IJDB Records Database
I wrote a Python program to scrape and analyze the data from an archived version of the IJDB records database. This program only scraped part of the IJDB. It may be much larger.
The records are saved in a single spreadsheet now, request if interested. If anyone has the real spreadsheet, I'd like to look at it. I'll provide help anonymizing it if necessary.
Here is the code and output: https://imgur.com/a/r58VX.
Some statistics from the IJDB records data:
# of records in the ijdb database: 6270
# of individual tricks in the ijdb database: 189
# of users in the ijdb database: 853
Showing records for: Gatto (trick, then number of catches)
7 Ball 2 Hand Cascade ss:7 1000
5 Club 2 Hand Cascade Double Spin ss:5 10800
7 Ring 2 Hand Cascade ss:7 368
8 Ball 2 Hand Asynch Fountain ss:8 324
9 Ball 2 Hand Cascade ss:9 305
8 Ring 2 Hand Synch Fountain ss:(8,8) 368
7 Club 2 Hand Cascade Triple Spin ss:7 750
9 Ring 2 Hand Cascade ss:9 332
10 Ring 2 Hand Synch Fountain ss:(a,a) 68
11 Ring 2 Hand Cascade ss:b 22
12 Ring 2 Hand Synch Fountain ss:(c,c) 16
10 Ball 2 Hand Asynch Fountain ss:a 15
6 Club 2 Hand Asynch Fountain Triple Spin ss:6 nan
8 Club 2 Hand Asynch Fountain Triple Spin ss:8 21
Top Ten Most Commonly Logged Tricks (trick, number of log entries):
5 Ball 2 Hand Cascade ss:5 458
7 Ball 2 Hand Cascade ss:7 405
6 Ball 2 Hand Asynch Fountain ss:6 321
9 Ball 2 Hand Cascade ss:9 214
5 Club 2 Hand Cascade Double Spin ss:5 202
8 Ball 2 Hand Asynch Fountain ss:8 165
4 Ball 2 Hand Asynch Fountain ss:4 156
4 Ball 2 Hand Shower ss:71 142
3 Ball 2 Hand Cascade ss:3 141
4 Club 2 Hand Asynch Fountain Double Spin ss:4 133
#statistics
Thanks to 7b_wizard for posting a link to the data.
Little Paul - - Parent #
I don't know where you scraped the data from, but I think I' still have a MySQL dump of the records data (and the appropriate bits of the other tables you need to make sense of it) which you could probably have access to if you want.
Here's a csv of the IJDb records data that I have which was taken from the MySQL dump LP mentions which I use to allow users to import their IJDb records to the Edge. The last 4 columns are my additions to match up IJDb trick definition to the Edge equivalent.
There will be some over lap between the two record sets because of importing. Unfortunately I haven't kept a record of what has been imported vs entered so the only way to filter will be to test for duplicates.
Stephen Meschke - - Parent #
Thank you for the anonymized IJDb Records Data.
Here is a summary of the complete IJDb records data:
# of records in the ijdb database: 34419
# of individual tricks in the ijdb database: 432
# of users in the ijdb database: 1149
TEN MOST COMMONLY LOGGED TRICKS (trick name, number of times logged):
5b cascade 2643
7b cascade 2035
6b fountain 1430
5c cascade 1080
4b fountain 627
4c fountain doubles 591
4b Mills Mess 576
4b shower 558
3b in one hand 551
5b (6x,4)* 539
I have noticed a few interesting things comparing the IJDb Records Data with the Juggling Edge Records data:
- 1. The IJDb records are for less difficult tricks.
- 2. Cascade, fountain, and shower patterns are more common in the IJDb, while siteswaps are more common in Juggling Edge. (results: https://i.imgur.com/VtCvare.png
)- 3. People work on tricks in pairs.
1 & 2 are probably due to differences in record input methods. The list of available tricks at the IJDb was hard coded, to get a new trick added you had to suggest the new trick which was then reviewed by an editor who added it to the list of available tricks. Here we have a defined language for describing tricks which is validated automatically so getting a new trick into the Edge database is a much faster process.
I do also think it reflects a trend in juggling. Many young technical jugglers today are very familiar with many siteswaps, siteswap theory and are of a higher general level than ever.
Stephen Meschke - - Parent #
I made this graph: https://imgur.com/ugOWm23 to compare the general level of the IJDb members to the Juggling Edge members.
