The first three cells are published guidance, graded in the source table below; none of them is computed by this site. The fourth comes from this site's public payoff engine. Money and score are kept in separate columns on purpose: this site can check the first and cannot check the second.
What is actually published, and by whom
The honest answer to "how many points" is that the sources run from a tight one to a very loose one, and the loose ones do not say where their numbers came from. So instead of averaging them, here is each claim next to its source and a grade for how much weight it can carry. Grade A means the scoring model's own publisher says it. B means a named publisher citing the model, with no number. C means an aggregator summarizing a regulator's language, with no number. D means a specific number with no method or sample — usable as a claim to check, not as evidence.
| Claim | Source | Grade |
|---|---|---|
| One additional hard inquiry lowers a FICO score by fewer than five points in most cases | myFICO, "How a Balance Transfer Impacts Your Credit" (FICO's own publisher) | A |
| Balance transfer cards generally want a FICO score of 670 or higher | myFICO, same page | A |
| Amounts owed is 30% of a FICO score; length of credit history is roughly 15% — and only those two weights are stated on that page | myFICO, same page | A |
| The other factor weights (payment history 35%, new credit 10%, mix 10%) are widely quoted but not on that page, so they are not used anywhere here | No first-party source opened for this page | — |
| The new account lowers average account age, which is "roughly 15%" of a FICO score, and the score effect "won't usually be huge" — no number given | myFICO, same page | B |
| A hard inquiry stays on the report two years; score models use inquiries from roughly the past 12 months | Card issuer guidance summarizing the bureau practice (Chase education page) | B |
| A hard inquiry "might cost you up to five points according to FICO", and for VantageScore the cost is "likely to be even more" — no figure for VantageScore | NerdWallet (2025), citing FICO | C |
| The Federal Reserve's guidance on credit inquiries caps the score damage from inquiries as small — a hard limit, not a point figure. The primary document was not opened for this page, so this row is here as a pointer, not as evidence | Aggregators cite it; CFPB's page on inquiries returned 404 on the check | C |
| An inquiry costs 3 to 7 points; the first month dips 5 to 15 points; a missed promotional period costs 60 to 100+ points | Aggregator blog, no model, sample or method published | D |
| A balance transfer drops the score by 20, 50, 110 points (various pages) | Aggregator blogs, no method; the ranges disagree with each other | D |
The D rows are the ones for sale: they are specific, they are memorable, and none of them survives the question "measured on what sample, under which model version". A 3-to-7 and a 20-to-110 cannot both be the effect of one inquiry on the same population. This page keeps them as claims to check, and does not build anything on them.
The one number with a source, and what it covers
FICO publishes "fewer than five points" for one additional hard inquiry, in most cases, and that is the number to hold on to. It covers exactly one event. It does not cover the new account, it does not cover the change in average account age, and it does not cover the shift in per-card utilization when the balance lands on the new card. Each of those is a separate input, and only the inquiry has a published figure attached to it. Note also what the same page implies about timing: the inquiry stops counting well before the two-year reporting window closes, so the dip is a quarter or so of your score history, not a permanent mark.
One thing the published guidance says that is easy to miss: several hard inquiries in a short window can compound. That is the case where this stops being a few points. One application that is approved is a small, finite event. Three applications in a week, two of them declined, is a different event and the guidance treats it as such.
What one transfer's utilization change does, computed
Here the arithmetic is not a model, it is division, so this part can be checked line by line. Start with $5,000 on a $10,000 limit: 50% utilization on one card. Move all of it to a new card with a $15,000 limit and keep the old card open at $0. The total limit is now $25,000 against the same $5,000, which is 20%. The per-card figure on the new card is 33.33%. The old card reads $0, which is the best utilization a card can show.
What that does to the score is not published as a point figure by anyone with a method, so this page does not attach one. What is published is that amounts owed carries 30% of the FICO weight, and that lower utilization generally means a higher score. The direction is well supported; the size is not, and the difference between those two things is the whole point of this page.
