Why Numbercade works: the research behind every game

The logic model

The principles we used to build Numbercade on and the research behind them

Each row pairs a teaching principle with how our games apply it, followed by the supporting research. We tag every study by its method and its strength. For the method, we rely on randomized controlled trials (RCTs), meta-analyses, or high-quality quasi-experiments (QEDs) that measure actual causal impact. For the strength grade, we label large trials and meta-analyses as Strong, while smaller trials and solid quasi-experiments are marked Good. If the U.S. Department of Education's What Works Clearinghouse has reviewed the practice, we include their rating as an additional data point.

Layer A: Research that identifies best practices in general learning

The principle	How Numbercade does it	The evidence
Spreading practice across several days and blending problem types outperforms marathon study sessions, and actively recalling an answer builds stronger retention than rereading it.	Fifteen-minute sessions, worlds sized to a single sitting with a daily cap, mixed problem types, and a Review Missed mode that resurfaces questions a child got wrong days earlier.	Meta-analysis + RCTsStrong Classroom randomized trials in middle-school math found interleaved practice roughly doubled scores a month later (Rohrer 2015, 2020), and a meta-analysis of 317 experiments confirms distributed practice beats massed (Cepeda 2006). The What Works Clearinghouse independently rates retrieval practice Strong (Pashler 2007).
Studying worked examples that contain mistakes, and explaining what went wrong, deepens understanding more than correct examples alone.	A Wrong Map detective mode where kids find and fix errors seeded from the mistakes students actually make.	RCTsStrong Randomized classroom experiments in beginning algebra improved conceptual understanding, with the largest gains for the students furthest behind (Booth 2013, replicated in a 2015 trial of roughly 6,000 students).
Beginners learn most from fully worked solutions, and withdrawing that support gradually as skill grows beats removing it all at once.	Lessons show complete worked solutions, the step-by-step solver supports each move, then challenges and open practice take the rails off.	Meta-analysisStrong A meta-analysis of 43 studies found worked examples reliably improve mathematics performance (Barbieri 2023), and research reviews show the benefit is largest for beginners and fades as expertise grows (Kalyuga 2003).
Asking learners to explain why a step works, or why a wrong move is wrong, produces deeper understanding than practice alone.	"Pick the why" moments ask which reason makes a step wrong, and lesson checkpoints ask why a rule holds.	Meta-analysisStrong A meta-analysis of self-explanation prompts found a moderate overall effect (g=0.55; Bisra 2018), and inside an educational game, menu-based explanation prompts markedly improved transfer to new problems (Mayer & Johnson 2010).
Immediate feedback that explains the error teaches more than a bare right-or-wrong mark.	Every answer is marked instantly, every miss shows a short explanation of the specific mistake, and missed questions return for review.	Meta-analysisStrong Meta-analyses of computer-based practice find feedback that explains substantially outperforms right/wrong marks alone (0.49 versus 0.05; Van der Kleij 2015), with larger gains the more the feedback explains (Wisniewski 2020).
Letting a wrong answer become another chance to practice, rather than a failure that ends the session, sustains effort and learning.	Running low on hearts pauses the run and routes to review; answering earlier missed questions restores them, and the copy stays focused on effort and strategy.	Meta-analysesStrong The healing loop is retrieval practice, which the What Works Clearinghouse rates Strong. Allowing productive struggle before instruction improves learning (Sinha & Kapur 2021), and low-stakes quizzing has been shown to lower test anxiety rather than raise it (Yang 2023).
Small rewards can motivate without crowding out a child's own interest, as long as they stay modest and tied to real work.	One gem per correct answer in practice, capped at fifteen a day, plus badges. Nothing to buy, no ads, no prize economy.	Meta-analysisStrong Gamification meta-analyses find positive effects on learning (Sailer & Homner 2020), while the motivation literature warns that large expected prizes can undermine intrinsic interest (Deci 1999), which is why our rewards stay small and capped.
Wrapping the same math in an appealing story, with humor built into the concept rather than draped over it, raises both motivation and learning.	The wily fox's "deals" teach percents and the villain's wrong answers teach error-spotting, so each game runs on a story instead of a worksheet.	RCTs + meta-analysisStrong In a randomized trial, wrapping identical math in a fantasy context raised both motivation and how much fourth and fifth graders learned (Cordova & Lepper 1996), and randomized studies of warm, content-linked visuals show gains in comprehension and memory (Mayer & Estrella 2014; meta-analysis Brom 2018).
The strongest gains from educational games come from specific, tested design features, not from "gamification" in general.	Vocabulary cards before each challenge, one concept per screen, and characters who talk like people.	Meta-analysisStrong Research reviews and meta-analyses find that pretraining, personalization, coaching, and self-explanation are the game features with the best evidence (Mayer 2019), and that thoughtfully augmented games outperform standard versions of the same game (Clark 2016).