With the exception of the IJDb being much better with 5 rings, general level seems similar. 7 isn't the new 5.
Little Paul - - Parent #
How much of the ijdb being "better a 5 rings" was down to Norbi?
Stephen Meschke - - Parent #
There are 455 observations for 5 ring cascade in the IJDb data. It's unlikely that a single juggler could have the much influence.
Looking through the 5 ring cascade data I found this observation:User: 11756, Catches: 2100, Duration: 660 seconds, Date: 2012-02-24
Little Paul - - Parent #
I think I can probably say that 11756 was Arttu Lahtinen (not a name I remember to be honest, presumably he was one of those who hung out on the records section and didn't post on r.j or chat in the chatroom)
Norbi was userid 843.
Other people have mentioned that there is crossover between the user base for the two record sets, but I don't think anyone has explicitly pointed out that the edge has (had? I've not checked recently) an "import my records from the ijdb" function - which will have directly duplicated part of the data set.
What we really need is a proper statistician to look at the data and dig out the useful/interesting bits... now where could we find one of those?
Normally I would accept the page LP.. but having just spent more than 3 hours making an 8 minute video on basic statistical distributions for engineers (damned captioning of mathematical terminology) I have far too much work on my hands at the moment...
Funny thing, I'm working/learning about mixtures of exponential distributions. Do you know any helpful webplace?
I'm afraid that mixtures of exponentials can be a lot more complicated than just a single exponential distribution [which is one of the most basic distributions around!]
Dee has the good reply. Wiki is better than Wolfram for the single exponential distribution.
I'm afraid a good lot of the Edge members have also been IJDb members and previously logging their records there, and now here, as the Edge is a successor of the IJDb (it's not two wholly different \ separated groups of members). So, the sense in comparing can mainly be to e.g. see a development in time or to find out e.g. how the new platform Edge impacted on logged personal records? Or what are you aiming at?
I mean .. the data would then be redundant.
[your graph-pic direct link : ] https://i.imgur.com/ugOWm23.png
Even merged, it would partly contain old and newer records of the same users, and partly only old, partly only new data.
It looks like Juggling Edge leads in 10 out of 26 categories and IJdB is the overall winner. I suspect that more sophisticated statistics would confirm that hypothesis. Of course they aren't independent samples so there isn't anything to be said about comparing the two overlapping populations. It's more likely different sampling techniques or different ways of entering the raw data.
Stephen Meschke - - Parent #
The IJDb records data contains only unique users ID's, not users names. There is no way to relate users data between the two databases. This is the method I used to generate the results for the graph:
##read Juggling Edge data
je_df = pd.read_csv('/home/sm/Desktop/records.csv')
##read IJDB data
ijdb_df = pd.read_csv('/home/sm/Desktop/IJDB.csv')
##compare these tricks
for trick in tricks:
##find the average of the IJDb records
idjb_records = []
for record in df.values:
if record[10]==trick[0] and record[9]==trick[1] and record[12]==trick[2]:
idjb_records.append(record[3])
print 'Average IJDb record:', np.average(idjb_records)
##find the average of the Juggling Edge records
je_records = []
for record in je_df.values:
if record[3]==int(trick[0]) and record[4]==trick[1] and record[6]==trick[2]:
je_records.append(record[7])
print 'Average Juggling Edge record:', np.average(je_records)
I do have a grad student looking at stuff relating to partially overlapping samples (so not independent, but not completely matched either)... I'd have to have a good think about how to relate this to the IJDb records
Stephen Meschke - - Parent #
By the way, the documentation in Records Help (https://www.jugglingedge.com/help/recordshelp.php) is really nice. Thanks.
Stephen Meschke - - Parent #
7b_wizard sent me this link: https://web.archive.org/web/20120412063629/https://www.jugglingdb.com/records/record.php?trick=16. I went from there.
6 ball fountain color change transitions: https://youtu.be/rvYHKllbu-o.
In the 6 ball fountain pattern, three balls are juggled in one hand and three in the other. The balls don't cross or mix. Left always juggles one set of three, and right always juggles the other set. I use three red balls and three blue balls to make the pattern easy to see.
In this video (link above), I use three different transitions to cross the balls to the opposite hand. The siteswaps for the second and third transitions are 7777770 and 777771. What's the siteswap for the first one?
Are there any other transitions?
