{"id":76063,"date":"2023-08-28T13:44:00","date_gmt":"2023-08-28T17:44:00","guid":{"rendered":"https:\/\/www.sailingworld.com\/?p=76063"},"modified":"2023-08-28T16:01:06","modified_gmt":"2023-08-28T20:01:06","slug":"a-data-driven-race-winning-formula","status":"publish","type":"post","link":"https:\/\/www.sailingworld.com\/how-to\/a-data-driven-race-winning-formula\/","title":{"rendered":"A Data-Driven Race Winning Formula"},"content":{"rendered":"\n        <section class=\"hydra-container\">\n\n\t\t\t                <div class=\"hydra-canvas\">\n\t\t\t\t\t<img fetchpriority=\"high\" decoding=\"async\" width=\"800\" height=\"534\" src=\"https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/230713_P24TE_Marseille_L2_2145-1024x683.jpg\" class=\"hydra-image disable-lazyload\" alt=\"Paris 2024 Olympic Sailing Test Event, Marseille, France.\" sizes=\"(max-width: 1400px) 100vw, 1400px\" fetchpriority=\"high\" srcset=\"https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/230713_P24TE_Marseille_L2_2145-1024x683.jpg 1024w, https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/230713_P24TE_Marseille_L2_2145-300x200.jpg 300w, https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/230713_P24TE_Marseille_L2_2145-768x512.jpg 768w, https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/230713_P24TE_Marseille_L2_2145-50x33.jpg 50w, https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/230713_P24TE_Marseille_L2_2145.jpg 1200w\" \/>                <\/div>\n\t\t\t\n\t\t\t\n            <figcaption class=\"caption margin_top_xs full border_1 hydra-figcaption\">\n                <span class=\"hydra-image-caption\">The author&#8217;s data bears out one known element of a race win: a good start is essential, followed in priority by a good first leg.<\/span>\n                <span class=\"article_image_credit italic margin_right_xs\">Sailing Energy\/World Sailing<\/span>\n\n\t\t\t\t            <\/figcaption>\n        <\/section>\n\t\t\n\n\n<p>Over six months of thesis research and statistical analyses with professors at Yale and Harvard and the head of sports analytics at ESPN, I derived a mathematical model that predicts sailing race performance. This formula reveals which skills, such as turn rate in tacks or a boat\u2019s VMG upwind, actually matter and to what degree they matter toward winning a sailing race.<\/p>\n\n\n\n<p>We know this formula can tell us how to win because the numbers don\u2019t lie. To understand the model, we first need to define the most fundamental component to winning a sailing race: crossing the finish line before your competitors. To do so, we need to have the shortest elapsed race time. With this in mind, we can produce a model to reveal which skills are the strongest predictors of a shorter race time. In other words, we can use this model to tell us which sailing skills help win races.<\/p>\n\n\n\n<p>To guide you through this model let\u2019s use the 49erFX during the last Olympic Quad. To get enough data to be able to make any discernible conclusions we take SAP boat tracking data from every 49erFX race sailed in every major regatta between August 2016 and August 2021. To decide which skills are necessary to win, we break down a sailing race into 38 individual skills that contribute to a boat\u2019s performance. These are a combination of speed, VMG, time, distance, and maneuver-specific metrics, each in total and broken down for the individual legs of the race and<img decoding=\"async\" width=\"403\" height=\"162\" src=\"\" alt=\"Image\"> the start. To better grasp the data behind the model let\u2019s look at a snippet from the final data set below.<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table><tbody><tr><td><strong>Class<\/strong><\/td><td><strong>49erFX<\/strong><\/td><td><strong>49erFX<\/strong><\/td><td><strong>49erFX<\/strong><\/td><\/tr><tr><td>boat<\/td><td>USA50<\/td><td>ESP23<\/td><td>CAN25<\/td><\/tr><tr><td>regatta<\/td><td>Europeans 2019<\/td><td>Worlds 2020<\/td><td>Olympics 2021<\/td><\/tr><tr><td>race<\/td><td>1<\/td><td>6<\/td><td>3<\/td><\/tr><tr><td>race.finish<\/td><td>15<\/td><td>3<\/td><td>7<\/td><\/tr><tr><td>regatta.finish<\/td><td>36<\/td><td>1<\/td><td>18<\/td><\/tr><tr><td>uw_dist<\/td><td>4614<\/td><td>5469<\/td><td>5196<\/td><\/tr><tr><td>distance.m.l1<\/td><td>2374<\/td><td>2937<\/td><td>1992<\/td><\/tr><tr><td>vmg.l2<\/td><td>12.59<\/td><td>13.59<\/td><td>8.56<\/td><\/tr><tr><td>jibe.speed.loss.gain.kts<\/td><td>2.43<\/td><td>1.96<\/td><td>1.54<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p>Now let\u2019s dive into the stats. To get our race win model we can run a multiple linear regression (a statistical technique that uses multiple variables to predict the outcome) to see which skills are the strongest predictors of a faster race. We use total race time as our dependent\/outcome variable and race skills as our independent\/explanatory variables. In layman\u2019s terms, we look at all the race skills to see which combination of individual skills produces a shorter race time. The model<img decoding=\"async\" width=\"603\" height=\"383\" src=\"\" alt=\"Image\"> can be illustrated through the graph below.