วีดีโอ

Absolutely. For example the German Goalkeepers are quite similar in skill but ter Stegen age 32y is worth €28mio vs Neuer age 38y €4m. I am planning to do a bigger analysis with all 622 players where I also show the R code. There I will consider age, games played and other factors. Also, Mbappe worth €180m moved from Paris to Madrid this year without that payment because his contract ended.

@@TheDataDigest cool 👍🏻

Guter Kommentar :) Besten Dank zurück. Hatte mich einfach interessiert wie aussagekräftig der Marktwert bei den Ergebnissen ist.

Thanks for watching my video and the nice compliment. I think I can still improve but I am glad you found the instructions helpful.

Can you give me a timestamp to the plot you are referring to? You can always add a line with "+ geom_vline(xintercept = median(data$column))". Or geom_hline(yintercept = ...) for horizontal lines.

What error message are you getting? Did you load the tidytext package to have the reorder_within() function available?

This information was not included in the tournament csv files that I analyzed for this video. There are however files for each tournament that contain the game information. I am also interested in such questions like "flagging %" and decisive games depending on titles and rating differences. And the bongcloud opening or the cow :) But to analyze these huge text files I have to write some specific functions first (or let ChatGPT do it) :D

@@TheDataDigest Thanks for the swift answer! I am not able to sift through that data myself, so I was hoping that someone had done the job for me ;-) I am curious because of the "clash of claims" match - a comparison between 3+1 and 3+2 flagging rate would be interesting. I suspect that Kramnik wants 3+2 because he's afraid of flagging, and that makes the "clash" a bit weird since it's not titled tuesday format. But without data I am disinclined to propose that theory.

@@kristianfagerstrom7011 Titled Tuesday is 3+1, right? Would you have examples of 3+2 tournaments where I could download some games to check your hypothesis? I want to look into these game analyses very soon.

@@TheDataDigest Yes, TT is 3+1, and no, sorry I don't have any tournament data, but maybe chess.com would be willing to share?

Wow, what a great compliment. Glad you liked. It took me a long time to make and I am happy when people find it useful, share it and leave a comment.

Glad you found my video helpful then :) If you need more help with your stats project you can always reach out to me via: question@thedatadigest.email

Thanks. It is my best performing video so far. Maybe I should make one about ggpairs which is also a good first visualization package.

Absolutely, if it's not too much hassle....they're of great help

Use the str() function on your data object, like "str(my_data)" and tell me what it says. It probably is not a data.frame. But you can turn it into a data.frame with the "as.data.frame(my_data)" or simply "data.frame(my_data)" functions.

Thanks for the compliment :)

I think in this video I only used ggplot2 library. It is part of the tidyverse package, which also includes dplyr. I also often use forcats (to work with factors), lubridate (to work with dates), and stringr (to manipulate strings). I also saw this tweet today about using ggdist package for distribution plots (twitter.com/MatthewBJane/status/1780952366541385910). Hope that helps :)

Feedback like this is highly appreciative. Thanks for leaving a comment. I do cut out all unnecessary speech so my videos are actually quite fast. But I want to show a lot and one can always rewind and watch again.

Should I make an update for 2024?

Glad you liked it and left a comment. Helps the channel with engagement and promotion.

What a cool first comment :) Thank you very much. Glad you liked it.

Thank you so much for leaving a comment. I am glad you liked it. It was definitely fun to do the analysis.

I think you are capable! If you are curious about a topic, or like to learn some form of data analysis there is always a simple step to be taken to get a little bit better. Start with some easy text books that have loads of examples. Or listen to david robinsons tidy tuesday screencasts on youtube.

Well, thank you :) But in all honesty i don't think i could. My intelligence is almost decent, but the work you put in, the research and hours. it's also beyond me.@@TheDataDigest

Oh wow, these are really good suggestions. I like the rating comparisons by different brackets of difference. And whether it makes a difference who has the white pieces. I think this would be great for some models. I will look into it!

Yes of course. @Kivallara also asked for it. I am currently moving it into my GitHub and then pin the link to it in the top comment.

Dataset can be found here: github.com/TheDataDigest/EDA/tree/main/Chess Have fun analyzing.

@@TheDataDigest thanks a lot and keep doing the amazing things

I missed that controversy but found some interesting articles discussing the matter. Thanks for bringing it up. Now I better understand another comment that mentioned Kramnik and Hikaru before. But the winning streak analysis did not show any big surprises. Six players with 12 games, Hikaru with 15 and Magnus with 17. Seems fine given that Magnus is the GOAT and Hikaru has an incredible Blitz rating and participates so many times in TT.

Glad you liked it Patrick and thank you for leaving a comment.

Hi Tristan, first of all, you are correct. Second of all, thanks for leaving such a nice comment. I am glad you enjoyed the video. I did some more analysis and only found two cases of players both playing und men and women titles. 1) Meri-Arabdize (from GEO) participated 26x as WGM and 8x as IM. 2) Jiner Zhu (from CHN) participated 10x as WGM and 5x as GM. The best placed women gets $100 in each tournament, so I assume that most women will play under the women title to qualify for that. Unless chess.com has another way of know or they don't care that much about the prize money.

