Top-ranked Teams Research Articles

Upper limb movement control strategy of top-tier elite male badminton players when smashing for different distances

PurposeMastering the skill of smashing in a game can be quite challenging. Unfortunately, it’s not always easy to control the distance of the smash. The present study investigated the impact of different smashing distances on muscle activation patterns and kinematics of the upper limbs.Material and methodsTen collegiate athletes from a top-ranked team in Korea participated in the present study with a mean age, weight, and height of 21 ± 1 years, 71.9 ± 9.1 kg, and 1.78 ± 2.24 m, respectively. Three valid data points for each subject at three different distances were analyzed using a one-way analysis of variance (ANOVA).ResultsThe results demonstrated that with increasing smashing distance, the internal rotation angle of the elbow and wrist joints decreased, while the angular velocity of flexion and internal rotation of the shoulder, elbow, and wrist joints increased. Muscle activation increased for all muscles except the biceps brachii (BICLO) with increasing distance. The greatest changes in activation levels were observed in the anterior deltoid (DELTA), extensor carpi radialis (ECR), and flexor carpi ulnaris (FCU).ConclusionsDuring tasks with different smashing distances, muscle activation in the upper limbs (except the biceps) increased with increasing distances. Additionally, the internal rotation angle of the upper limbs decreased with increased distances, while the angular velocity of flexion and internal rotation of the shoulder, elbow, and wrist joints increased.

Physiological parameters of professional football players in teams of various levels

Background and study aim. Soccer is a demanding sport which, in addition to technical and motor characteristics, also requires exceptional functional abilities of athletes to play at a high level. The aim of this study: 1) To analyze the disparities in physiological parameters among professional soccer players in teams of varying levels. 2) To explore potential connections between team ranking positions and the physiological attributes of professional soccer players. Material and Methods. 91 professional soccer players who compete in Serbian Super league (age 24.14±5.15) were included in the study. Participants were classified into two groups. The first group consisted of players from teams that were ranked from third to sixth place at the end of 2021/2022 competitive season (n=49). The second group of participants consisted of the players of the teams that were ranked in the last two places at the end of the above-mentioned season (n=42). Maximal oxygen consumption (VO2max), as well as running speeds and heart rates at first and second ventilatory thresholds (VT1 and VT2) were obtained by performing maximal cardiopulmonary exercise test. Results. The results showed a statistically significant difference in terms of VO2max values, running speeds and heart rates at VT1 and VT2 when the top and bottom ranked teams were compared (p < 0.001). Conclusion. The results of the study imply that playing high-level soccer requires exceptional functional abilities, which go beyond technical and tactical characteristics. These exceptional abilities enable players to continuously perform at high intensities during the games in order to achieve top results. Coaches should be aware of the importance of high aerobic capacity for achieving top results, especially if the team consists of less talented players.

Quantitative assessment of the HVAC system of zero-energy houses of the Solar Decathlon Middle East 2021

HVAC systems consume up to 50% of the total energy demanded by buildings. This paper aims to provide quantitative assessment of the HVAC solutions used on the highly efficient houses competing in the Solar Decathlon Middle East 2021. This international competition challenges university students to design, build, and operate sustainable zero-energy houses. The analysis includes the system selection, capacity, and coefficient of performance (COP), as well as the monitored indoor temperature, relative humidity, and CO2 levels. The university teams’ selection capacity (systems availability) and budget were affected by the COVID-19 pandemic. However, they designed their houses to respond appropriately to arid climates and reduce HVAC consumption. The study evaluates the HVAC solutions of all eight projects, providing more information about the four top-ranked teams. Most homes use air-to-air, decentralized, and multizone air-conditioners. The teams made the best effort to select systems that significantly exceed the COP required by the local regulations. Some also exceed the local energy codes regarding refrigerants’ global warming potential. The average COP (at T1 i.e., Moderate Climate Conditions) of air-to-air systems was 3.71 kW/kW, and the air-to-water system was 3.42 kW/kW. The lower installed cooling capacity per area of air-to-water HVAC systems was 57 W/m2 and 122 W/m2 in the air-to-air ones. In several cases, the HVAC systems’ consumption was affected by the short assembly period (15 days), nonprofessional student construction, and the lack of a testing period before starting the competition. Nevertheless, these houses exhibited excellent performance, and their analysis brought relevant lessons for buildings in arid climates.