<\/p>\n\n\n\n        <section class=\"hydra-container\">\n\n\t\t\t                <div class=\"hydra-canvas\">\n\t\t\t\t\t<img decoding=\"async\" width=\"800\" height=\"524\" src=\"https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/table-for-B-lindsey-image-1024x671.jpg\" class=\"hydra-image\" alt=\"Graph showing metric vs correlation in racing\" sizes=\"(max-width: 1400px) 100vw, 1400px\" srcset=\"https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/table-for-B-lindsey-image-1024x671.jpg 1024w, https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/table-for-B-lindsey-image-300x197.jpg 300w, https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/table-for-B-lindsey-image-768x503.jpg 768w, https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/table-for-B-lindsey-image-50x33.jpg 50w, https:\/\/www.sailingworld.com\/wp-content\/uploads\/2023\/08\/table-for-B-lindsey-image.jpg 1200w\" \/>                <\/div>\n\t\t\t\n\t\t\t\n            <figcaption class=\"caption margin_top_xs full border_1 hydra-figcaption\">\n                <span class=\"hydra-image-caption\">Metric Correlation with Race Time.<\/span>\n                <span class=\"article_image_credit italic margin_right_xs\">B Lindsay<\/span>\n\n\t\t\t\t            <\/figcaption>\n        <\/section>\n\t\t\n\n\n<p>Here, each bar represents a different skill. The length of the bar represents numerically how much a skill is correlated with total race time. The color of the bar illustrates the strength of the correlation; lighter blue indicates a strong, positive correlation and darker blue indicates a strong, negative correlation. A long, light blue bar means a skill has a strong positive correlation with total race time. This tells us as that value increases, total race time increases and you theoretically sail a worse race.<\/p>\n\n\n\n<p>While the graph is nicer to look at, the actual model shown below is more important to us: <strong>total.time.min = 34.58 \u2014 2.0vmg.l1 \u2014 .75dw.speed.kts \u2014 .66vmg.l3 \u2014 .39maneuvers.l4&nbsp;&nbsp;+ .35vmg.l2 \u2014 .29vmg.l4 \u2014 .22maneuvers.l1 \u2014 .17tack.turn.rate.deg.s&nbsp;&nbsp;+ .094total.maneuvers + .060distance.behind.line.bl + .0059dw_dist \u2014 .00065uw_dist<\/strong><\/p>\n\n\n\n<p>On the left of our mathematical formula we see <strong>total.time.min<\/strong> or the total race time. We are trying to reduce this value because less time equals a better race. On the right, we first see the number is 34.58. This is our intercept or the mean for the outcome when all of the explanatory variables are zero. Meaning, if all race skills were somehow zero, the average time it would take a 49erFX to finish a race is 34 minutes and 58 seconds. To say that race skills won\u2019t impact race time is a ridiculous concept so we don\u2019t need to pay too close attention to that number. What we do need to pay attention to is the value of the coefficient, the number in front of each skill, and the sign of this number, positive or negative.<\/p>\n\n\n\n<p>The first skill to look at is <strong>vmg.l1<\/strong> or VMG on leg 1, the first upwind. <strong>Vmg.l1<\/strong> has a value of -2.0 which is the largest magnitude and almost ten times that of any other skill in the model. Its negative value tells us that as VMG on leg 1 increases, total race time decreases. What does all this mean? That VMG on leg 1 is by far the most important contributor to the total time spent on a race; having more VMG on the first upwind beat is worth more, ten times more, to a winning boat than any other skill on the race course.<\/p>\n\n\n\n<p>Next, factor in <strong>distance.behind.line.b<\/strong> the distance behind the line at the starting signal. Its value of 0.060 tells us that the more distance you are behind the line at the start, the longer your race will be. Combine this with our conclusion about VMG on leg 1 and we see that having a good start and an efficient first beat is the most important combination to sailing a shorter race. This is something you probably knew, but the numbers confirm it is the most important factor to sailing a winning race.<\/p>\n\n\n\n<p>Now let\u2019s focus on maneuvers. First, glance at <strong>total.maneuvers<\/strong>, the total number of maneuvers. The coefficient is .094. This positive value tells us that more maneuvers equals a longer race. Again, this is pretty intuitive and you could guess that without all of the complicated calculations. However, this isn\u2019t the part to focus on. Look more closely at the other two maneuver metrics in our model: <strong>maneuvers.l1<\/strong> and <strong>maneuvers.l4<\/strong>\u2014maneuvers on leg 1 and leg 4. These skills both have negative values, -.22 and -.39, respectively. Increasing either of these values will subtract a larger number from the average race time, 34.58, which decreases the total race time. What this means is doing more maneuvers on the first upwind or the final downwind decreases your total race time\u2014sailing a theoretically better race.