How we rank.
A verdict you can't audit is just an opinion with confidence. So here is every source, every refresh interval, every threshold and every weight that decides what LazyComparo tells you to buy.
Nothing here is ranked by money
No ads, no affiliate links, no sponsored placement, no pay-to-rank. A store link earns us nothing, so the ordering has nothing to protect. See About for what the site costs and who pays for it.
Where the data comes from
| What | Source | How fresh |
|---|---|---|
| Game price & current discount (SGD) | Steam's official store API, Singapore region | Live per request, cached ~30 minutes at the edge |
| Steam review score | Steam's official store API | Same request, same cache |
| Epic & GOG prices, all-time lows | IsThereAnyDeal | Live per request, cached ~30 minutes |
| This week's free games | Epic Games' official promotions feed, SG region | Live per request |
| Cover art | Steam's CDN, keyed off the game's AppID | Static |
| Hours-to-beat, pros/cons, co-op support | Editorial — our own estimates | Reviewed when the catalogue changes |
| Phone specs, spec scores and street prices | Editorial — compiled by hand for the SG market | Reviewed when the catalogue changes |
The split matters: the game prices you see are machine-read from the store seconds ago, while the phone figures and the hours-to-beat numbers are considered human estimates. We never present one as the other.
A note on currency
Steam prices arrive in SGD directly. Non-SGD store prices from IsThereAnyDeal are converted at a live public FX rate and shown with a ~ prefix to mark them as approximate. If the conversion step fails, off-currency listings are dropped from the comparison entirely rather than shown next to SGD figures as if they were comparable.
Deal heat: is this price actually good?
Every price carries one of four heat levels, and the level means the same thing everywhere it appears — on a card, in the compare table, in the price ticker on the home page.
| Heat | When |
|---|---|
| Hot | Free, or at its all-time low across tracked stores, or discounted 60% or more |
| Warm | Discounted 30–59% |
| Mild | Discounted 1–29% |
| None | Full price |
An all-time low outranks the raw percentage on purpose. 20% off a game that has never been cheaper is a better moment to buy than a routine 50% sale that comes round every quarter — and a percentage-only badge can't tell you the difference.
Dollars per hour
The value score is deliberately boring: price ÷ hours to beat. A S$70 game you'll play for 60 hours costs less per hour than a S$30 game you'll finish in four. Lower is better, and the bar next to it is scaled so the cheapest-per-hour game in your comparison is the only one in colour.
The honest caveat: hours-to-beat is an editorial estimate of a typical main-story playthrough. It is not your playtime, and for endless games it is close to meaningless.
The game verdict
When you compare games, the banner at the top picks a winner from exactly the same rows the table below renders — so the headline and the per-row award icons can never disagree. Each row a game wins adds its weight to that game's tally:
| Row | Weight | Why |
|---|---|---|
| Review rating | 2 | The strongest single signal of whether it's good |
| Dollars per hour | 2 | What you actually get for the money |
| Price | 1.5 | Budget is real, but cheap-and-bad isn't a win |
| Cheapest store | 1 | Rewards being available somewhere better |
| Hours to beat | 1 | Length, on its own terms |
| Combat / story / replay | 0.5 each | Editorial spec scores — deliberately light |
| Release year, stores, player count | 0 | Excluded — see below |
Release year is deliberately worth nothing. "Newer" is not a reason to buy a game, and when year was comparable the award icons started outnumbering the actual scoreline.
Two more rules keep the verdict honest:
- Ties award nobody, in the tally and in the table both.
- If the winning margin is under 12% of the total weight on the table, we don't declare a winner — the banner says "It's close" instead of "Get X". A ranking engine that always produces a confident answer is lying about its own precision.
- Reasons are concrete deltas against the runner-up, never adjectives, and the caveat line names the spec where the runner-up genuinely beats the winner.
The phone verdict
The phone advisor scores each candidate against the four priorities you set — camera, battery, performance, display — then subtracts the cost of leaving your current ecosystem:
raw score = Σ (your priority weight × spec score) ÷ total weight
ecosystem penalty = 0 staying on the same ecosystem · 5 moving to a different Android brand · 15 crossing between iOS and Android
adjusted score = raw score − ecosystem penalty
That penalty is the part no spec-ranking site counts, and it's the reason this tool exists. Switching costs you real things — iMessage threads, AirDrop, DeX, your accessories, an afternoon of migration — and a phone that wins on paper by two points does not win once you price that in. Crossing platforms is weighted three times heavier than changing Android brands because the losses are structural rather than habitual.
The verdict then compares the best candidate against the phone you already own:
| Gap vs your current phone | Verdict |
|---|---|
| 8 points or more | Worth upgrading |
| 3 to 8 points | Marginal — worth it only if you'll keep it 3+ years |
| Under 3 points | Wait a generation |
Most people who run this tool on a two-year-old flagship get told to keep it. That is the correct answer, and it is the answer a site paid by phone sales would never give you.
What this method can't do
- The catalogues aren't exhaustive. A game or phone missing from a comparison isn't a judgement on it — it just isn't in the list yet.
- Prices move. A sale can end between our cache refreshing and you clicking through. The store's own page is always the authority; check it before you pay.
- Regional pricing is Singapore-first. Figures are SGD for the SG storefront and won't match what you see elsewhere.
- Editorial scores are judgements. Spec scores, pros/cons and hours-to-beat are ours, they're weighted lightly on purpose, and reasonable people will disagree with some of them. Tell us when we're wrong.
When this changes
Weights and thresholds get adjusted as the tools improve, and this page is updated in the same change with the date below. The figures here are transcribed from the constants the site actually runs on, and the code is public — so if this page and the site ever disagree, the code is right and this is a bug worth reporting.
Last reviewed 9 August 2026 · Weights and thresholds current as of this date.