Contextual factors and match-physical performance of international-level footballers with cerebral palsy

ABSTRACT This study aimed to explore the match-physical response of international-level footballers with cerebral palsy (CP) according to contextual factors such as team ranking, the quality level of opposition, and match outcome. Eighty-two male outfield footballers with CP from top-ranked teams (n = 26) and bottom-ranked teams (n = 56) took part in this study. The match-player response was recorded using global positioning devices considering the total distance, distance covered at different velocities and number of short-term actions. Players from top-ranked teams covered more distance at low intensities compared to players from bottom-ranked teams (p < 0.05; 0.45 < ES < 0.49). Higher sprint distance, moderate and high accelerations/decelerations were found in players from top-ranked teams playing against teams of similar levels (p < 0.05; 0.73g< 0.86). When players from bottom-ranked teams played against teams of equivalent rank, more physical requirements were presented in total distance, running from medium to sprint intensities, and the number of short-term actions (p < 0.0.05; −0.55S < −0.89). Players from the top-ranking teams presented greater sprint demands in losing matches (p<0.05; dg=-1.01). Considering the bottom-ranking opposition, players presented more running demands in matches with losing results in moderate running, sprinting, and moderate-to-high deceleration (p < 0.05; −0.61 < dg < −0.64). The match-physical response of players with CP varies according to the contextual factors studied. The results provide advice for coaches and classifiers to understand more comprehensively the competing demands in CP football. This knowledge could help practitioners to prepare competition matches and planning training load or post-match recovery strategies, but also for the observation assessments of the classification process in this team para-sport.

Quantifying training load and intensity in elite male ice hockey according to game-related contextual variables.

We aimed to quantify training load (TL) and intensity during practice sessions according to game-related contextual variables (game outcome, opponent standard, game location) in an elite male ice hockey team. Practice data were collected using a wearable 200-Hz accelerometer, heart rate (HR) recording, and session-rating of perceived exertion (s-RPE) throughout 23 sessions (n = 306 files). The reference team performed a greater number of accelerations, decelerations, spent longer time > 85% maximum HR (t85%HRmax) and reported greater s-RPE after losing a game compared to a win (r = 0.13-0.19). Moreover, a lower number of accelerations, decelerations, t85%HRmax and s-RPE (r = 0.15-0.45) were found before playing against a top-ranked opponent. In contrast, more accelerations, decelerations, longer t85%HRmax and greater s-RPE were observed after playing against a top-ranked team opponent (r = 0.15-0.41). The players performed more accelerations/min, spent more t85%HRmax and reported greater s-RPE before playing an away game (r = 0.13-0.22). Weekly TL seems to slightly increase after losing a game, when preparing a game against a weaker opponent, after playing against a stronger opponent, and when preparing an away game. On the other hand, training intensity seems not to be affected by game-related contextual variables. Thus, ice hockey practitioners involved with TL monitoring should consider the interplay of the numerous variables that influence the volume of prescribed training and the actual training responses in each individual player.

