Fastest Solvers

[euler]

For some time now we have been experimenting with live data as members have been solving new problems, and since problem 277 we have logged the fastest twenty solvers. Those observant ones amongst you may have noticed yesterday evening that these tables went "live". To view the tables you will be able to see that the number in the "Solved By" column on the Problems page has become a clickable link. If you follow this link you will be taken to a "Fastest Solvers" table.

It is our intention to change the current scoring tables so that they will be based on the impressive achievements of members who appear in these fastest twenty solvers lists. We are still refining the system, but we hope to make an announcement in the near future.

[TripleM]

Hey, cool. Obviously everything depends on timezones but I always look up the main tables to see which of the 1st places solved it faster than me, so definitely worth it.

Small bug - the first 7 problems on the list are linked, even when sorted by difficulty - in which case some of the links lead to a 'no data found' page.

[euler]

Thanks, TripleM. Bug fixed.

[robertking]

very cool, some fast solvers.

[phsr]

While honoring the fastest solvers, maybe there is a place to honor the newest dedicated solvers. If I count correctly, there are about 8 people in the Euler list who have solved fewer than 50 problems in total.
Having worked to get there, so understanding what it takes to get there, I think that those who are more dedicated than I was to this goal deserve to be recognized.
phsr

[albert]

The fastest solvers system certainly favours the students over the people who have to go to the supermarket each saterday with their wives. ;-(

Groetjes Albert

[phsr]

Maybe those of us who live in the wrong time zone or have other pressing obligations on Saturday mornings (or whose brains are no longer as swift as they might have been when students) would be more than satisfied if the scores displayed within each level were sorted by highest 'current performance' rather than by total problems solved.
I am currently #837 on the list of Level 2 members, but my 'current performance' might put me in the top 20. At least that would give me some credibility with the younger generation .

[GenePeer]

Just out of curiosity, since I'm still solving the first problems, is the time to solve counted from the time a user opens the problem or when it's made accessible?

[euler]

It's a measure of time since the problem went live. So, for example, Problem 320 (View Problem) went live on 15 Jan 2011 at 10pm (GMT) and the first solver had read the problem, designed an algorithm, programmed it, and entered his solution in 16 minutes and 44 seconds. It almost defies belief how quickly some of the "fastest solvers" do it, doesn't it?

[GenePeer]

It most definitely does!

[browni3141]

I like the Eulerlians section, but it seems a shame that previous bests aren't recognized somewhere. How about a page for the top 100 highest scores with the best rank achieved
Example:
Name: Anton_Lunyov
Highest score achieved: 177
Highest rank achieved: 1

[Stache]

I think there's an error in the fastest solver list. It's currently showing that I appear in 1/30 scoring tables, but I believe I'm in two tables (problems 328 and 330).

btw, whatever happened to the idea of increasing the tables to the fastest 50 instead of 20? Supposedly that data was being collected and retained, no?

cheers,

[euler]

You might appear in two tables, but every ten days the score obtained from that table is reduced by one point. You are 18th for #328 which is worth 3 points, but the problem is 32 days old which means that 3 points is subtracted giving you no points from that table. You are 19th in #330 which is worth 2 points, but the problem is 17 days old so 1 point is subtracted. This gives you a cumulative total of 1 point. The Eulerians table gives the cumulative totals and the performance bar indicates the number of scoring tables you appear in.

[Stache]

You might appear in two tables, but every ten days the score obtained from that table is reduced by one point. You are 18th for #328 which is worth 3 points, but the problem is 32 days old which means that 3 points is subtracted giving you no points from that table. You are 19th in #330 which is worth 2 points, but the problem is 17 days old so 1 point is subtracted. This gives you a cumulative total of 1 point. The Eulerians table gives the cumulative totals and the performance bar indicates the number of scoring tables you appear in.

Ah, got it...thanks for the clarification.

[sivakd]

Another suggestion to increasing the fastest solvers to 50 from 20 is

1) In the profiles page, show the list of problems the user is part of the fastest solvers list. This can be a separate table or use a different icon (instead of green) against the problem with a tooltip that indicates the position.

2) Start showing the profiles public like in the past, but only show for the fastest solutions. This is anyway public info, just that it would be consolidated

3) Provide the count for fastest solutions in the profile badge (image).

