All toolsTan Zhi Wei, Property AdvisorTalk to Zhi Wei
Backtest · Your 11 SGX picks

What S$10,000 in each of your 11 stocks became

Bought at the closing price on 2 Jan 2020, whole shares only, broker fees in and out. Valued at the 7 Oct 2026 close, after the cost of selling.

Yearly growth rate, dividends included (not reinvested)

Stock by stock

Tap a column to sort. Tap a stock to see the workings.

Stock Current portfolio Dividends collected Net profit Yearly growth

The three banks, side by side

From each bank's own results for the six months to 30 Jun 2026. Dividend yield and price-to-book use the 7 Oct 2026 closing price. Green marks the strongest of the three. Price-to-book is not coloured. A lower figure means the shares cost less against the bank's net assets. It does not show which bank is stronger. Where your earlier table differed, its figure is shown underneath.

Measure DBS OCBC UOB

What each measure means. Net interest margin: what the bank earns on loans minus what it pays on deposits. Bad-loan ratio (NPL): share of loans not being repaid. Cost-to-income: costs as a share of income (lower is better). CET1: the bank's safety cushion of core capital (higher is safer). Price-to-book: share price ÷ the bank's net assets per share.

Capital cushion caution. DBS and OCBC quote CET1 under transitional rules; on the fully phased-in basis it is 14.6% and 14.0%. UOB's release doesn't say which basis its 15.4% uses, so compare with care.

UOB's ROE is not in UOB's own release; 11.6% comes from The Smart Investor's comparison of the banks' results.

Company check-up: your 11

Latest 12 months of results (mostly to 30 Jun 2026), each measure turned into a plain answer. Green healthy · Amber watch · Red concern · Grey neutral. Tap a column to sort, tap a stock for the full explanation.

Stock Industry Govt-linked owner Price vs its usual
P/E
Profit per S$1 owned
ROE
Beats 8% cost of money?
ROIC
Years to repay debt
replaces Debt/Equity
Can pay this year's bills?
Current ratio
Cash per S$1 profit
Operating cash flow
Dividend covered?
Free cash flow

Price vs its usual (P/E). Today's P/E against the stock's own median over the last 4 financial years. More than 20% above is "pricier than usual", more than 20% below is "cheaper". Cheaper is not automatically better: the market may expect profits to fall.

Profit per S$1 owned (ROE). Green from 12 cents, amber 8–12, red below 8. The arrow compares with 4 years ago. Where a company owes more than it owns, part of the return comes from borrowing.

Beats 8% cost of money? (ROIC). 8% is my rough estimate of what money costs a Singapore company today, given high interest rates. Above it, the business creates value; below, it doesn't.

Years to repay debt. Borrowings minus cash, divided by a year's operating cash. Green up to 3 years, amber 3–6, red above 6. "Net cash" means it holds more cash than debt.

Can pay this year's bills? (Current ratio). Short-term assets ÷ bills due within a year. Below 1 is amber, not red, when the business collects customers' money upfront.

Cash per S$1 profit. Average over 4 years, with the latest 12 months alongside. Below S$1 for years means reported profit isn't turning into cash.

Dividend covered? Spare cash after investment spending (free cash flow) ÷ dividends paid, latest 12 months. Below 1 means part of the dividend is paid from savings or borrowing.

Banks show "n/a" for the last five: a bank's debt is mostly customer deposits, so those measures don't apply. Use the bank comparison above instead.

Government-linked owner covers Temasek and other government bodies. GIC need not disclose holdings under 5% and none were disclosed, so "None" means no disclosed holding.

When do these stocks tend to rise or fall?

Tested on 20 large SGX stocks (your 11 plus 9 others from the original list) and the Straits Times Index, Jan 2010 to Sep 2026: 4,192 trading days. The intraday test uses 30-minute prices, available for the last 60 trading days only. Returns include dividends unless marked "index only".

1 · Month of the year Clearest pattern

Average monthly return of the 20 stocks, and how many of the years tested that month was up (17 years for January to September, 16 for October to December).

What it means: April has been the best month, up in 16 of 17 years and averaging +4.1%. July (+2.5%) and February (+2.4%) come next, and September is the weakest. The index looks weak in May and August only because the banks pay their dividends then: the price drops by the dividend, but shareholders receive the cash. With dividends counted, those months are flat, not negative.

2 · Same month, next year Strong

A stock that beats the others in a given calendar month tends to beat them in that same month the following year, and the year after. The link is about 6 times stronger at 12 and 24 months than at other gaps. This matches Heston and Sadka's 2008 finding for US stocks. The likely reason in Singapore is that each company's results and dividend dates fall in the same months every year.

Months apart 12 24 36 48 60 All other gaps
Link strength (correlation) 0.107 0.098 0.045 0.011 0.051 0.016

3 · Day to day Strong

A stock that beats the others today tends to give some of it back tomorrow (correlation −0.074, the strongest result in this study). There is no weekly rhythm: a stock's return 5 trading days ago tells you nothing useful.

4 · Day of the week Weak

Tuesdays have been the best day and Mondays the weakest. The gap is about 0.1–0.2% a day, smaller than the cost of a single trade.

5 · Within the day (the Heston, Korajczyk and Sadka test) Too little data

Their paper found a stock's return in a given half-hour repeats in that same half-hour on following days. Singapore shows the same sign: the same half-hour 1 to 3 days earlier is positively linked (correlation about 0.03), while neighbouring half-hours are not. The first half-hour of trading was the weakest (−0.12% on average). But this covers only 60 days that included last week's sell-off, so it is not reliable.

Can you profit from these? Mostly not. The daily and intraday effects are smaller than broker fees of 0.2% or more each way. The April effect is large, but it rests on just 17 Aprils, and patterns that become well known tend to fade. Treat these as timing hints for money you were going to invest anyway, not as a trading system.

How the numbers were worked out

Current portfolio is shares held × 7 Oct 2026 closing price, minus the cost of selling them today, plus cash from selling any extra shares or rights the company handed out. Dividends are kept separate.

Net profit is current portfolio minus S$10,000. Buying fees are already inside the S$10,000, and the few dollars left over after buying whole shares are treated as gone, as you asked.

Yearly growth is the steady yearly rate that turns S$10,000 into S$10,000 + net profit + dividends over the holding period. Dividends are not reinvested.

Broker fees: 0.12% commission (minimum S$10) plus 0.04% exchange fees, plus GST at the rate of the day (7% to 2022, 8% in 2023, 9% from 2024).

Extra shares and rights (Sembcorp 2020, Keppel 2023 and 2026): no extra money put in. Anything handed out is valued at the price it was worth on the day it was handed out and sold, with selling fees.

DBS issued 1 free share for every 10 in April 2024; the share count reflects this.

Sources: prices from Yahoo Finance, corrected back to the real traded prices; dividends from Dividends.sg (2022 onwards) and Yahoo Finance (2020–2021); company announcements on SGX for Keppel's hand-outs; timing patterns from Yahoo Finance daily and 30-minute prices. Past results do not predict future returns.