Skip to content
violetelegy edited this page Oct 12, 2012 · 6 revisions

While testing the HoltWintersNew() function using table 7.2, I found some errors in table 7.2:

for example:

when alpha =0.2

it gives:

l_5 = alpha y_5 + (1-alpha) * 1_4 = 0.2 * 456.3 + (1-0.2)* 444.5 = 446.86

if we round to the nearest tenth, we should get l_5 = 446.9 but not 446.8 in the table.

when alpha=0.6,

l_8 = 0.6485.1 + 0.4434.6 = 464.9 but not 459.9 in the table.

there are many of this kind of errors in table 7.2, especially when alpha = 0.86

below is the comparison between table 7.2 and my calculation:

alpha=0.2

table7.2, my calculation

446.7 446.7

448.2 448.26

449.7 449.748

444.5 444.5184

446.8 446.8747

445.6 445.6198

441.5 441.5558

450.3 450.2647

461.4 461.4117

474.5 474.4894

482.5 482.4515

484.8 484.8012

alpha = 0.6 (a lot of dis-match)

table7.2, my calculation

446.7 446.7

450.6 451.38

453.1 453.972

438.4 435.7488

447.3 448.0795

444 443.5918

434.6 432.6167

459.9 464.1067

483 489.2427

504.9 511.7771

509.6 513.2908

501.9 501.8363

alpha=0.89

table7.2, my calculation

446.7 446.7

453.6 453.642

455.4 455.4736

427.1 427.1061

453.1 453.0887

441.9 441.9738

427.1 427.1341

478.9 478.7238

503.1 502.9996

524.2 524.182

515.3 515.387

496.5 496.5306