18 |
19 | 
37 | 
39 | 
41 | 
43 |
46 |
54 |
57 |
58 |
68 |
--------------------------------------------------------------------------------
/style.css:
--------------------------------------------------------------------------------
1 | .caption {
2 | padding: 8% 0 5px 0;
3 | font-size: 26px;
4 | color: #636161;
5 | font-family: "Verdana", sans-serif;
6 | }
7 |
8 | #back_button {
9 | background-color:#d9d5cc;
10 | display:inline-block;
11 | cursor:pointer;
12 | color:#ffffff;
13 | font-family:Arial;
14 | font-size:13px;
15 | font-weight:bold;
16 | padding:18px 30px;
17 | text-decoration:none;
18 | position: absolute;
19 | /*margin: 40px 0 0 40px;*/
20 | }
21 | #back_button:hover {
22 | background-color:#aba79d;
23 | }
24 |
25 | .arrow_left {
26 | border: solid white;
27 | border-width: 0 5px 5px 0;
28 | display: inline-block;
29 | padding: 5px;
30 | transform: rotate(135deg);
31 | -webkit-transform: rotate(135deg);
32 | }
33 |
34 |
35 | .error {
36 | color: #a31515;
37 | }
38 |
39 | .center_img {
40 | display: block;
41 | margin-left: auto;
42 | margin-right: auto;
43 | width: 60%;
44 | }
45 | .table_img {
46 | width: 100%;
47 | }
48 |
49 | p, .content {
50 | font-size: 15px;
51 | color: #7c7c7c;
52 | font-family: "Verdana", sans-serif;
53 | }
54 |
55 | .content {
56 | padding: 0 40px;
57 | color: #212121;
58 | }
59 |
60 | #coffee_btn {
61 | color: #0a677a;
62 | cursor: pointer;
63 | }
64 |
65 | #bottom_text {
66 | font-family: "Verdana", sans-serif;
67 | font-size: 17px;
68 | width: 100%;
69 | text-align: center;
70 | position: relative;
71 | top: 18%;
72 | }
73 |
74 | #donate_links {
75 | display: none;
76 | }
77 |
78 | a {
79 | color: #0a677a;
80 | text-decoration: none;
81 | }
82 |
83 | .container, .about_container {
84 | height: 90%;
85 | width: 70%;
86 | margin: auto;
87 | box-shadow: 0px 0px 20px 1px rgba(0,0,0,0.3);
88 | background-color: #ffffff;
89 | padding: 0 0 30px 0;
90 | }
91 |
92 | .about_container {
93 | height: auto;
94 | }
95 |
96 | .input_container, .center {
97 | text-align: center;
98 | }
99 |
100 | input {
101 | font-size: 19px;
102 | text-align: center;
103 | border-width: 0px;
104 | border: none;
105 | padding: 10px;
106 | margin: 10px 0 20px 0;
107 | }
108 |
109 | .username:focus {
110 | outline: none;
111 | box-shadow: 0px 0px 20px 1px #c4c4c4;
112 | }
113 |
114 | body {
115 | background-color: #dbdad9;
116 | }
117 |
118 | /* progress circle from https://www.codingnepalweb.com/2020/07/circular-progress-bar-using-html-css.html */
119 |
120 | .circular {
121 | height: 200px;
122 | width: 200px;
123 | position: relative;
124 | top: 5%;
125 | margin: auto;
126 | }
127 |
128 | .circular .inner, .circular .outer, .circular .circle {
129 | position: absolute;
130 | z-index: 6;
131 | height: 100%;
132 | width: 100%;
133 | border-radius: 100%;
134 | box-shadow: inset 0 0 3px 2px rgba(0,0,0,0.3);
135 | }
136 | .circular .inner {
137 | z-index:8;
138 | top: 50%;
139 | left: 50%;
140 | height: 180px;
141 | width: 180px;
142 | margin: -90px 0 0 -90px;
143 | background-color: #ffffff;
144 | border-radius: 100%;
145 | box-shadow: 0 0 3px 2px rgba(0,0,0,0.4);
146 | }
147 |
148 | .circular .inner:hover, #FIDE_rating_container:hover {
149 | cursor: pointer;
150 | background-color: #f0f0f0;
151 | }
152 |
153 | .circular .circle {
154 | z-index: 1;
155 | box-shadow: none;
156 | }
157 |
158 | .circular #FIDE_rating_container, .circular #calculate_container {
159 | position: absolute;
160 | top: 50%;
161 | left: 50%;
162 | transform: translate(-50%, -50%);
163 | z-index: 10;
164 | font-size: 25px;
165 | color: 7c7c7c;
166 | font-family: "Verdana", sans-serif
167 | }
168 |
169 | .circular #calculate_container {
170 | font-size: 18px;
171 | }
172 |
173 | .circular .bar {
174 | position: absolute;
175 | height: 100%;
176 | width: 100%;
177 | background: #dbdad9;
178 | -webkit-border-radius: 100%;
179 | clip: rect(0px, 200px, 200px, 100px);
180 | }
181 | .circle .bar .progress {
182 | position: absolute;
183 | height: 100%;
184 | width: 100%;
185 | -webkit-border-radius: 100%;
186 | clip: rect(0px, 100px, 200px, 0px);
187 | }
188 | .circle .bar .progress {
189 | background: #2cdb00;
190 | }
191 |
192 | .circle .left .progress {
193 | z-index: 1;
194 | animation: left 0.5s linear both;
195 | }
196 | @keyframes left {
197 | 100%{
198 | transform: rotate(360deg);
199 | }
200 | }