[albert]

Today I was late with looking at the new problem, and I had no idea how long ago
the problem was posted.
It would be nice if with the fast solvers scores there is a mention when (or how long
ago) the problem was entered.
It would also be nice (because i never end into the fast solvers) that with the congratulations screen is mentioned how many have solved the problem before you.

Groetjes Albert

[euler]

I've modified the scripts to display the publication date/time on the page that displays the list of problems: hover your mouse over the problem description/title. On the actual problem page if you hover over the short date at the top of the problem then it will display the same data.

I've also added a feature which displays, "You are the __st/nd/rd/th person to have solved this problem", on the "Congratulations" page. This will be real time so the problems list page may not reflect this immediately.

[sivakd]

It would be good to see that the difficulty of the problem is reflected in the scores assigned to the top 20 fastest solvers. One way is probably to use the total time taken to get to the 20th solver and the more it is, the higher the scores for that problem relative to other problems which took lesser time. Also, for the top 20, the scores would again be reflective of time taken for each position rather than a constant decrement by 1.

[sivakd]

Thanks for incorporating the speed into the score in the Eulerians table.

[sivakd]

In the Eulerians section the scoring for speed says

"For the twelve most recent problems the difference, d, in the length of time to solve the problem (in minutes) between each member and the slowest in the table is calculated and log2(max(d,2)) points are awarded."

Till I paid close attention to how the scores look like I didn't realize (IMHO) that this scoring is not right. Intuitively, the distribution of the intermediate solvers is not considered and importance is given only to the last solver. It may be computationally expensive and hence this was chosen, but I feel that averaging the same log value from i+1 to last-person for the ith person should give a better score.

Edit: I computed how the scores would look like for Problem 370 and here is the table. The last column is the score.

Code: Select all

Rank	Time -Days	Time - Hours	Time - Min	Time - Seconds	Total Seconds	Minutes	Score
1	0	1	46	32	6392	4124	12
2	0	2	0	6	7206	4111	12
3	0	2	6	29	7589	4104	12
4	0	3	20	7	12007	4031	12
5	0	3	31	43	12703	4019	12
6	0	3	45	26	13526	4005	12
7	0	4	22	11	15731	3969	12
8	0	4	52	13	17533	3938	12
9	0	5	1	13	18073	3929	12
10	0	5	8	27	18507	3922	12
11	0	5	45	20	20720	3885	12
12	0	6	50	43	24643	3820	12
13	0	7	50	56	28256	3760	12
14	0	9	2	13	32533	3688	12
15	0	9	44	36	35076	3646	12
16	0	9	49	45	35385	3641	12
17	0	10	41	44	38504	3589	12
18	0	10	51	13	39073	3579	12
19	0	11	5	16	39916	3565	12
20	0	11	28	23	41303	3542	12
21	0	11	51	50	42710	3519	12
22	0	12	39	43	45583	3471	12
23	0	12	43	24	45804	3467	12
24	0	13	25	33	48333	3425	12
25	0	13	31	18	48678	3419	12
26	0	13	39	1	49141	3412	12
27	0	13	59	29	50369	3391	12
28	0	17	40	53	63653	3170	12
29	0	18	53	27	68007	3097	12
30	0	21	29	20	77360	2941	12
31	0	22	13	59	80039	2897	12
32	1	4	7	42	101262	2543	11
33	1	4	48	8	103688	2503	11
34	1	9	44	42	121482	2206	11
35	1	10	16	20	123380	2174	11
36	1	11	31	22	127882	2099	11
37	1	17	29	16	149356	1741	11
38	1	21	21	57	163317	1509	11
39	1	21	21	59	163319	1509	11
40	2	1	18	45	177525	1272	10
41	2	1	33	18	178398	1257	10
42	2	1	33	22	178402	1257	10
43	2	2	17	2	181022	1214	10
44	2	3	14	59	184499	1156	10
45	2	7	2	6	198126	929	10
46	2	9	35	2	207302	776	10
47	2	13	9	33	220173	561	9
48	2	16	2	23	230543	388	9
49	2	19	47	52	244072	163	7
50	2	22	30	41	253841	0	1