Layer B: Research that identifies best practices in math learning

The principle	How Numbercade does it	The evidence
Explicit, step-by-step instruction paired with immediate corrective feedback.	Every game teaches in small steps and reacts to each move; the equation solver has kids choose each step and marks it right away.	Meta-analysis + QEDStrong A meta-analysis of 50 controlled evaluations of step-based tutoring systems found a median gain of 0.66 standard deviations over conventional instruction (Kulik & Fletcher 2016), reinforced by a matched algebra classroom study (Witzel 2003). The What Works Clearinghouse rates explicit, systematic instruction Strong in two separate practice guides (Gersten 2009; Fuchs 2021).
Visual representations, with the number line at the center, make mathematical concepts easier to grasp.	Number line, tape diagram, percent bar, and symbols shown side by side and translated between.	RCTsStrong Randomized trials in the target age band found representation-rich instruction improved fraction learning (Fuchs 2013, N=259 fourth graders; Cramer 2002, 66 classrooms). The practice is rated Strong in two separate What Works Clearinghouse guides (Woodward 2012; Fuchs 2021).
Introducing an idea with physical objects first, then pictures, then symbols: the concrete, representational, abstract (CRA) sequence.	Every lesson arc runs concrete, then visual, then abstract, on purpose: a treasure chest, then a diagram, then the symbols.	QED + RCTGood A matched algebra classroom study found the concrete-to-abstract sequence outperformed conventional instruction (Witzel 2003), and randomized concreteness-fading experiments confirm the fading step matters (McNeil & Fyfe 2012). The What Works Clearinghouse rates the practice Strong (Fuchs 2021).
Reading the structure of an expression before computing, so kids stop working blindly left to right or acting on only the first term.	"Read the structure first" prompts, and Order Up! targets the classic left-to-right and first-term mistakes.	RCTsGood Randomized trials show that teaching children the meaning of mathematical notation, such as what the equals sign represents, improves their grasp of expression structure (McNeil 2019; Donovan 2022, grades 4 to 5). The reflective-prompt component is rated Strong by the What Works Clearinghouse (Woodward 2012).
Comparing two worked strategies side by side builds flexible, transferable problem-solving.	"Deckhand vs. Navigator" and "Two Cooks" have kids compare two worked strategies and judge them.	RCTGood A randomized experiment with seventh graders found that comparing strategies improved procedural skill and flexibility (Rittle-Johnson & Star 2007). The What Works Clearinghouse rates the practice Moderate (Star 2015).
Teaching the real mathematical words and notation explicitly, rather than only informal nicknames.	Frayer-style glossaries where the real term ("isolate") leads and the friendly gloss follows, never the reverse.	Practice guidesStrong Rated Strong in the What Works Clearinghouse elementary guide (Fuchs 2021), with a supporting recommendation on connecting informal language to formal concepts and notation (Woodward 2012).
Writing out and monitoring your own steps, rather than guessing at an answer, improves problem-solving.	"Grab scratch paper" prompts tell kids to write out the steps before answering.	Meta-analysis + experimentStrong A self-explanation meta-analysis (g=0.55; Bisra 2018) and a classroom experiment teaching eighth graders to self-question as they solve (Kramarski & Mevarech 2003) show that externalizing and monitoring reasoning improves math problem-solving. The What Works Clearinghouse rates monitoring-and-reflection prompts Strong (Woodward 2012).

