The 2024 Chessable Research Awards had five student winners, including Alex Knopps, the author of this guest blog post. Knopps explores whether solving chess puzzles alone or with a partner leads to better outcomes. His research also accounted for the difficulty of puzzles. The results indicate that there wasn’t much difference between the number of correctly solved chess puzzles in both individual and collaborative settings. However, working in a group led to fewer errors. Check out Alex Knopps’ post:

Check-mate? Exploring how collaboration impacts strategy and success in chess by Alex Knopps

Introduction 

Throughout day-to-day life, individuals are asked to solve problems and think creatively—oftentimes, this occurs in the context of others. Intuitively, one might assume that collaboration helps problem-solving and creative processes. In actuality, research indicates that collaboration can impair these processes (Basden et al., 1997; Kohn & Smith, 2011; Knopps & Wissman, under revision). However, it is an open question as to whether working with another player will facilitate or hinder performance when playing the game of chess.

Background

One increasingly studied area of cognitive psychology research is evidenced-based learning strategies, as understanding how and under what circumstances people learn best is an important part of everyday life. One such method is collaborative learning, which involves individuals learning and working together in groups. Research from this area has focused on investigating the costs and benefits that emerge from collaboration (for a review, see Marion & Thorley, 2016). One robust outcome from collaborative learning research has shown a cost to performance, such that individuals working together in a group recall less information compared to individuals working alone (referred to as collaborative inhibition: Basden et al., 1997; Blumen & Rajaram, 2008). 

Interestingly, no research has explored how collaboration impacts solving chess puzzles. It is an open question as to whether working with another player to capture a target chess piece will facilitate or hinder performance. To solve a chess puzzle, a player must understand how each piece moves on a board and make a move that captures their target piece while simultaneously ensuring their opponent does not capture a desirable piece. In addition, each player must be able to think through multiple scenarios at once and make a decision on the best possible move. In other words, chess players must strategize to be successful. On the one hand, one might argue that working with another player would be beneficial, in that players’ strategies may complement and inspire one another. On the other hand, one might argue that working with another player would be detrimental, in that players’ strategies may be disruptive to one another.

From a research perspective, outcomes from this project will provide novel and foundational evidence for how collaboration impacts performance when playing chess collaboratively versus individually. From an applied perspective, outcomes from this project will provide insights on how strategizing and successful game play is affected when working with another player, which in turn affords the possibility of discovering ways to increase problem solving and creativity for chess players. 

Hypotheses

We designed the current study to investigate the impact of working collaboratively or individually on chess problem-solving. We proposed the following hypotheses:

• Hypothesis #1: Participants working collaboratively compared to individually would make more mistakes when solving easy chess puzzles 
• Hypothesis #2:  Participants working collaboratively compared to individually would make fewer mistakes when solving difficult chess puzzles   

Methods 

Participants and Design

A 2 (condition: collaborative versus individual) x 2 (puzzle: hard versus easy) design was implemented, with condition manipulated between-subjects and puzzle manipulated within-subjects. Participants were asked to work individually or collaboratively (as a dyad of two players) to solve both easy and hard puzzles. Participants included undergraduate students from a large, midwestern University who participated for course credit. The current data set contains 18 participants (72% male; average age: 19 years); 8 individual participants and 10 collaborative individuals (making a total of 5 groups).

Materials

Materials consisted of 20 chess problem-solving puzzles. Ten puzzles were categorized as easy, and 10 puzzles were categorized as hard determined by their rating on Lichess.org. This determination is based on the number of players that have attempted to solve the puzzle and how many mistakes players make while attempting to solve the puzzle. Chess puzzles presented participants with a series of white and black pieces in predetermined locations on the board with the goal of finding the best possible move. Each chess puzzle required 1-3 moves to be successfully solved. On average, the easy puzzles took 1.4 moves, and on average the hard puzzles took 2.4 moves to solve. To solve each puzzle, participants must select a piece they would like to move and select a desired location. If it is correct, participants are prompted to either keep going to find the next best move in the puzzle indicating they are not yet done with the puzzle, or they will see a “success” message indicating they have completed the puzzle. If it is incorrect, participants are prompted to try something else and the puzzle resets to the last move. Lichess puzzles are released under the Creative Commons CC0 license. This allows the use of them for research, commercial purpose, and/or publication purposes. 

