* Segmented Analysis of Russian Casualties in Ukraine on Thu Apr 02 15:29:45 2026 - input data directory: ./results - results directory: ./results - transcript to: ./results/segmented-ukr-rus-casualties-transcript.txt Loading updated data: --------------------- * Updated data file: ./results/russian-casualties-in-ukraine-updated-2026-04-02.tsv - Found 124 rows x 3 columns: - Columns: DayNum, Date, Soldiers Doing 3-fold crossvalidated segmented fits: ------------------------------------------- - Segmented fit, testFold = 1 o Train data: 82 points o Test data: 42 points o Simple linear fit, no kinks: Call: lm(formula = Soldiers ~ DayNum, data = trainData) Residuals: Min 1Q Median 3Q Max -85195 -7929 3396 9245 47654 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 1.054e+05 2.172e+03 48.53 <2e-16 *** DayNum 9.813e+02 9.676e+00 101.42 <2e-16 *** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 17140 on 80 degrees of freedom Multiple R-squared: 0.9923, Adjusted R-squared: 0.9922 F-statistic: 1.028e+04 on 1 and 80 DF, p-value: < 2.2e-16 o Davies test for need of kink: Davies' test for a change in the slope data: formula = Soldiers ~ DayNum , method = lm model = gaussian , link = identity segmented variable = DayNum 'best' at = 433.67, n.points = 10, p-value < 2.2e-16 alternative hypothesis: two.sided o Segmented fit: ***Regression Model with Segmented Relationship(s)*** Call: segmented.lm(obj = linearFit) Estimated Break-Point(s): Est. St.Err psi1.DayNum 404.118 12.476 Coefficients of the linear terms: Estimate Std. Error t value Pr(>|t|) (Intercept) 122569.14 1037.73 118.11 <2e-16 *** DayNum 698.58 15.49 45.11 <2e-16 *** U1.DayNum 525.40 17.89 29.36 NA --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 4483 on 78 degrees of freedom Multiple R-Squared: 0.9995, Adjusted R-squared: 0.9995 Boot restarting based on 9 samples. Last fit: Convergence attained in 2 iterations (rel. change 5.2427e-10) Est. CI(95%).low CI(95%).up psi1.DayNum 404.118 379.281 428.955 $DayNum Est. St.Err. t value CI(95%).l CI(95%).u slope1 698.58 15.4880 45.106 667.75 729.42 slope2 1224.00 8.9584 136.630 1206.10 1241.80 - Segmented fit, testFold = 2 o Train data: 83 points o Test data: 41 points o Simple linear fit, no kinks: Call: lm(formula = Soldiers ~ DayNum, data = trainData) Residuals: Min 1Q Median 3Q Max -93617 -7815 3807 10070 42553 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 1.048e+05 2.352e+03 44.54 <2e-16 *** DayNum 9.886e+02 9.419e+00 104.95 <2e-16 *** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 18800 on 81 degrees of freedom Multiple R-squared: 0.9927, Adjusted R-squared: 0.9926 F-statistic: 1.102e+04 on 1 and 81 DF, p-value: < 2.2e-16 o Davies test for need of kink: Davies' test for a change in the slope data: formula = Soldiers ~ DayNum , method = lm model = gaussian , link = identity segmented variable = DayNum 'best' at = 358.33, n.points = 10, p-value < 2.2e-16 alternative hypothesis: two.sided o Segmented fit: ***Regression Model with Segmented Relationship(s)*** Call: segmented.lm(obj = linearFit) Estimated Break-Point(s): Est. St.Err psi1.DayNum 391.861 16.573 Coefficients of the linear terms: Estimate Std. Error t value Pr(>|t|) (Intercept) 122525.07 1378.97 88.85 <2e-16 *** DayNum 699.43 20.71 33.77 <2e-16 *** U1.DayNum 496.74 22.73 21.86 NA --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 5998 on 79 degrees of freedom Multiple R-Squared: 0.9993, Adjusted R-squared: 0.9992 Boot restarting based on 6 samples. Last fit: Convergence attained in 2 iterations (rel. change 3.7462e-10) Est. CI(95%).low CI(95%).up psi1.DayNum 391.861 358.873 424.849 $DayNum Est. St.Err. t value CI(95%).l CI(95%).u slope1 699.43 20.7130 33.768 658.2 740.66 slope2 1196.20 9.3588 127.810 1177.5 1214.80 - Segmented fit, testFold = 3 o Train data: 83 