{"url":"https://etfiq.com/methodology/scores","section":"methodology","title":"How the ETFIQ scores are computed | ETFIQ","description":"How ETFIQ works out its 7 category scores: the question each answers, the funds it compares and how the percentile is worked out.","as_of":"Oct 27, 2026","structured_data":[{"@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"ETFIQ","item":"https://etfiq.com/"},{"@type":"ListItem","position":2,"name":"Methodology","item":"https://etfiq.com/methodology/"},{"@type":"ListItem","position":3,"name":"The scores","item":"https://etfiq.com/methodology/scores"}]},{"@type":"WebPage","name":"How the ETFIQ scores are computed","description":"How ETFIQ works out its 7 category scores: the question each answers, the funds it compares and how the percentile is worked out.","url":"https://etfiq.com/methodology/scores","dateModified":"2026-10-27","isPartOf":{"@type":"WebSite","name":"ETFIQ","url":"https://etfiq.com"},"speakable":{"@type":"SpeakableSpecification","cssSelector":[".lede"]},"publisher":{"@id":"https://etfiq.com/#org"}}],"characters":16039,"truncated":false,"content":"ETFIQ / Methodology / The scores \n On this page\n The method, in full Why a percentile and not a rating The rules that apply to all seven The seven What is not computed Where to next \n Recomputed every trading night. 1637 funds scored across 7 categories, as of Oct 27, 2026 .\n How the ETFIQ scores are computed\n One score per category. Each is a fund\u2019s percentile on one published number, among the funds that can fairly be compared on it.\n 1,637 Funds scored \n 7 Categories scored \n Nightly Rebuilt from the raw sources \n\n The method, in full\n For a fund with value v in a set of n values, the score is\n score = 100 * (count(x <= v) - count(x == v) / 2) / n A score is a rank turned into a percentile, so no fund scores 100 or 0. The top of a set of n is 100 * (n - 0.5) / n, which is 96.2 for the smallest set here, 13 funds. The whole file is published , so anyone can recompute it.\n Scores are rounded to one decimal. Tied funds get the same score. 22 income funds sit within 0.1 of another on coverage, and ranking them apart would invent a difference. Where a score blends two numbers, each is scored this way first, and the two are averaged at the stated weights. The weights are the only judgment in the method. They are printed beside every score they affect.\n Why a percentile and not a rating\n It has no weights to argue with A percentile is a count. Six of the seven scores have no weights. Each ranks one published number. It survives the outliers a linear scale cannot Income coverage runs from \u2212377% to 2,200%. A linear 0 to 100 scale would put nine tenths of the category within 15 points of each other. It can be recomputed by anyone The number, the set and the method are published. The file carries every score, component and weight. It moves when the set moves A fund can slip two points on a day it did nothing, because two other funds improved. That is how a percentile works. So the underlying number is printed beside every score. The rules that apply to all seven\n One score per category Each score ranks a fund only against similar funds in its own category. A minimum set of 12 Below twelve funds, a percentile says more about the set than the fund. A smaller set is not ranked, and its funds get no score. The file names each one, with its set and why. A high score is not a recommendation A low score is not a warning. Each score page says what its top and bottom mean. For two of these scales, neither end is the good end. A fund with no data has no score It gets no zero and no median. It is left out of the set, and the set size beside every score counts only the funds in it. The score never enters another number Nothing on this site is ranked, filtered or weighted by a score. A score comes out of the data and never goes back in. The seven\n Downside Cover Score, for buffer ETFs\n The question: If the market falls into a bear market from here, how much does the buffer absorb?\n The limit: Today's position against one 20% fall; it ignores smaller falls, and stock buffers are never ranked against bond or commodity ones.\n The set: 292 buffer ETFs on stocks with a live outcome period (scores 0.2 to 99.8). A fund outside that set has no score.\n Not scored: 2 funds with the number, in sets under twelve: 1 buffer ETFs on a commodity with a live outcome period and 1 buffer ETFs on bonds with a live outcome period.\n What it is computed from: \n share of a 20% fall the buffer absorbs \n What it does not say: It is measured over a 20% fall, the conventional bear market line, so it says nothing about a 5% one, where a buffer that has not yet engaged absorbs nothing. Buffers on bonds or a commodity are ranked only among their own kind, and only where there are twelve. The upside left is printed beside it rather than blended into it.\n Every fund in this category, ranked by it \u00b7 Method \n Distribution Coverage Score, for income ETFs\n It replaced the Return Stability Score on Oct 9, 2026, which blended the gap to the benchmark with the price change.\n The question: How much of what it paid out did its total return cover?\n The limit: One year only, and a ratio of two published numbers: it says nothing about where a distribution came from.\n The sets: 165 income ETFs not writing on a single company (scores 0.3 to 99.7), and 74 income ETFs writing on a single company (scores 0.7 to 99.3), ranked separately. A fund outside both has no score.\n Not scored: 17 funds with the number, in 17 with no one-year number against their benchmark.