Procedure

Prior to signing up for the experiment, individuals were instructed to participate only if they understood game rules and had chess playing experience. When participants arrived at the lab, they were given a consent form that provided information about their rights as participants and the research assistant on duty asked if they would like to continue with the experiment. Participants were then assigned to work on puzzles collaboratively (with one other participant) or individually. All participants attempted to solve the chess puzzles at a computer station. For those working individually, participants worked at their own computer stations to select and submit solutions. For those working collaboratively, both participants were placed at the same computer station. Participants were instructed to work together however they see fit, and one individual was randomly chosen to select and enter responses for the group. Experimental instructions and tasks were presented via Qualtrics., an online survey platform. In the Qualtrics program, participants selected a button that would then open the Lichess chess puzzle in a new tab. Participants were given up to three minutes to solve each puzzle. If participants solved a chess puzzle in less than three minutes, they were instructed to close out of the chess window and return to the previous survey window. If the puzzle had not been solved within the three-minute time window, participants heard a chime sound effect, which indicated to participants that time was up and they should close out of the puzzle window to return to the survey program. All 20 puzzles were presented in this manner with the easy and hard puzzles presented in a random order. After all 20 chess puzzles were attempted, participants answered a questionnaire on their task experience, prior chess experience, and a standard demographic questionnaire. Finally, participants were given a debriefing sheet that included information about the experiment and thanked for their participation. 

Results

Outcomes of primary interest suggest that errors were substantially reduced when solving problems in a group compared to alone. This pattern was observed for both easy and hard puzzles. When attempting to solve the easy puzzles, nearly twice as many errors were made per puzzle when working individually (M = 3.9, SD = 2.3) compared to collaboratively (M = 2.0, SD = 1.3). Similarly, when attempting to solve the hard puzzles, nearly twice as many errors were made per puzzle when working individually (M = 10.2, SD = 6.6) compared to collaboratively (M = 5.6, SD = 4.8). It was also evident that the hard puzzles were in fact more challenging than the easy puzzles for all participants, as the overall number of errors per puzzle was substantially greater for these puzzles. These results provide initial evidence that working collaboratively during chess problem-solving may yield some benefit. 

Exploratory outcomes revealed that the number of puzzles successfully completed was approximately the same for participants who worked in a group compared to alone. Specifically, the individual group successfully solved 18 puzzles on average and the collaborative group successfully solved approximately 19 puzzles on average. Concerning the easy puzzles, both the individual group and the collaborative group successfully solved 10 puzzles on average, and for the hard puzzles, participants working individually successfully solved 8 puzzles on average and participants working collaboratively successfully solved 9 puzzles on average.

Discussion

The goal for the current research was to investigate how collaboration impacts the solving of chess puzzles. Results revealed that working collaboratively compared to individually leads to a similar number of correctly solved chess puzzles, but that fewer errors were made when working in a group. This pattern of results was observed regardless of item difficulty. 

What might explain the observed pattern of results? One possible explanation relates to research that has shown group members can eliminate errors through the collaborative process and feedback, known as error pruning (Ross et al., 2008). For example, in the context of the current study, a group member may have suggested a piece and its move, but the other group member may have believed another move was better and shared that suggestion. This process of deliberation and ultimate decision-making would not have been available for those individuals working alone. 

In sum, outcomes from the current research provide novel evidence that working collaboratively compared to individually may reduce the numbers of errors made when solving both easy and difficult chess puzzles. Future research will need to further evaluate the replicability and generalizability of the effects observed here, while also extending investigations to explore other factors, such as time on task, chess expertise, and/or opening versus mid-game versus end-game time points in the chess game.

References

Basden, B. H., Basden, D. R., Bryner, S., & Iii, R. L. T. (1997). A comparison of group and individual remembering: Does collaboration disrupt retrieval strategies? Journal Experimental Psychology Learning, Memory, and Cognition, 23(5), 1176-91. doi: 10.1037//0278-7393.23.5.1176 

Blumen, H. M., & Rajaram, S. (2008). Influence of re-exposure and retrieval disruption during group collaboration on later individual recall. Memory, 16(3), 231–244. https://doi.org/10.1080/09658210701804495

Knopps, A. G., & Wissman, K. T., (Under Revision). Are two heads better than one? Investigating the influence of collaboration on creativity. Manuscript submitted for publication

Kohn, N. W., & Smith, S. M. (2011). Collaborative fixation: Effects of others’ ideas on brainstorming. Applied Cognitive Psychology, 25(3), 359–371. https://doi.org/10.1002/acp.1699

Marion, S. B., & Thorley, C. (2016). A meta-analytic review of collaborative inhibition and post collaborative memory: Testing the predictions of the retrieval strategy disruption hypothesis. Psychological Bulletin, 142(11), 1141–1164. https://doi.org/10.1037/bul0000071

Ross, M., Spencer, S. J., Blatz, C. W., & Restorick, E. (2008). Collaboration reduces the frequency of false memories in older and younger adults. Psychology and Aging, 23(1), 85–92. https://doi.org/10.1037/0882-7974.23.1.85

Chessable looks forward to applications for the 2025 cycle of the Chessable Research Awards. Applications are open until May 15, 2025. For more information, please visit the Chessable Research Awards site.