points o Test data: 41 points o Simple linear fit, no kinks: Call: lm(formula = Soldiers ~ DayNum, data = trainData) Residuals: Min 1Q Median 3Q Max -90728 -7718 3541 9713 50805 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 105160.17 2428.05 43.31 <2e-16 *** DayNum 981.19 10.13 96.89 <2e-16 *** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 19340 on 81 degrees of freedom Multiple R-squared: 0.9914, Adjusted R-squared: 0.9913 F-statistic: 9387 on 1 and 81 DF, p-value: < 2.2e-16 o Davies test for need of kink: Davies' test for a change in the slope data: formula = Soldiers ~ DayNum , method = lm model = gaussian , link = identity segmented variable = DayNum 'best' at = 433.11, n.points = 10, p-value < 2.2e-16 alternative hypothesis: two.sided o Segmented fit: ***Regression Model with Segmented Relationship(s)*** Call: segmented.lm(obj = linearFit) Estimated Break-Point(s): Est. St.Err psi1.DayNum 396.416 14.401 Coefficients of the linear terms: Estimate Std. Error t value Pr(>|t|) (Intercept) 122457.02 1213.78 100.89 <2e-16 *** DayNum 700.73 18.36 38.17 <2e-16 *** U1.DayNum 501.86 20.15 24.91 NA --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 5311 on 79 degrees of freedom Multiple R-Squared: 0.9994, Adjusted R-squared: 0.9993 Boot restarting based on 6 samples. Last fit: Convergence attained in 2 iterations (rel. change 4.4846e-10) Est. CI(95%).low CI(95%).up psi1.DayNum 396.416 367.752 425.081 $DayNum Est. St.Err. t value CI(95%).l CI(95%).u slope1 700.73 18.3580 38.17 664.19 737.27 slope2 1202.60 8.3031 144.84 1186.10 1219.10 - Segmented fit, testFold = NA o Train data: 124 points o Test data: 124 points o Simple linear fit, no kinks: Call: lm(formula = Soldiers ~ DayNum, data = trainData) Residuals: Min 1Q Median 3Q Max -91893 -8011 3589 9665 47634 Coefficients: Estimate Std. Error t value Pr(>|t|) (Intercept) 1.051e+05 1.889e+03 55.64 <2e-16 *** DayNum 9.840e+02 7.927e+00 124.13 <2e-16 *** --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 18400 on 122 degrees of freedom Multiple R-squared: 0.9921, Adjusted R-squared: 0.9921 F-statistic: 1.541e+04 on 1 and 122 DF, p-value: < 2.2e-16 o Davies test for need of kink: Davies' test for a change in the slope data: formula = Soldiers ~ DayNum , method = lm model = gaussian , link = identity segmented variable = DayNum 'best' at = 357.67, n.points = 10, p-value < 2.2e-16 alternative hypothesis: two.sided o Segmented fit: ***Regression Model with Segmented Relationship(s)*** Call: segmented.lm(obj = linearFit) Estimated Break-Point(s): Est. St.Err psi1.DayNum 396.002 11.96 Coefficients of the linear terms: Estimate Std. Error t value Pr(>|t|) (Intercept) 122516.84 1000.80 122.42 <2e-16 *** DayNum 699.58 15.03 46.53 <2e-16 *** U1.DayNum 505.33 16.70 30.26 NA --- Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 Residual standard error: 5330 on 120 degrees of freedom Multiple R-Squared: 0.9994, Adjusted R-squared: 0.9993 Boot restarting based on 6 samples. Last fit: Convergence attained in 2 iterations (rel. change 1.0633e-09) Est. CI(95%).low CI(95%).up psi1.DayNum 396.002 372.322 419.682 $DayNum Est. St.Err. t value CI(95%).l CI(95%).u slope1 699.58 15.0350 46.531 669.81 729.35 slope2 1204.90 7.2692 165.750 1190.50 1219.30 * Crossvalidation and final whole-dataset fit results: TestFold Kink sd.Kink. Slope1 sd.Slope1. Slope2 sd.Slope2. lm.Adj.R2 1 1 404.118 12.476 698.58 15.488 1224.0 8.958 0.992 2 2 391.861 16.573 699.43 20.713 1196.2 9.359 0.993 3 3 396.416 14.401 700.73 18.358 1202.6 8.303 0.991 4 NA 396.002 11.960 699.58 15.035 1204.9 7.269 0.992 Adj.R2 lm.RMSE RMSE 1 0.999 20596.01 7080.426 2 0.999 17662.80 3906.167 3 0.999 16404.75 5384.350 4 0.999 18246.72 5243.688 * Segmented Analysis of Russian Casualties in Ukraine completed Thu Apr 02 15:29:45 2026 (0.4 sec elapsed).