Who will Win the Data Science Competition? Insights from KDD Cup 2019 and Beyond

Data science competitions are becoming increasingly popular for enterprises collecting advanced innovative solutions and allowing contestants to sharpen their data science skills. Most existing studies about data science competitions have a focus on improving task-specific data science techniques, such as algorithm design and parameter tuning. However, little effort has been made to understand the data science competition itself. To this end, in this article, we shed light on the team’s competition performance, and investigate the team’s evolving performance in the crowd-sourcing competitive innovation context. Specifically, we first acquire and construct multi-sourced datasets of various data science competitions, including the KDD Cup 2019 machine learning competition and beyond. Then, we conduct an empirical analysis to identify and quantify a rich set of features that are significantly correlated with teams’ future performances. By leveraging team’s rank as a proxy, we observe “the stronger, the stronger” rule; that is, top-ranked teams tend to keep their advantages and dominate weaker teams for the rest of the competition. Our results also confirm that teams with diversified backgrounds tend to achieve better performances. After that, we formulate the team’s future rank prediction problem and propose the Multi-Task Representation Learning (MTRL) framework to model both static features and dynamic features. Extensive experimental results on four real-world data science competitions demonstrate the team’s future performance can be well predicted by using MTRL. Finally, we envision our study will not only help competition organizers to understand the competition in a better way, but also provide strategic implications to contestants, such as guiding the team formation and designing the submission strategy.

Effect of tournament format change on team performance of Chinese Football Super League during COVID-19 pandemic.

The outbreak of COVID-19 forced Chinese Football Association to adopt a special format for the 2020 Chinese Football Super League (CSL) after 5-month suspension. All 160 matches (decreased from 240) were divided into two stages (round robin; two-legged knockout) and played behind closed doors in two neutral cities (Suzhou and Dalian). The purpose of this study was to investigate the influence of absence of spectators and format change on teams’ winning percentage and match performance. Technical-tactical and event data of all 400 matches during 2019 and 2020 seasons played by 18 teams were collected. A total of 26 performance indicators were extracted and standardized from the original data. The independent t-test results showed that there were statistically significant differences in 13 performance indicators between 2019 season home win and 2020 season win: shot, shot not on target, simple pass, pass, pass accurate, key pass, corner, corner accurate, dribble and take on (p&lt;0.025, absolute ES: 0.28-0.74) for player performance as well as defensive fouls, defensive third foul and yellow cards for referee decision-making (p=0.001, absolute ES: 0.42-0.45). And there were statistically significant differences in 16 performance indicators between 2019 season home not win and 2020 season not win: shot, shot not on target, touch, simple pass, pass, pass accurate, key pass, corner, corner accurate, dribble, defensive duel, penalty shot and clearance for player performance (p&lt;0.04, absolute ES: 0.21-0.93) as well as defensive third foul, defensive foul and yellow card for referee decision-making (p&lt;0.015, absolute ES: 0.29-0.43). The findings implied that the change of match locations and absence of team supporters led to a decrease in teams’ match performance and an increase in referees’ decision. Although top-ranked teams seemed to be unaffected and the low-ranked teams showed an increase in winning percentage, the fairness of such tournament format needs to be reconsidered.

Variations in Attack Patterns between Female and Male outside hitters in top-level Volleyball

In volleyball outside hitters have wide participation in the offensive game. This study aims to analyze the variables that predict attack effectiveness in top-level volleyball depending on gender. Inferential statistics, multinomial logistic regression and multiple correspondence analysis were applied to analyze 1931 attacks (1154 for women and 777 for men) from the six top-ranked teams in the Men and Women World Championships 2018. The analysis revealed that the two genders vary in some aspects of the game at the top level. In order to increase the odds for a winning attack, male volleyball teams must pass the ball to the setter accurately, setting in 2nd fast tempo to zones 4 and 6, enriching offensive options across all the net to avoid attack against an organized block. On the other hand, female teams must set in 2nd fast tempo to the offensive zone (positions 4 &amp; 2), to avoid off-speed attacks, to spike primarily parallel and to take advantage of no block situations. A similarity between genders is that attack against a single block is more difficult and causes lower effectiveness than spike versus open double block.