Layer C: Research that identifies best practices for each math topic

Topic	The evidence	How Numbercade uses it
Fractions, ratios, and percents (Fruit Market)	RCTs + meta-analysisStrong Randomized trials with at-risk fourth and fifth graders improved fraction skills (Fuchs 2013, N=259; Jayanthi 2021, N=205), and a meta-analysis of 43 rational-number studies found large gains for struggling learners (Rojo 2023).	Fruit Market puts every fraction on a number line before pie slices, and teaches ratios through unit rates before cross-multiplication.
Equations and algebra (Equation Pirates)	RCTsStrong The error-analysis experiments (Booth 2013, 2015), the strategy-comparison experiment (Rittle-Johnson & Star 2007), and the concrete-to-abstract study (Witzel 2003) were all run in algebra or equation-solving classrooms.	Equation Pirates teaches solving as keeping a scale balanced, sets two solver strategies side by side, and turns its boss fight into an error-analysis puzzle.
Order of operations (Order Up!)	RCTsGood Randomized trials show that teaching children the meaning of mathematical notation, such as what the equals sign represents, improves their grasp of expression structure (McNeil 2019; Donovan 2022, grades 4 to 5). The specific left-to-right and first-term errors are well documented (Linchevski & Livneh 1999).	Order Up! trains kids to read an expression's structure before computing, targeting the classic left-to-right and first-term mistakes.
Slope and linear relationships (Slope Quest)	Cluster RCTStrong Large school-randomized experiments found that connecting graphs, tables, and equations improved rate and linear-function learning in seventh grade, and the result replicated across settings (SimCalc: about 1,600 students, effect sizes 0.50 to 0.63; Roschelle 2010).	Slope Quest links the graph, the table, and the equation for the same line, so rate of change is something kids watch move.
Geometry and measurement (Master Builder)	RCTs + meta-analysisStrong A randomized study of about 17,000 children found that on-screen spatial training improved math learning (Judd & Klingberg 2021), a smaller trial linked spatial practice to solving missing-number problems (Cheng & Mix 2014), and a meta-analysis confirms spatial training improves math on average (Hawes 2022).	Master Builder trains spatial reasoning through building and measuring tasks that ask kids to picture, hold, and rotate shapes in their heads.
Probability and statistics (Loot Lab)	Randomized experimentsGood Teaching probability as whole-number counts ("3 out of 10") rather than abstract percentages sharply improves correct reasoning, a format shown to work in fourth to sixth graders (Zhu & Gigerenzer 2006; Sedlmeier & Gigerenzer 2001). The equiprobability and representativeness biases are robustly documented (Lecoutre 1992; Kahneman & Tversky 1972).	Loot Lab builds chance from counting real outcomes rather than percentages, and directly confronts the "it's all just luck" misconception.

Why these skills at all? Fraction and division knowledge around age 10 uniquely predicts algebra and overall math achievement around age 16, controlling for IQ, family income, and more (Siegler 2012, two longitudinal cohorts). Although the data here does not establish a strict causal relationship, the underlying intuition is clear: these early concepts are the essential building blocks required to master advanced subjects like algebra and calculus.

Which topics made the cut, and why only seven: the curriculum page walks through the full rationale. The short version is that the games target the National Mathematics Advisory Panel's Critical Foundations of Algebra (whole numbers, fractions, and the geometry that slope depends on), a 2008 expert panel synthesis rather than trial evidence, and the documented misconceptions in this table concentrate in exactly those topics.

The receipts

Full reference list

Every source below was verified against its primary publication before it appeared here or in the games.