<\/p>\n\n\n\n<p>At first thought, these negative values seem to contradict our positive <strong>total.maneuver<\/strong>s value that says doing more maneuvers means a longer and worse race. However, if you think about these two values in the context of a race, they actually make a lot of sense.<\/p>\n\n\n\n<p>The first and last legs are where boats are typically the most clustered together; either right after the start or closing into the finish line. With boats closer together, these legs give more opportunity to gain or lose boats. It\u2019s where you might double tack to leebow a close boat or do an extra jibe onto another boat\u2019s wind to slow them. If we go back to our model, we see it\u2019s telling us that doing these extra maneuvers on leg 1 or leg 4 can actually improve our race\u2014and that does make sense. So, the next time you\u2019re racing don\u2019t be afraid to throw in that extra tack on the first upwind or jibe on the last downwind; the math tells us these can actually help win the race.<\/p>\n\n\n\n<p>Now, if the numbers lost you somewhere along the way, just remember the following:<\/p>\n\n\n\n<p>How do we win sailing races?\u00a0<br>Sail less time<\/p>\n\n\n\n<p>How do we do that?\u00a0<br>Better use and execute your maneuvers<\/p>\n\n\n\n<p>Do fewer maneuvers over the race<\/p>\n\n\n\n<p>Don\u2019t be afraid to do an extra maneuver for strategic reasons on the first upwind or last downwind<\/p>\n\n\n\n<p>Get a good start and sail a fast first beat<\/p>\n\n\n\n<p>Be on the start line at \u201cgo\u201d<\/p>\n\n\n\n<p>Have better VMG on the first upwind<\/p>\n\n\n\n<p>That\u2019s it. If you follow the numbers, winning a sailing race really is that simple.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>What if I could tell you exactly how to win your next race without ever having to watch you sail? Well, I can, and it\u2019s easier than you think.<\/p>\n","protected":false},"author":24,"featured_media":76065,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"BS_author_type":"BS_author_is_guest","BS_guest_author_name":"B Lindsay","BS_guest_author_url":"","hydra_display_date":"","hydra_display_updated":false,"_yoast_wpseo_primary_category":"156","_yoast_wpseo_metadesc":"A professional sailor and data analyst with the US Sailing Team extrapolates the winning variables of a sailboat race.","_yoast_wpseo_title":"","_yoast_wpseo_meta-robots-noindex":"","arc_story_id":"","arc_website_url":"","custom_permalink":"","arc_subtype":"","arc_exclude_from_feeds":false,"sponsored":false,"sponsored_label":"Sponsored Content","sponsored_display_label":false,"sponsored_image":false,"post_right_rail":true,"post_right_rail_ad_1":true,"post_right_rail_ad_2":true,"post_right_rail_ad_3":false,"post_right_rail_ad_4":false,"post_right_rail_recirc":true,"fixed_anchor_ad":true,"post_top_ad":true,"post_off_ramp":true,"post_taboola":false,"labels":true,"apple_news_api_created_at":"","apple_news_api_id":"","apple_news_api_modified_at":"","apple_news_api_revision":"","apple_news_api_share_url":"","apple_news_coverimage":0,"apple_news_coverimage_caption":"","apple_news_is_hidden":false,"apple_news_is_paid":false,"apple_news_is_preview":false,"apple_news_is_sponsored":false,"apple_news_maturity_rating":"","apple_news_metadata":"\"\"","apple_news_pullquote":"","apple_news_pullquote_position":"","apple_news_slug":"","apple_news_sections":"\"\"","apple_news_suppress_video_url":false,"apple_news_use_image_component":false,"footnotes":"","ad_targeting":"","ad_settings_ads_on_this_page":true,"ad_settings_automatic_ad_injection_into_the_content":true,"sponsored_url":"","social_share":true},"categories":[156],"tags":[183],"class_list":["post-76063","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-how-to","tag-sailing-how-to"],"acf":[],"apple_news_notices":[],"_links":{"self":[{"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/posts\/76063","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/users\/24"}],"replies":[{"embeddable":true,"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/comments?post=76063"}],"version-history":[{"count":0,"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/posts\/76063\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/media\/76065"}],"wp:attachment":[{"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/media?parent=76063"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/categories?post=76063"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.sailingworld.com\/wp-json\/wp\/v2\/tags?post=76063"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}