For a trick where they all swap, b97531 would be a nice one with your 2 sets of balls. Also 999333 could be cool.
Stephen Meschke - - Parent #
I tried b97531 today. It did look really cool, even though I could only run it to 5 (out of 6) gathers.
Thanks for posting your video, Stephen. Your 6 ball pattern is very solid with high throws. It would be great if you could try to break the current "slow juggling" record for 6 balls (i.e. fewest catches in 1 minute).
The current record on Juggle Wiki is 240 catches in 1 minute by Steve Hogan (video).
Stephen Meschke - - Parent #
I doubt I will try.
There was some interesting data on the Speed Juggling Records page you linked to. I made a graph: https://i.imgur.com/x407e97.png.
#statistics
Orinoco: Rendering a graph in the text box would be awesome!
That seems to mirror the invert proportional influence of gravity for more balls: a) balls get faster coming from greater height, so b) greater height is needed to even space them.
Thanks for the graph. I agree that the "narrowing" of the gap between slowest and fastest is based on fundamental physical limitations. The "premature narrowing" at 6 balls is, I think, due to the fact that jugglers haven't intentionally tried to do fast and slow versions of 6 balls (unlike 5 balls), and that 6 balls is generally less popular than the odd ball patterns, at least for endurance. I found those 6 ball videos through searching for every video of a 6 ball over 1 minute and taking the fastest & slowest.
Stephen Meschke - - Parent #
Due to fundamental physical limitations the fast and slow juggling records will converge where numbers juggling starts at 8 balls.
This source contains inadequate data. The records are for runs lasting 60 seconds, that's too long for 6 and 7 balls. Changing the duration of the record would increase the quality of your data. The duration of the record should be something like 10 times the number of balls, in catches. For example, the 5 ball speed records would be based on 50 catches.
Brook Roberts - - Parent #
https://youtu.be/5wqR76KZl4g?start=69 has some swaps in the same vein.
Top ten most frequently logged records in JugglingEdge Records.
There are 10273 records logged.
There are records for 1084 unique tricks.
The ten most logged tricks are:
7b cascade was logged 432 times.
5b cascade was logged 430 times.
6b fountain was logged 282 times.
5c cascade was logged 264 times.
3b back crosses was logged 133 times.
5b 744 was logged 132 times.
8b fountain was logged 121 times.
6b synchronous fountain was logged 117 times.
9b cascade was logged 114 times.
3b in one hand was logged 112 times.
#statistics
Little Paul - - Parent #
The clustering is interesting there.
Do people work on tricks in pairs of something?
Little Paul - - Parent #
I can't afford the airfare to France every time I want to learn something new.
Little Paul - - Parent #
I may have misread my own post - it's one of those days!
My immediate thought is, what is 3b back crosses doing in there amongst all those 'numbers' tricks?
Can we also get "people that logged X"? Something like "7b cascade was logged by xxx different people".
https://www.jugglingedge.com/records.php?PropNumber=7&PropType=b&Trick=cas .. 50 people. .. [ Main Menu "Records" --> linky "search records" --> "Number of props": drop-down-choice "7" + "Prop type": "ball" + Siteswap: [empty or ``7´´] + "Pattern": "cascade" ] .. ;o]p
Stephen Meschke - - Parent #
There are 73 different people that posted 432 records for 7 ball cascade.
I am using Python to look at this data, here is a link to an image of the code's text: https://i.imgur.com/tcUnZi2.png.
I would really like to just type the 6 lines of Python code in to this text box, but I can't find a way.
You can use the <code> tag:
import pandas as pd
records = pd.read_csv('/home/sm/Desktop/records.csv')
sbc = []
for record in records.values:
if 7==record[3] and 'b'==record[4] and 'cas'==record[5]:
sbc.append(record[2])
print np.unique(sbc)
print 'there are', len(np.unique(sbc)), 'jugglers working on 7 ball cascade
I'll let the typographical error in that be an exercise in debugging for the reader...
I see three, but I'm not sure if two of them are due to the formatting of the code block or your copying.
Stephen Meschke - - Parent #
There are 4: 2 code block formatting, 1 typographical, and Orinoco forgot to import numpy as np.
Ah, I ignored that as it an expert from the whole code :P
Another one would be that the code block allows wrapping, which would have lead to multiple addition issues...
