@September 30, 2022 | @Untitled
- 以下內容適用於所有DSE科目。
- No cut off are ‘HKEAA Official’ , 永遠都係放榜後買卷嘅考生將自己成績自發放上網再同更多放上網嘅成績整合。 「我今年DSE，XX科要考幾多分先有5*？」考之前、考當日、放榜嗰刻，都係冇人知，出卷嗰班都唔知，只能靠舊數據估。
- 齋睇 cut-off 冇意義，要睇埋當年達標人數先有用。點解？下個point講
- From 考評局：
- 即係 Lv1-5 用分數cut, 唔理幾多人考到, 分數應該年年差唔多, 至於幾多人達標就睇嗰年啲考生自己造化(🪖Supposedly🪖)； 但 5* 5** 就 拉Curve, 係 Lv5果堆考生入面Top 10% 就 5**, next 30% 5*，即係嗰年多人拎 5 就自然多人拎5**，比例上年年一致(🪖Supposedly🪖) 詳情呢度
- 我估啦，就係留返個彈性比佢哋每年「輕微調整」：即係根據返當年考生水平修改啲數字，令到出嚟盤數同以往唔爭太遠/符合佢哋心中嘅期望甚至agenda 🪖 純屬虛構對白🪖 ”aiyaa 今年又有好多考生中文陰公豬 2 咗啦，丫 唔知我哋應該點做先可以幫到班學生呢~ ” ”eee 今年純粹因為疫情而完全冇其他原因，我哋取消咗廣東話口試嚟緊取消埋聆聽綜合，咦多咗考生拎3喎~ ”
OK 人話時間：
Anyway個真相係點都相信你 too tired to care 㗎啦😅 你要嘅資料就係呢度：
- 事實係咪咁呢？先假設佢哋冇講大話🪖🪖，成件事最大嘅變數就係：每年 幾多分 先攞 3 / 4 / 5？ 咩嘢成績攞咩Level 佢哋其實有講嘅，但係用文字表述，有興趣可以睇吓。但若果你睇完都唔知自己睇左乜嘢，放心有問題嗰個唔係你。
DSE Math Core
「咁姐係MC要啱幾多題呀？」
我係咪好醒估到你想問乜呢？梗係唔係因為我識讀心啦，你估吓我教咁多年書（10＋年）有幾多學生問過同樣嘅問題？唔使估啦幫你用 excel 計埋喇：
5** | 91.5% | 93.9% | 168 | 90.4% | 169 | 94.0% | 171 | 90.4% | 170 | 89.5% | 170 | 92.9% | 170 | 93.4% | 171 | 92.3% | 169 | 92.4% | 170 | 88.6% | 178 | 88.9% | 176 | |
5* | 85.1% | 86.6% | 155 | 84.0% | 157 | 86.8% | 158 | 84.0% | 158 | 83.2% | 158 | 86.9% | 159 | 87.4% | 160 | 86.9% | 159 | 87.0% | 160 | 81.1% | 163 | 81.8% | 162 | |
5 | 78.3% | 79.3% | 142 | 75.9% | 142 | 79.1% | 144 | 76.1% | 143 | 76.8% | 146 | 82.0% | 150 | 79.8% | 146 | 79.8% | 146 | 79.3% | 146 | 74.1% | 149 | 78.8% | 156 | |
4 | 64.5% | #VALUE! | X | 62.0% | 116 | 62.1% | 113 | 65.4% | 123 | 62.6% | 119 | 66.7% | 122 | 66.7% | 122 | 65.0% | 119 | 65.2% | 120 | 61.2% | 123 | 67.7% | 134 | |
3 | 53.9% | #VALUE! | X | 55.1% | 103 | 49.5% | 90 | 53.2% | 100 | 55.3% | 105 | 54.6% | 100 | 60.1% | 110 | 51.9% | 95 | 48.9% | 90 | 52.2% | 105 | 58.1% | 115 |
LQ (Paper1)
Lvl | Avg | 2022 | 2021 | 2020 | 2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 | 2012 | |
5** | 96 | 99 | 95 | 99 | 95 | 94 | 98 | 98 | 97 | 97 | 93 | 93 | |
5* | 89 | 91 | 88 | 91 | 88 | 87 | 91 | 92 | 91 | 91 | 85 | 86 | |