Match running performance in Brazilian professional soccer players: comparisons between successful and unsuccessful teams

BackgroundThis study aimed to compare the match running performance between bottom- and top-ranked teams in professional soccer players over the 2020 season of the Brazilian National 2nd Division League. In addition, this study verified the independent and interactive effects of playing position and contextual factors on running outputs between these teams.MethodsForty-eight professional male outfield soccer players participated in this study (top-ranked team, n = 24; bottom-ranked team, n = 24). The distance- and accelerometry-based measures were recorded during 69 matches using a global positioning system (10 Hz) integrated with an accelerometer (400 Hz).ResultsThe top-ranked team covered greater total distance [median (interquartile range); 10,330.0 m (1430.0)] and high-acceleration [97.0 m (32.0)] than the bottom-ranked team, in home and away matches [p < 0.05, effect size (ES) = small]. The midfielders of the top-ranked team covered higher total distance, high-speed running (> 18 km h−1), high acceleration (≥ 3 m s−2), high-deceleration (≤ −3 m s−2), and performed more sprints [(> 25 km h−1) compared to midfielders of the bottom-ranked team (p < 0.05, η2 = small-moderate]. The matches against top-level opponents required high values of high-acceleration and number of sprints only for the top-ranked team (p < 0.05, ES = small). Independent analysis showed that match outcome (loss vs. draw vs. win) was not influenced by running performance for both bottom- and top-ranked teams (p > 0.05; η2 = small). However, the top-ranked team covered greater total distance, high-acceleration/deceleration than bottom-ranked team in loss matches (p < 0.05, η2 = small).ConclusionsThese findings should be considered when the coaches and practitioners interpret the match running outputs and when evaluating the effects of training intervention on these performance indicators.

Variations in attack patterns between female and male opposite players in top-level volleyball

In volleyball, the opposite player is the most requested hitter since she/he is the player with a higher probability of successfully carrying out attacks. The main objective of this study was to analyze variables that predicted attack effectiveness, in top-level teams depending on gender. Inferential analysis and multinomial logistic regression were applied to analyze 1512 attacks of men and women opposite players from 79 sets of the six top-ranked teams in the Men and Women 2018 World Championships. The analysis revealed that in female volleyball to increase the odds for a winning attack from the opposite player, teams have to pass the ball more accurately, setters have to set in a faster tempo and opposites avoid off-speed attack and spike in a diagonal direction from position 2. On the contrary, in male volleyball the odds for a winning attack from the opposite player are increased when male setters set accurate to the pre-agreement point on the net independent of the quality of the previous pass, opposite players hit in parallel and volleyball coaches select proper line-up and process tactical substitutions to increase the number of rotations with an opposite player in the offensive zone.

The 2017 and 2018 Iranian Brain-Computer Interface Competitions.

This article summarizes the first and second Iranian brain–computer interface competitions held in 2017 and 2018 by the National Brain Mapping Lab. Two 64-channel electroencephalography (EEG) datasets were contributed, including motor imagery as well as motor execution by three limbs. The competitors were asked to classify the type of motor imagination or execution based on EEG signals in the first competition and the type of executed motion as well as the movement onset in the second competition. Here, we provide an overview of the datasets, the tasks, the evaluation criteria, and the methods proposed by the top-ranked teams. We also report the results achieved with the submitted algorithms and discuss the organizational strategies for future campaigns.

The first world championship in cybernetic building optimization

ABSTRACTThe World Championship in Cybernetic Building Optimization (WCCBO) was held in 2019 to test participants’ ability to optimize buildings cybernetically. Office buildings with a total floor area of 10,000 m2 were built in cyberspace, one for each of the 33 participating teams. The cyber buildings were controlled by BACnet, and the participants competed to show their operational skills by tuning the HVAC system of their respective cyber buildings online. The ability of optimization was evaluated in terms of both energy consumption and thermal comfort, and their scores were published online in real-time. A total of 339 different operations were tested during the two-month competition period. The top-ranked team succeeded in reducing energy consumption and thermally dissatisfied occupant rate by 12.1% and 21.0%, respectively. In this paper, we report on the examination of the rules and schedule of this championship as well as the analysis of the participants’ scores.

REFUGEChallenge: A unified framework for evaluating automatedmethods for glaucomaassessment from fundus photographs.