Stephen Meschke - - Parent #
What's your preferred method sharing Python code? I don't think I'll ever be able to get it into this text box, which would be ideal. My preferred alternative is posting an image (like I did) of the code running, the drawback is copy/paste doesn't work. I don't like Github either.
Little Paul - - Parent #
I'd have gone github (either sharing the repo, or a gist) - but if you don't like that you won't like my bitbucket/gitlab.com/assembla suggestions either...
If you're not particularly bothered about people being able to contribute changes back to you, there's always pastebin
7b is 21 times harder than 5b (data from jugglingedge.com/Records)
According to data from the Records section of this site, 7b cascade is 21 times harder than 5b cascade (see full data and calculations in the spreadsheet below).
My calculations are simple. I merged the data from 5b cascade Records with 7b cascade Records.
To find the relative difficulty for each juggler, I took 5b Personal Record, devided by 7b PR.
For example, Orinoco has a 5b PR of 1412, and a 7b PR of 72. His relative difficulty is 19.6:
7b PR / 5b PR = Relative Difficulty = 1412 / 72 = 19.6
The average relative difficulty for all jugglers is 21.
I hope this provides solace for those learning the 7b cascade. On average a juggler with a 5b cascade PR of 210 only has a 7b cascade PR of 10 catches.
Jugglingedge.com records can be downloaded in one .csv file (at the bottom of the Records page). My hope is to use this data to find my strengths and weaknesses, and to more efficiently allocate my practice time.
Link to spreadsheet: https://docs.google.com/spreadsheets/d/10DM2Xw_MWW2gJQF1j5Wp-mzeiRSaKqvW60ySwzdND84/edit?usp=sharing
5b data: https://www.jugglingedge.com/records.php?PropNumber=5&PropType=b&Trick=cas&
7b data: https://www.jugglingedge.com/records.php?PropNumber=7&PropType=b&Trick=cas&
um........why do I have the lowest "relative difficulty"???? Is that bad news doc? How long do I have to live???!!! ;)
it means that you can (more) easily improve 5b endurance (rather than improve with 7b not equally improving with 5b) .. easier said: with your level of 7b cascade, you would - in average O,o - expect to endure way longer with 5b.
Over the years, I found a whole bunch of differences between learning the 5b and the 7b cascade (where you might think, there wasn't any, but "higher & faster & more narrow") ..
- posture: you don't have to necessarily look up and put your head into your neck soo much with 5b,
- posture: you can keep 5b up juggling (too far) ahead in front of you (more or less easily) for a whole while
(30-50-100 throws?) which - with 7b - exhausts after two or three periods - when first getting to that level -
and makes your arms burn after a while (so with 7b, you'll have to soon or late come near the body).
- pattern: the top of the pattern is way more compact with 7b .. there's two to three balls at the top to cope
with (find a `tunnel´) and you can partly throw (\aim) at(!) the last ball (that will be out of the way when
your throw gets there) .. also you can throw 5b in a way less acute angle, if you don't juggle so high,
- the balls, with 7b, - coming from pretty much higher - land considerably faster, thus, you'll have to either
react earlier (the ball being then higher) or faster,
- having to throw higher with faster handmovement requires another throwing technique (or more force or
rowing more and faster with arms),
- five to six balls in the air doing 7b is really more than the brain-eye-coordination can take when still lear-
-ning (with three to four balls in the air doing 5b, you have one about to land and two high up out of the
way and out of `panic-threshold´, one about to reach its highest point - when learning 7b, there's
always two to four balls in a `blind´ zone, where you can only rely on having thrown them well in height
and beat),
- ( .. ? )
.. more differences arise when considering fast correction throws to be made or saving a pattern out of beat or out of shape (explosiveness, fast & efforty-ish skills).
(( = all in all 21 x harder ;o]) .. ))
Stephen Meschke - - Parent #
I updated the spreadsheet (linked in the thread starter) to compare the difficulty of 645 and (6x,4)*. I calculated that 645 is twice as hard as (6x,4)*. Those results look a lot better. This method works better for patterns that are of a closer in difficulty level.
After some consideration, 5b cascade and 7b cascade are such different levels of difficulty that no real comparison can be drawn. Aside from the points that you raised, general physicality may be a problem for a lot of jugglers. Going over 100 catches with 7 balls requires quite a bit of physical stamina. I think that a (physically) average joe can juggle 5 balls for quite some time, but it takes athleticism (from training) to keep 7 balls in the air for longer than 20 seconds.