Barbieri, C. A., Miller-Cotto, D., Clerjuste, S. N., & Chawla, K. (2023). A meta-analysis of the worked examples effect on mathematics performance. Educational Psychology Review, 35(1), Article 11.
Bisra, K., Liu, Q., Nesbit, J. C., Salimi, F., & Winne, P. H. (2018). Inducing self-explanation: A meta-analysis. Educational Psychology Review, 30(3), 703-725.
Booth, J. L., Lange, K. E., Koedinger, K. R., & Newton, K. J. (2013). Using example problems to improve student learning in algebra. Learning and Instruction, 25, 24-34.
Booth, J. L., Oyer, M. H., Paré-Blagoev, E. J., et al. (2015). Learning algebra by example in real-world classrooms. Journal of Research on Educational Effectiveness, 8(4), 530-551.
Brom, C., Stárková, T., & D’Mello, S. K. (2018). How effective is emotional design? A meta-analysis. Educational Research Review, 25, 100-119.
Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354-380.
Cheng, Y.-L., & Mix, K. S. (2014). Spatial training improves children’s mathematics ability. Journal of Cognition and Development, 15(1), 2-11.
Clark, D. B., Tanner-Smith, E. E., & Killingsworth, S. S. (2016). Digital games, design, and learning: A systematic review and meta-analysis. Review of Educational Research, 86(1), 79-122.
Cordova, D. I., & Lepper, M. R. (1996). Intrinsic motivation and the process of learning. Journal of Educational Psychology, 88(4), 715-730.
Cramer, K. A., Post, T. R., & delMas, R. C. (2002). Initial fraction learning by fourth- and fifth-grade students. Journal for Research in Mathematics Education, 33(2), 111-144.
Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6), 627-668.
Donovan, A. M., Stephens, A., Alapala, B., et al. (2022). Is a substitute the same? Learning from lessons centering different relational conceptions of the equal sign. ZDM Mathematics Education, 54(6), 1199-1213.
Fuchs, L. S., Schumacher, R. F., Long, J., et al. (2013). Improving at-risk learners’ understanding of fractions. Journal of Educational Psychology, 105(3), 683-700.
Fuchs, L. S., Newman-Gonchar, R., Schumacher, R., et al. (2021). Assisting Students Struggling with Mathematics: Intervention in the Elementary Grades. What Works Clearinghouse (WWC 2021006).
Gersten, R., Beckmann, S., Clarke, B., et al. (2009). Assisting Students Struggling with Mathematics: Response to Intervention (RtI) for Elementary and Middle Schools. National Center for Education Evaluation (NCEE 2009-4060).
Hawes, Z. C. K., Gilligan-Lee, K. A., & Mix, K. S. (2022). Effects of spatial training on mathematics performance: A meta-analysis. Developmental Psychology, 58(1), 112-137.
Jayanthi, M., Gersten, R., Taylor, M. J., et al. (2021). The impact of a fractions intervention on fifth graders with mathematics difficulties. Exceptional Children, 88(1), 81-100.
Judd, N., & Klingberg, T. (2021). Training spatial cognition enhances mathematical learning in a randomized study of 17,000 children. Nature Human Behaviour, 5, 1548-1554.
Kahneman, D., & Tversky, A. (1972). Subjective probability: A judgment of representativeness. Cognitive Psychology, 3(3), 430-454.
Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect. Educational Psychologist, 38(1), 23-31.
Kramarski, B., & Mevarech, Z. R. (2003). Enhancing mathematical reasoning in the classroom: The effects of cooperative learning and metacognitive training. American Educational Research Journal, 40(1), 281-310.
Kulik, J. A., & Fletcher, J. D. (2016). Effectiveness of intelligent tutoring systems: A meta-analytic review. Review of Educational Research, 86(1), 42-78.
Lecoutre, M.-P. (1992). Cognitive models and problem spaces in “purely random” situations. Educational Studies in Mathematics, 23(6), 557-568.
Linchevski, L., & Livneh, D. (1999). Structure sense: The relationship between algebraic and numerical contexts. Educational Studies in Mathematics, 40(2), 173-196.
Mayer, R. E. (2019). Computer games in education. Annual Review of Psychology, 70, 531-549.