The other side of the trade, which is computed
| Route on the same $5,000 | Monthly | Months | Cost of interest |
|---|---|---|---|
| 24% card, $125 minimum | $125 | 82 | $5,159.49 |
| 24% card, paying $300 a month | $300 | 21 | $1,143.34 |
| 0% card, $5,000 cleared in 18 months | $277.78 | 18 | $150.00 fee, no interest |
| 0% card, $5,000 cleared in 21 months | $238.10 | 21 | $150.00 fee, no interest |
| 0% card, only the $125 minimum during the 21 months | $125 | 46 | $150.00 fee + $642.65 interest |
The last row is where the score question changes character — and it is also the one that keeps the balance from being an interest story at all. Paying only the minimum behind a 0% window does not clear $5,000 in 21 months; it clears it in 46, with $642.65 of interest charged after the window closes, on top of the $150.00 fee. That is still cheap next to $5,159.49, and it is not cheap next to zero. Staying on the minimum is not a score event, it is a money event: $5,159.49 of interest and 82 months. A missed promotional period is both — and by then the C-grade and D-grade pages are describing something real, because the balance is sitting at high utilization on two cards. The transfer did not cause that. The balance that never came down did.
Pricing one point against one month of interest
If a 5-point hit is the working assumption, the $5,009.49 net of the fee, set against the minimum-payment route, works out to about $1,001.90 of net saving per point. At a 20-point hit it is about $250.47 per point. That arithmetic is this site's, from the engine's numbers and the published inquiry figure; it is a way to size the trade, not a claim about what your score will do. The point of it is the ratio: on a balance this size, the money side moves in thousands and the score side moves in single digits.
Where the ratio flips is when the score is about to be used. A mortgage application next month prices one point very differently from a card you will keep for years, and nothing on this page can tell you which one you are in. That is a calendar question, not a model question.
What this page refuses to do
It refuses to publish a middle figure for a whole balance transfer. The tempting move is to say "about 10 points, sometimes 15", which reads better than a table of sources and grades. That number would be the only figure on the page with nothing behind it, and it would be the one readers repeat. The sources either give a tight number for one event (the inquiry) or a loose one for the whole thing with no method (the 20-to-110 pages). Bridging that gap with an average is not measuring, it is deciding what the answer should be and then writing it down.
What this page does instead is separate the two questions. The score side: a few points, published, graded, with the gaps left as gaps. The money side: computed, reproducible, and larger by two orders of magnitude on a $5,000 balance. Both belong in the decision; only one of them can be verified from here.
Reproduce the money side
Every money figure above comes from /api/payoff. Post {"debts": [{"name": "Card", "balance": 5000, "apr": 24, "minPayment": 125}], "extra": 0} and it returns 82 months and $5,159.49; change extra to 175 and it returns 21 months and $1,143.34, which is the $300-a-month row. The 0% rows are the same balance divided by the number of months, with a 3% fee of $150.00 added once: 5000 × 3 ÷ 100 = 150, then 5000 ÷ 18 = 277.78. The post-promotional row uses the same monthly loop at 24% with the payment fixed at $238.10, run until the balance is at or below a cent and rounded only at the end. No score model appears anywhere in this arithmetic, because none is reproducible from public information.
Frequently asked
How many points does a balance transfer drop your score?
There is no single published number. The one figure with a named source is the inquiry: FICO's own guidance says one additional hard inquiry lowers a FICO score by fewer than 5 points in most cases. The new account and the lower average account age are described by the same source as not usually huge, without a number. Any 20, 50 or 110-point claim comes from a page that does not say how it was measured.
Why do some sites say a balance transfer costs 20 to 110 points?
Because those pages print specific numbers without publishing the model, the sample or the method. The ranges disagree with each other, and the high end describes a different event: a missed promotional period and a balance that got spent again, not the transfer itself. A number with no source cannot be checked, so this page does not repeat one.
How long does the drop last?
The inquiry stays on the report for two years, and the sources that describe its effect put that effect at roughly twelve months — the score impact ends before the record does. The new-account and age effect fades as the account ages.
Is a few points worth it?
On this balance, the money side is two orders of magnitude larger. Minimums alone cost $5,159.49; an 18-month 0% route costs $150.00 in fees and no interest. That is $5,009.49 net of the fee, against a sub-5-point inquiry, or about $1,001.90 per point. The score cost is real — the ratio is what makes it small here, not the size of the hit.
Does applying for several cards in one week stack the points?
Yes, and that is the case where inquiries start to matter. FICO's guidance says several hard inquiries over a short period can compound. One approved application is a few points; three applications, two declined, is a different event with a different cost.
What is this page not doing?
It is not computing a score. No public model is detailed enough to reproduce, this site has no score data, and a made-up middle figure would be the least checkable number here. Every external figure carries a source and a grade; the only arithmetic is the money side, which comes from the payoff engine.