There is absolutely no way 645 is twice as hard as (6x,4). 645 is if anything slightly easier and if you look at little more closely at your data, you will see better jugglers tend to have longer runs of 645 than (6x,4). You also have data such as lucasgabd, where it looks like he just hasn't really spend time on 645, skewing the data.
| lucasgabd | 116 | 18 |
I thought that would copy & paste the statistic in for lucasgabd... His runs are 116 (6x,4) and 18 (645)
One could strike most outstanding data falling out of the average and would surely need lots of data to get such meaningful average.
(html injection attack edited out :P)
Also don't delete the outliers!
Mybe (6x,4)° is slower, therefore easier? .. and you have the 6 liable to collide with the 5.
Mike Moore - - Parent #
I'm surprised you feel that 645 is easier...my longest run with it was probably 18-30 catches, while I can run (6x,4)* for a while (>1 minute). I've practised (6x,4)* for around 2-3 times as long as 645.
I'm also surprised that (as you point out) the very good jugglers in the spreadsheet tend to find them about even in difficulty!
It is a fundamental flaw of this kind of analysis: people need to take the patterns about equally seriously. I don't think that's the case with 645 and (6x,4)*.
Stephen Meschke - - Parent #
This analysis works best if the patterns are very popular, taken seriously, and can be run for 100+ throws.
If only the data from 'good' jugglers (100+ catches of each pattern) is considered, 645 is only 1.14 times harder than (6x,4)*.
Relative difficulty changes for pattern duration. For example, it's probably several times easier to qualify (6x,4)* than 645, but only slightly easier to run (6x,4)* for 100+ throws.
It's also a difference, if someone is still learning or, or if they `got it´ and doing it as matter of endurance .. even though already enduring for many catches (e.g. with effort or ardouus savings), wholly mastering a pattern easily fluently isn't yet granted. The compared patterns - once mastered - might give different difficulties, than when still learning them ..
If "taken seriously" could be taken into the statistic by evaluating frequently logged data e.g. by a factor or by striking all single logs ..
Daniel Simu - - Parent #
I find running (6x,4) muuuch easier than running 645, I've practised both just a little bit but I can't do more than 2 rounds of 645 while I can easily hold 6x4 for at least 10 seconds, maybe half a minute
For the 7 ball versions I find (8x,6)* easier than 867 because the collision problem becomes more of a problem.
.. are \ were easier for you to learn and \ or to endure or perform (both 5b- & 7b-versions) ?
Regarding the statistical approach, I'm afraid, there won't be enough data to be found of jugglers doing 7b-versions (and both of them and equally `seriously´), .. so statements of jugglers doing these is worth way more as a #difficultyRating, than any #statistics could be for time being.
I see both approaches (statistics + polls) complement and or fruitfully compete with one another.
Daniel Simu - - Parent #
Yes.
6x4 has less different heights to control
6x4 has crossing 6's, high throws are easier for me when they cross
6x4 is less collision prone
6x4 repeats itself more quickly
It is a statistically valid approach to rate difficulty by catches done. It should, though, then be considered, personal favorings or popularity of patterns distorting the results. And interpretation will not give clear statements. As additional to polls or jugglers' ratings, it could be very interesting to compare.
JackJuggles - - Parent #
That's pretty smart of you to think of that. I've been doing that stuff in school(not the juggling part, just the math part) and that never came close to my mind
JackJuggles - - Parent #
That info could help a lot. If you recorded how many hours it took to learn the 5 ball cascade, let's say 1000 catches, the average person would have to calculate how many hours multiplied by 21 to find roughly how many hours it would take to learn 7 to 1000 catches. I will experiment this if I ever juggle 7 balls
Daniel Simu - - Parent #
This is amazing! I tried yesterday to see what my records are, I tried 5b and 7b both for about 5 minutes, I got a run of 603 with 5 and 30 with 7! That is very very close to your statistic, this seems super realistic!
Cool :D, now I'll know how to explain how hard 7b is in the future :).. And I know how far away I am from 100 catches.. :(
Stephen Meschke - - Parent #
Thank you.
Please log your progress every couple weeks in the Records section of this site. This will improve the quality of the data.
My hope is to use the data to find efficient practice strategies. I want to find links between things like practicing 744 and 7 ball cascade progress.