Glaucoma is one of the leading causes of irreversible but preventable blindness in working age populations. Color fundus photography (CFP) is the most cost-effective imaging modality to screen for retinal disorders. However, its application to glaucoma has been limited to the computation of a few related biomarkers such as the vertical cup-to-disc ratio. Deep learning approaches, although widely applied for medical image analysis, have not been extensively used for glaucoma assessment due to the limited size of the available data sets. Furthermore, the lack of a standardize benchmark strategy makes difficult to compare existing methods in a uniform way. In order to overcome these issues we set up the Retinal Fundus Glaucoma Challenge, REFUGE (https://refuge.grand-challenge.org), held in conjunction with MICCAI 2018. The challenge consisted of two primary tasks, namely optic disc/cup segmentation and glaucoma classification. As part of REFUGE, we have publicly released a data set of 1200 fundus images with ground truth segmentations and clinical glaucoma labels, currently the largest existing one. We have also built an evaluation framework to ease and ensure fairness in the comparison of different models, encouraging the development of novel techniques in the field. 12 teams qualified and participated in the online challenge. This paper summarizes their methods and analyzes their corresponding results. In particular, we observed that two of the top-ranked teams outperformed two human experts in the glaucoma classification task. Furthermore, the segmentation results were in general consistent with the ground truth annotations, with complementary outcomes that can be further exploited by ensembling the results.

Factors Affecting the Anthropometric and Physical Characteristics of Elite Academy Rugby League Players: A Multiclub Study.

To investigate the factors affecting the anthropometric and physical characteristics of elite academy rugby league players. One hundred ninety-seven elite academy rugby league players (age = 17.3 [1.0]y) from 5 Super League clubs completed measures of anthropometric and physical characteristics during a competitive season. The interaction between and influence of contextual factors on characteristics was assessed using linear mixed modeling. All physical characteristics improved during preseason and continued to improve until midseason, whereafter 10-m sprint (η2 = .20 cf .25), countermovement jump (CMJ) (η2 = .28 cf .30), and prone Yo-Yo Intermittent Recovery (Yo-Yo IR) test (η2 = .22 cf .54) performance declined. Second (η2 = .17) and third (η2 = .16) -year players were heavier than first-years, whereas third-years had slower 10-m sprint times (η2 = .22). Large positional variability was observed for body mass, 20-m sprint time, medicine-ball throw, CMJ, and prone Yo-Yo IR1. Compared with bottom-ranked teams, top-ranked teams demonstrated superior 20-m (η2 = -.22) and prone Yo-Yo IR1 (η2 = .26) performance, whereas middle-ranked teams reported higher CMJ height (η2 = .26) and prone Yo-Yo IR1 distance (η2 = .20) but slower 20-m sprint times (η2 = .20). These findings offer practitioners who design training programs for academy rugby league players insight into the relationships between anthropometric and physical characteristics and how they are influenced by playing year, league ranking, position, and season phase.

Benchmark on Automatic 6-month-old Infant Brain Segmentation Algorithms: The iSeg-2017 Challenge.

Accurate segmentation of infant brain magnetic resonance (MR) images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) is an indispensable foundation for early studying of brain growth patterns and morphological changes in neurodevelopmental disorders. Nevertheless, in the isointense phase (approximately 6-9 months of age), due to inherent myelination and maturation process, WM and GM exhibit similar levels of intensity in both T1-weighted (T1w) and T2-weighted (T2w) MR images, making tissue segmentation very challenging. Despite many efforts were devoted to brain segmentation, only few studies have focused on the segmentation of 6-month infant brain images. With the idea of boosting methodological development in the community, iSeg-2017 challenge (http://iseg2017.web.unc.edu) provides a set of 6-month infant subjects with manual labels for training and testing the participating methods. Among the 21 automatic segmentation methods participating in iSeg-2017, we review the 8 top-ranked teams, in terms of Dice ratio, modified Hausdorff distance and average surface distance, and introduce their pipelines, implementations, as well as source codes. We further discuss limitations and possible future directions. We hope the dataset in iSeg-2017 and this review article could provide insights into methodological development for the community.