Mayer, R. E., & Estrella, G. (2014). Benefits of emotional design in multimedia instruction. Learning and Instruction, 33, 12-18.
Mayer, R. E., & Johnson, C. I. (2010). Adding instructional features that promote learning in a game-like environment. Journal of Educational Computing Research, 42(3), 241-265.
McNeil, N. M., & Fyfe, E. R. (2012). Concreteness fading promotes transfer of mathematical knowledge. Learning and Instruction, 22(6), 440-448.
McNeil, N. M., Hornburg, C. B., Brletic-Shipley, H., & Matthews, J. M. (2019). Improving children’s understanding of mathematical equivalence. Journal of Educational Psychology, 111(7), 1023-1044.
Pashler, H., Bain, P., Bottge, B., et al. (2007). Organizing Instruction and Study to Improve Student Learning. National Center for Education Research (NCER 2007-2004).
Rittle-Johnson, B., & Star, J. R. (2007). Does comparing solution methods facilitate conceptual and procedural knowledge? Journal of Educational Psychology, 99(3), 561-574.
Rohrer, D., Dedrick, R. F., & Stershic, S. (2015). Interleaved practice improves mathematics learning. Journal of Educational Psychology, 107(3), 900-908.
Rohrer, D., Dedrick, R. F., Hartwig, M. K., & Cheung, C.-N. (2020). A randomized controlled trial of interleaved mathematics practice. Journal of Educational Psychology, 112(1), 40-52.
Rojo, M., King, S. E., Gersib, J., & Bryant, D. P. (2023). Rational number interventions for students with mathematics difficulties: A meta-analysis. Remedial and Special Education, 44(3), 225-238.
Roschelle, J., Shechtman, N., Tatar, D., et al. (2010). Integration of technology, curriculum, and professional development for advancing middle school mathematics. American Educational Research Journal, 47(4), 833-878.
Sailer, M., & Homner, L. (2020). The gamification of learning: A meta-analysis. Educational Psychology Review, 32, 77-112.
Sedlmeier, P., & Gigerenzer, G. (2001). Teaching Bayesian reasoning in less than two hours. Journal of Experimental Psychology: General, 130(3), 380-400.
Siegler, R. S., Duncan, G. J., Davis-Kean, P. E., et al. (2012). Early predictors of high school mathematics achievement. Psychological Science, 23(7), 691-697.
Sinha, T., & Kapur, M. (2021). When problem solving followed by instruction works: Evidence for productive failure. Review of Educational Research, 91(5), 761-798.
Star, J. R., Caronongan, P., Foegen, A., et al. (2015). Teaching Strategies for Improving Algebra Knowledge in Middle and High School Students. National Center for Education Evaluation (NCEE 2015-4010).
Van der Kleij, F. M., Feskens, R. C. W., & Eggen, T. J. H. M. (2015). Effects of feedback in a computer-based learning environment on students’ learning outcomes. Review of Educational Research, 85(4), 475-511.
Witzel, B. S., Mercer, C. D., & Miller, M. D. (2003). Teaching algebra to students with learning difficulties. Learning Disabilities Research & Practice, 18(2), 121-131.
Wisniewski, B., Zierer, K., & Hattie, J. (2020). The power of feedback revisited: A meta-analysis. Frontiers in Psychology, 10, 3087.
Woodward, J., Beckmann, S., Driscoll, M., et al. (2012). Improving Mathematical Problem Solving in Grades 4 Through 8. National Center for Education Evaluation (NCEE 2012-4055).
Yang, C., Li, J., Zhao, W., Luo, L., & Shanks, D. R. (2023). Do practice tests (quizzes) reduce or provoke test anxiety? A meta-analytic review. Educational Psychology Review, 35, Article 87.
Zhu, L., & Gigerenzer, G. (2006). Children can solve Bayesian problems. Cognition, 98(3), 287-308.

Spotted an error in a citation, or a claim that outruns its evidence? Tell us: info@numbercade.com. We treat corrections as a feature.

We show our homework.

The principles we used to build Numbercade on and the research behind them

Layer A: Research that identifies best practices in general learning

Layer B: Research that identifies best practices in math learning

Layer C: Research that identifies best practices for each math topic

Full reference list

Play the arcade