There's some more data around on the web .. e.g. on archived IJDb https://web.archive.org/web/20120418145054/https://www.jugglingdb.com/records/index.php.
And I found a chinese records site https://juggling-donuts.org/record/local/patio_jug/index.html. The sections for 5b cascade [in the first paragraph] is 5ボールカスケード , for 7b cascade it's 7ボールカスケード and 5b-744: https://juggling-donuts.org/record/local/patio_jug/index.cgi?mode=seeenmokudata&enmokuid=2-0-744 .. but it's lotsa work to decode the chinese signs per e.g. babelfish https://www.babelfish.de/dict?query=%E3%83%AB%E3%82%AB%E3%82%B9%E3%82%B1%E3%83%BC%E3%83%89&src=auto&dst=de&submit=%C3%BCbersetzen or some online-translator
Stephen Meschke - - Parent #
I don't understand. Who are you quoting? and why?
I'm quoting 7b_wizard, because he made a mistake.
...& his posts have been much more readable recently!
'kay. However. I'll make sure then to include some general information for everyone in every my replies ;o])
pumpkineater23 - - Parent #
All this blocking is making me paranoid..
Stephen Meschke - - Parent #
I apologize for your paranoia. Ignoring users posts was a gaffe. I'll consider others ideas more carefully in the future.
pumpkineater23 - - Parent #
Thanks Stephen but it's not just you.. with all this blocking going on I wonder if anyone has blocked me. I think it might be helpful if a blocked person was notified that they had been blocked. Not just to avoid confusions but also for their own personal benefit.
Little Paul - - Parent #
and he's not even a real plumber
I'm not sure how that's relevant
pumpkineater23 - - Parent #
Just as I expected LP! I knew you'd try to cover up all your sneaky blocking with a joke. So then, how many of us have you blocked? Who has access to the special LP clique?
Little Paul - - Parent #
I haven't blocked anyone.
How can I take the piss out of what I can't see?
pumpkineater23 - - Parent #
Typical blockers response LP. And I wouldn't believe Orinocco for one one second either.. he's another classic blocker, an obvious member of the clique with all the blocking power at his fingertips no less.
Don't worry, you are currently being blocked by all of zero people. In the grand scheme of things I think there is very little blocking going on. I'm reluctant to add notifications of being blocked because then it runs the risk of becoming a tool to spite someone.
Daniel Simu - - Parent #
Stephen, I second Orin! You can unblock 7b_wizard :)
lol .. I've become a real polarizing stumbling stone here by using some punctuation and long nested sentences haha
Stephen Meschke - - Parent #
Hello 7b_wizard,
Thank you for your post. The data on the IJDb looks good. It will be easy to scrape the data and save it in a CSV file. I'll just copy the format that is used on Juggling Edge.
Daniel Simu - - Parent #
That would be amazing! Finally good reasons to update records :)
I like the idea of statistics telling me that there's a trick or siteswap that I didn't think of, that I could easily learn or do many catches of.
Daniel Simu - - Parent #
Ugh, today I tried 5b again, when I was warm but still full of energy, and got 900 catches on the first and only try.. I hope I can match my 7b record (43 catches) soon!
At least you give me a motivation to update my records, and try some long runs :)
It's not only about "long", but by doing long runs, you will partly have to fight for the pattern to stay up, which will train your recovering abilities a lot. And after saving the pattern from flaws with effort or under high tension, you will want to recover and find back into a ground-state that is as relaxing as possible to further endure it, .. so long runs will smoothen your ground-states and merge into swifter and poised technique in general. - At least, that's the experience, I'm going through at present.
Daniel Simu - - Parent #
Ugh, and today the third time I try 5b I get 1468 catches. My 7b record remains the same (30), but my believe that the 5b cascade and 7b cascade are very correlated is gone...
Daniel Simu - - Parent #
Nope, I can't decode " %gt;=1000 c"
Wanna rephrase?
yeh sry .. typo (meant: > which is html-code for "greater than") >=1000 c.
I set my 5b cascade goal to 1000 catches or more recently, but it turned out to be hard after fighting myself up to almost 900 throws.
Daniel Simu - - Parent #
So your smileys look like typos, and your typos look like they are supposed to mean something? wow!
I'll set the record to 2000, soon.... I don't suffer from much physical fatigue, and so far the mental fatigue has improved so quickly that I can improve the record by 1.5 every time I tried!