201 | .circle .right {
202 | z-index: 3;
203 | transform: rotate(180deg);
204 | }
205 | .circle .right .progress {
206 | animation: right 0.5s linear both;
207 | }
208 | @keyframes right {
209 | 100%{
210 | transform: rotate(-360deg);
211 | }
212 | }
213 |
214 | @media screen and (max-width: 650px) {
215 | .container, .about_container {
216 | height: 95%;
217 | width: 100%;
218 | margin: none;
219 | box-shadow: none;
220 | }
221 | .about_container{
222 | height: auto;
223 | }
224 | .circular {
225 | position: relative;
226 | height: 150px;
227 | width: 150px;
228 | margin: auto;
229 | top: 20px;
230 | left: 0%;
231 | transform: translate(0%, 0%);
232 | }
233 | .circular .inner {
234 | height: 135px;
235 | width: 135px;
236 | margin: -67.5px 0 0 -67.5px;
237 | }
238 | .circular .bar{
239 | clip: rect(0px, 150px, 150px, 75px);
240 | }
241 | .circle .bar .progress {
242 | clip: rect(0px, 75px, 150px, 0px);
243 | }
244 | .caption {
245 | padding: 8% 0 5px 0; /*change or delete*/
246 | font-size: 18px;
247 | color: #636161;
248 | font-family: "Verdana", sans-serif;
249 | }
250 | body {
251 | background-color: #ffffff;
252 | }
253 | #bottom_text {
254 | top: 20%;
255 | }
256 | }
--------------------------------------------------------------------------------
TLDR
20 |This site uses your ratings in each time control category that you have at least 20 games as inputs to linear regression models in order to estimate a FIDE rating. The models were generated from a data set of about 28,000 users who have listed their FIDE rating on their lichess profile.
21 |Also, if you liked this site and would like to see it get more accurate I would be eternally grateful if you considered donating to the project.
22 |
23 |
34 | -
24 |
- 25 | PayPal: https://paypal.me/UnrealDuck 26 |
- 27 | Bitcoin: 1Gy8GQhvDBka6Sx1DwFPo1bh9Bs6gbZEEB 28 |
-
29 | Venmo: @ethleb
30 |
31 |
32 |
About the Data
35 |I started by downloading the public information of a little over a million lichess users (this took a while) and then created a dataset with only the users who listed a FIDE rating in their profile. Only about 3% of users included a FIDE rating in their profile, so the dataset came out to around 28,000 observations. Before fitting a linear model I removed every data point that had fewer than 50 games played in the time controls that I was using for the explanatory variables. Below is a scatterplot of FIDE rating plotted against blitz rating.
36 |
37 | Because the FIDE ratings are self reported it’s a little messy. There is clearly a relationship, but there’s a lot of cleanup that has to be done. First, I removed every data point with a FIDE rating above 2900 (the record is 2882) and below 1000 (because that’s the floor in the FIDE rating system). The plot is already looking better:
38 |
39 | There’s still a lot of noise in the top left, which makes sense because I would expect people to pretend to be stronger players than they are and not weaker than they are. To clean it some more I fit a linear model to the data and removed every point with a residual greater than two standard deviations from the median (chop at red lines):
40 |
41 | The result looks like this:
42 |
43 | Beautiful. If I were doing this data analysis for interpretive reasons that last step would likely have been me overstepping as an analyst, but all I’m trying to do is create a model with strong predictive power. The final step is to generate a least absolute distance model using r’s rq() function and call it a day.
44 |Unfortunately blitz rating isn’t the only explanatory variable I’m working with. I also have bullet, rapid and classical ratings, and every combination of the four (for a total of 15 models). For the other three single regression models I repeated the steps above. All four resulting scatter plots follow: 45 |
 |
48 | |
49 |
 |
51 |  |
52 |
The only thing I did differently for the multiple regression models (the models with more than one explanatory variable) was use the r’s step() function to remove redundant variables each time I fit a model. Unfortunately I can’t show the results in a scatterplot, because scatterplots only have two dimensions. The models have a Pearson correlation coefficient of 0.626 at best to 0.538 at worst.
55 |If you have any questions, comments, or words of encouragement feel free to shoot an email to duckunreal@gmail.com
56 |