
I'll play it later this evening when I'm not baked; some game skill lost for me with THC.
Some years ago, I had to covertly play 9X5s (half to one dozen/week); governess pushed for stiffer play.
joey, I gave your game a try:
| Mode: | Non-scored |
| Game: | 6x7 20033-12 |
| Elapsed: | 02:26 |
Definitely a challenging game (I used all of the freecells at one point). But, it's a type of game I'm usually pretty good at solving.
It's cool how the "Anon" clones are pushing the scores into more accurate numbers. Thanks, Hop, Guapo, whoever...
I managed to win it so it couldn't have been too hard. elo was 1557 after I finished with 3 plays and 3 wins
Here's a game where the Elo score might be too high:
| Mode: | Streak |
| Game: | 7x4 2591-5 |
| Game stats: | 243/49/20.2% |
| Player Elo: | 1834.401 ▲ 3.716 |
| Elapsed: | 01:38 |
| Streak: | 4 |
| Today: | 4 |
Thank you , Doctor. The fact that you zoomed through it in under three minutes kinda makes me think there was nothing to be proud of there for me. (sigh)
I predict the highest Elo on this site shall be achieved ca. late 80s...Hop's teenage great-granddaughter will reach 2917+...
...(by then of course play'll be fully mental).
(You read it here first).
7x5 3361-12...more like 1700 rating, sure...
7x5 31530-12...underrated somewhat, I'd say.
7x5 22722-12...overrated somewhat, I'd say.
7x5 14535-12...overrated by ~150...
7x5 10316-12...ibid...
| 12x0 10073-5 | |
| Game stats: | 60/18/30% |
Overrated by a lot.
| Game: | 27356 7x5-12 | |
| Plays: | 2 | |
| Wins: | 1 | |
| Percentage: | 50.00 | |
| Elo: | 1596.192 |
Heavily underrated, one man's opinion.
A demand of skills needed...
|