@40:42 I briefly show the winning percentage but I forgot to make a chart. Is it okay when I list the top 10 highest winning percentage with (wins/participations) in parenthesis? Below I will answer the same with the chance to place in the top 3. 1) Liem Le, VNM: 30% (3/10) 2) Hikaru Nakamura, USA: 24.3% (18/74) 3) Magnus Carlsen, NOR: 23.1% (9/39) 4) Maxime Vachier-Lagrave, FRA: 20% (5/25) 5) Platon Galperin, UKR: 16.7% (1/6) 6) Eduardo Iturrizaga, ESP: 14.3% (1/7) 7) Shakhriyar Mamedyarov, AZE: 12.5% (1/8) 8) Daniil Dubov, RUS: 12.2% (5/41) 9) Alexander Grischuk, RUS: 11.1% (3/27) 10) Nihal Sarin, IND: 1.11% (5/45)

Cool to see that Caruana and Artemiev make it in the highest % to place as top3. Thanks for the extra request/question. 1) Susanto Megaranto, IDN: 66.7% (2/3) 2) Magnus Carlsen, NOR: 46.2% (18/39) 3) Hikaru Nakamura, USA: 43.2% (32/74) 4) Eduardo Iturrizaga, ESP: 42.9% (3/7) 5) Liem Le, VNM: 40% (4/10) 6) Khumoyun Begmuratov, UZB: 33.3% (1/3) 7) Maxime Vachier-Lagrave, FRA: 32% (8/25) 8) Fabiano Caruana, USA: 26.8% (11/41) 9) Vladislav Artemiev, RUS: 25% (1/4) 10) Aram Hakobyan, ARM: 24% (12/50) I should probably also do a top 5 which stands for % winning prize money.

Great catch. Very precise use of language, but you are correct. So in order to show the average of what I meant I would have to look at the score of number 2 and then add half a point. Or when there is a tight live with the fact that the Sonnenborn-Berger Score will resolve this.

I would love that. I am actually planning to analyze all games to find out what the most common opening was and the rarest one and how often a player wins on time etc. Edit: I first did not understand the reference and the matching emoji. But another comment made me aware of certain accusations 😅

Cheating Tuesdays FTW!

You are of course right. Judit Polgar comes to mind. I checked the data again and I could only find two women that are listed with women and men titles. Maybe there are some that use the men title only but I wouldn't be sure from the name alone. I found a list of 41 women with GM title and did some spot check of the most recent ones but could not find their name in the data set. The 2 I found are: Meri-Arabidze (GEO) with WGM and IM and a best rating of 2745 Jiner Zhu (CHN) with WGM and GM and a best rating of 2692.

Hi there. Very good suggestion. I was thinking about several Shorts that just go through some of the results. Maybe with some interactive charts. But you are right, maybe a short ~5 min video would be best for that.

@@TheDataDigesthell yeah 🦅

Yes I can, here are the top 10 players, that participated at least 5 times in 2023, sorted by average rating with their average blitz rating and N-participations: 1. Hikaru Nakamura: 8.64 | 3289 | 74 2. Wesley So: 8.59 | 3102 | 22 3. Magnus Carlsen: 8.51 | 3268 | 39 4. Dmitry Andreikin: 8.35 | 3055 | 84 5. Denis Lazavik: 8.25 | 3049 | 40 6. Yu Yangyi: 8.14 | 3051 | 7 7. Bogdan Daniel Deac: 8.11 | 3033 | 59 8. David Navara: 8.08 | 2943 | 6 9. Aleksei Sarana: 8.06 | 3066 | 74 10. Jan-Krzysztof Duda: 8.04 | 3030 | 60

I will soon share the data and scripts on GitHub, then you could answer these kind of questions yourself, if you download R/R-Studio, which is free. Only if you are interested in data analysis of course :)

Thanks, that one dot really caught my interest, turns out it was Wesley So. Once again, great video keep it up 😁

@@TheDataDigest Well done video and looking forward to try to analyze the data myself. What's your github?

@@Kivallara github.com/TheDataDigest/EDA/tree/main/Chess But I also pinned it as top comment.

Thank you so much for your comment. I am glad you enjoyed it. It was actually performing really poorly in the first days with regards to click-through rate and views, but now it seems that the algorithm is slowly starting to find potential viewers that might enjoy it. So again, thanks for engaging by leaving a comment and subscribing.

6 month or 183days exactly :)

Hi Lucy, thanks for leaving a comment. If you run the following code: test_df <- data.frame(x = 1:10, y = 1:10, shapes = LETTERS[1:10]) ggplot(data = test_df, mapping = aes(x=x, y=y, shape = shapes)) + geom_point(size = 3) You get this warning message and now shapes for the letters H-J. Warning messages: 1: The shape palette can deal with a maximum of 6 discrete values because more than 6 becomes difficult to discriminate; you have 10. Consider specifying shapes manually if you must have them. You can manually overwrite this with the "scale_shape_manual()" function: ggplot(data = test_df, mapping = aes(x=x, y=y, shape = shapes)) + geom_point(size = 3) + scale_shape_manual(values = 1:10) However I recommend, if possible to either group your levels, or work with colors. Like using 4 shapes but with 2 different colors for a combination of 8. Or work with facet wrap to split your complexity into multiple charts.