Characterising game styles in the English Premier League using the “moments of play” framework

ABSTRACTThe aim of the current study was to identify different playing styles within 2015–16 English Premier League season. Game-related variables were allocated to the “moments of play” framework, and multidimensional game styles were investigated using cluster analysis and related to finishing position. Totally, 96 game-related variables were allocated into one of five discrete “moments of play”; set pieces (SP), established offence (EO), transition to offence (TO), transition to defence (TD) and established defence (ED). Moments of play variables were z-score transformed. The game-specific mean z-score for the five moments was used as K-means cluster analysis inputs. Relationships between game style clusters and finishing position were investigated using chi-square analysis. Three game style clusters were identified: (1) moderately favouring ED and prevalent among lower-ranked teams (n = 270 games); (2) dominant in TO and TD (n = 241) and represented by the top team, (3) strong in EO and SP and characterised by higher ranked teams (games n = 249). Success for the top-ranked team was associated with the dominance in transition moments. Higher ranked teams demonstrate control of EO and SP, whereas lower ranked teams clustered for poor performance in all moments except ED. Game style varies across teams, however, it also aggregates relative to finishing position.

Does the FIFA World Cup’s Congested Fixture Program Affect Players’ Performance?

Background: The tournament design of any FIFA World CupTM is characterized with a high amount of games within a short period of time with limited recovery periods. This study aimed to investigate alterations in match activity and fatigue patterns of players from top ranked teams during the 2014 FIFA World Cup in Brazil. Methods: The performance data from the final four ranked teams were recorded from the official game statistics of the federation internationale de football association. Inclusion criteria were: a) played during the seven matches, and b) took part in at least 85% of the total playing time of its respective team. Results: Seventeen players fulfilled the criteria (eight defenders, five midfielders and four attackers). Average exposure of the selected players was 97.4% ± 4.1% (range 86% - 100%) of the tournament regular official playing time. Total distance covered was lower (P < 0.05) in the fourth (round of 16) and fifth (quarterfinals) game than in the first and the sixth (semi-finals). Moreover, high-intensity running was lower in game five than one (P < 0.05), six (P < 0.01) and seven (finals and 3th - 4th place; P < 0.05). Additionally, players performed a lower number of sprints in game five than in game seven (P < 0.05). An intensity composite score revealed that game five was less intense (P < 0.05) than game one, six and seven. Additionally, substantially small decrement (Effect size = 0.4) in the number of sprints between post-peak match (subsequent game to peak-match) and average-five matches was observed. Conclusions: The analyses suggest that there wasn’t a general decline in player’s match activity during the 2014 FIFA World-cup in Brazil in a congested schedule.

Human blood gene signature as a marker for smoking exposure: Computational approaches of the top ranked teams in the sbv IMPROVER Systems Toxicology challenge

Crowdsourcing has emerged as a framework to address methodological challenges in omics data analysis and assess the extent to which omics data are predictive of phenotypes of interest. The sbv IMPROVER Systems Toxicology challenge was designed to leverage crowdsourcing to determine whether human blood gene expression levels are informative of current and past smoking. Participating teams were invited to use a training gene expression dataset to derive parsimonious models (up to 40 genes) that can accurately classify subjects into exposure groups: smokers, former smokers that quit for at least one year, and never-smokers. Teams were ranked based on two classification performance metrics evaluated on a blinded test dataset. The analytical approaches of the first- and third-ranked teams, that are presented in detail in this article, involved feature selection by moderated t-test or LASSO regression and linear discriminant analysis (LDA) and logistic regression classifiers, respectively. While the 12-gene signature of the top team allowed the classification of current smokers with 100% sensitivity at 93% specificity, discriminating former smokers from never-smokers was much more challenging (65% sensitivity at 57% specificity). Gene ontology molecular functions and KEGG pathways associated with current smoking included G protein-coupled receptor activity, signaling receptor activity, calcium ion binding, and the Neuroactive ligand-receptor interaction pathway. Selection of marker genes by either moderated t-test or multivariate LASSO regression followed by LDA or logistic regression, are robust approaches to classification with omics data, confirming in part findings of previous sbv IMPROVER challenges. While current smoking is accurately identified based on blood mRNA levels, smoking cessation for more than one year is accompanied by a “normalization” of the expression of certain mRNAs, making it difficult to distinguish former smokers from never-smokers.