5 | 82 | 83 | 80 | 83 | 80 | 81 | 86 | 84 | 84 | 83 | 78 | 83 | |
4 | 68 | #VALUE! | 65 | 65 | 69 | 66 | 70 | 70 | 68 | 68 | 64 | 71 | |
3 | 57 | #VALUE! | 58 | 52 | 56 | 58 | 57 | 63 | 55 | 51 | 55 | 61 |
MC (Paper2)
Lvl | Avg | 2022 | 2021 | 2020 | 2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 | 2012 |
5** | 41 | 42 | 41 | 42 | 41 | 40 | 42 | 42 | 42 | 42 | 40 | 40 |
5* | 38 | 39 | 38 | 39 | 38 | 37 | 39 | 39 | 39 | 39 | 36 | 37 |
5 | 35 | 36 | 34 | 36 | 34 | 35 | 37 | 36 | 36 | 36 | 33 | 35 |
4 | 29 | #VALUE! | 28 | 28 | 29 | 28 | 30 | 30 | 29 | 29 | 28 | 30 |
3 | 24 | #VALUE! | 25 | 22 | 24 | 25 | 25 | 27 | 23 | 22 | 24 | 26 |
然後你想問：「萬一我 paper1 得85分，咁MC要啱多幾多題先拉到5*啊？」估到你啦~ ~ 咁相信你都知 LQ 同 MC 有比重嘅，官方話係65% : 35%，事實係咪咁呢？再次感謝歷代考生自發提數據，令我哋有呢個表：
LQ : MC | 官方 | Avg | 2022 | 2021 | 2020 | 2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 | 2012 |
LQ 真分 | 冇講實 | 114.5 | 104 | 110 | 106 | 117 | 112 | 114 | 117 | 115 | 111 | 125 | 129 |
MC 真分 | 冇講實 | 72.8 | 75 | 77 | 76 | 71 | 78 | 69 | 69 | 68 | 73 | 76 | 69 |
LQ 真% | 65% | 61.10% | 58.1% | 58.8% | 58.2% | 62.2% | 58.9% | 62.3% | 62.9% | 62.8% | 60.3% | 62.2% | 65.2% |
MC 真% | 35% | 38.90% | 41.9% | 41.2% | 41.8% | 37.8% | 41.1% | 37.7% | 37.1% | 37.2% | 39.7% | 37.8% | 34.8% |
「咁即係點呀？」 即係兩個要點。第一：引用我朋友 @fergus.dsemath：”65:35 is a joke”。 首先睇平均值就知 LQ 重要性比官宣低，你再仔細睇更會發現佢唔單止低，仲係越嚟越低！最近嗰5年入面有4年 LQ 佔分甚至 < 59%！65%*0.9 = 58.5%，一句講哂：
LQ 打九折！
係，無錯，所以 MC 真係好緊要，it’s a fact. 所以 LQ 表現好過 MC 嘅同學，either you change, or you die.
第二點我覺得好實際：一題 MC 等於 LQ 幾多分？
已知 LQ滿分105 MC滿分45，假設 LQ : MC = 59 : 41，其實初中數學嚟姐，立志數學拎 5 或以上 嘅同學仔我希望你有能力自己計，計完先睇下面答案。
Solution
MC 1分 : LQ 1分 = (1/45)*0.41 : (1/105)*0.59 = 41/45 : 59/105 = 287 : 177 ~ 1.62 : 1 。
即係如果你 LQ跌咗5分 你 MC要啱多3題、8分LQ 約等於 5題MC。but again, no one actually knows. 我冇研究其他科，但數學core兩卷比重的確係年年轉。點解要轉？見上面 pt. 7 。
結語
「 一係唔做，一做做最好」應該係我多年來嘅作風同時係我嘅惡習。相信講 Cut Off 來講我應該係全香港做得最詳盡+講得最詳細同清晰。覺得呢篇文幫到你嘅咪俾個 like 我同 share 下囉~ d 數字之後有用咪 bookmark 埋囉~ 如果覺得冇用嘅話就……算啦
@September 30, 2022 | @Untitled