Species translatable blood gene signature as a marker of exposure to smoking: Computational approaches of the top ranked teams in the sbv IMPROVER Systems Toxicology Challenge

Crowdsourcing has been used to address computational challenges in systems biology and assess translation of findings across species. Sub-challenge 2 of the sbv IMPROVER Systems Toxicology Challenge was designed to determine whether a common set of genes can be used to identify exposure to cigarette smoke in both human and mouse. Participating teams used a training set of human and mouse blood gene expression data to derive parsimonious models (up to 40 genes) that classify subjects into exposure groups: smokers, former smokers, and never-smokers. Teams were ranked based on two classification performance metrics evaluated on a blinded test dataset. Prediction of current exposure to cigarette smoke in human and mouse by a common prediction model was achieved by the top ranked team (Team 219) with 89% balanced accuracy (BAC), while past exposure was predicted with only 57% BAC. The prediction model of the top ranked team was a random forest classifier trained on sets of genes that appeared best for each species separately with no overlap between species. By contrast, Team 264, ranked second (tied with Team 250), selected genes that were simultaneously predictive in both species and achieved 80% and 59% BAC when predicting current and past exposure, respectively. These performance values were lower than the 96.5% and 61% BAC estimates for current and past exposure, respectively, obtained by Team 264 (top ranked in sub-challenge 1) when using only human data. Unlike past exposure, current exposure to cigarette smoke can be accurately assessed in both human and mouse with a common prediction model based on blood mRNAs. However, requiring a common gene signature to be predictive in both species resulted in a substantial decrease in balanced accuracy for prediction of current exposure to cigarette smoke (from 96.5% to 80%), suggesting species-specific responses exist.

Surveying Professional Forecasters

This case illustrates how averaging point forecasts harnesses the wisdom of crowds. Students access data from the Survey of Professional Forecasters (SPF) and compare the performance of the crowd (i.e., the average point forecasts) to the average performance of the individual panelists and the best performer from the previous period.The case is intended for use in a class on forecasting, and the instructor can present it in three ways: with all necessary SPF data cleaned and preprocessed in a student spreadsheet (UVA-QA-0805X, provided with the case); with code (also provided in the student spreadsheet) written by the case authors in R, the statistical computing package, as well as a supplementary handout (UVA-QA-0805H, also provided with the case), which walks students through R code, explaining how to clean and analyze the SPF data; or as a team project to be worked on over several days, providing neither the spreadsheet nor the supplement. Excerpt UVA-QA-0805 Rev. Apr. 7, 2014 SURVEYING PROFESSIONAL FORECASTERS Since 1981, the Wall Street Journal surveyed economists for point forecasts of economic indicators such as gross domestic product (GDP), inflation, and unemployment. These forecasts were scored and ranked annually based on their accuracy. The top-ranked forecasters were celebrated in a special announcement of the complete ranking on the Journal's website, typically followed by a press release by the top forecasters' employers. In 2012, the Journal decided for the first time to average all the forecasts for each indicator and present that set of forecasts as an additional panelist. The economists' average forecast, or “the crowd,” was then ranked alongside the actual panelists. How had the economists' average forecast performed? How did the crowd compare to “chasing the expert”? Among the 49 panelists, it ranked 12th in 2012. Yet Don Leavens and Tim Gill, the top-ranked team in 2011, came in 5th in 2012 (see Exhibits 1 and 2). What did this say about the wisdom of the crowd? . . .