\n What it is computed from: \n total return as a share of the cash paid \n What it does not say: It is the year's total return as a share of the cash paid, so under 100% means the fund paid out more than it returned in total. It does not say where a distribution came from: only a Rule 19a-1 notice or the tax reporting says that. Funds writing on a single company are ranked among themselves.\n Every fund in this category, ranked by it \u00b7 Method \n S&P 500 Difference Score, for thematic ETFs\n Called the Off-Index Score until Oct 9, 2026.\n The question: How much of its weight does it not share with the S&P 500?\n The limit: Weight not shared with the S&P 500 only; it says nothing about returns, fees or whether the theme pays off.\n The set: 112 thematic ETFs with at least half their money in S&P 500 companies (scores 0.4 to 99.6). A fund outside that set has no score.\n Not scored: 277 funds with the number, in 277 with less than half their money in S&P 500 companies.\n What it is computed from: \n active share against the S&P 500 \n What it does not say: 100 minus weight overlap is active share, which is Cremers and Petajisto and not an ETFIQ invention; what is ETFIQ\u2019s is the percentile. Only funds with at least half their money in S&P 500 companies are ranked, because against the S&P 500 a small-cap or international fund scores high by construction. A high score says nothing about fees, returns or how many stocks the fund holds.\n Every fund in this category, ranked by it \u00b7 Method \n 3-Month Tracking Score, for leveraged ETFs\n Called the Decay Resistance Score until Oct 9, 2026.\n The question: Did it keep up with its own daily multiple, compounded day by day, over three months?\n The limit: Three months only, against the fund's own daily multiple; a different quarter can rank it very differently.\n The sets: 507 leveraged ETFs at 2x, over three months (scores 0.1 to 99.9); 75 inverse ETFs at 2x, over three months (scores 0.7 to 99.3); 35 inverse ETFs at 1x, over three months (scores 1.4 to 98.6); 31 leveraged ETFs at 3x, over three months (scores 1.6 to 98.4); 19 inverse ETFs at 3x, over three months (scores 2.6 to 97.4). They are ranked separately. A fund outside all of them has no score.\n Not scored: 5 funds with the number, in sets under twelve: 3 calendar-reset ETFs, over each reset period, 1 inverse ETFs at 1.5x, over three months and 1 leveraged ETFs at 1.25x, over three months.\n What it is computed from: \n points against its daily multiple, compounded \n What it does not say: It measures a fund against its own promise, a multiple of each daily move compounded day by day, over three months. What is left is the fee, the cost or income of financing, the slippage of rebalancing daily, the fund\u2019s price against its NAV on the two end days and, where an index is read through a fund that tracks it, that fund\u2019s own difference from the index. Each stated multiple is its own set, long and inverse apart, because a 3x fund decays faster than a 2x one for reasons that are not about how it is run.\n Every fund in this category, ranked by it \u00b7 Method \n Coin Capture Score, for crypto ETFs\n The question: How much of what holding the coin gave did a holder of the fund keep?\n The limit: Share of the coin's own price growth a holder kept; funds on different coins are ranked together.\n The sets: 28 spot crypto trusts listed less than a year (scores 1.8 to 98.2), and 23 spot crypto trusts with a full year (scores 2.2 to 97.8), ranked separately. A fund outside both has no score.\n Not scored: 17 funds with the number, in sets under twelve: 8 crypto futures and strategy ETFs with a full year, 6 crypto futures and strategy ETFs listed less than a year and 3 1940 Act crypto ETFs with a full year.\n What it is computed from: \n share of the coin's growth it kept \n What it does not say: These funds hold the same kind of thing, so the score is close to a fee ranking with the market\u2019s pricing of the shares in it. It says nothing about the coin. Trusts and 1940 Act funds are ranked apart, and so are funds listed less than a year, because a tracking gap compounds.\n Every fund in this category, ranked by it \u00b7 Method \n Stock Diversification Score, for alternatives ETFs\n Called the Diversifier Score until Oct 9, 2026.\n The question: How much does it diversify a stock portfolio?\n The limit: One year of weekly moves against the S&P 500, within one strategy; it ignores what the fund returned.\n The sets: 22 equity hedged alternatives ETFs with a full year (scores 8.0 to 97.7), and 17 systematic trend alternatives ETFs with a full year (scores 2.9 to 97.1), ranked separately. A fund outside both has no score.\n Not scored: 23 funds with the number, in sets under twelve: 9 long-short equity alternatives ETFs with a full year, 7 multistrategy alternatives ETFs with a full year, 3 event driven alternatives ETFs with a full year, 2 equity market neutral alternatives ETFs with a full year and 2 macro trading alternatives ETFs with a full year.\n What it is computed from: \n correlation with the S&P 500 , weighted 50%\n down-week capture against the S&P 500 , weighted 50%\n What it does not say: It is one year of weekly returns, and a fund that rose when stocks fell in that year need not in the next. Each strategy is ranked only among itself, and a strategy with fewer than twelve funds has no score. It leaves out what the fund earned, which is why return against T-bills sits beside it.\n Every fund in this category, ranked by it \u00b7 Method \n Rate Cushion Score, for bond ETFs\n The question: How far can rates rise in a year before the year's income is gone?\n The limit: One year's income against today's duration; it ignores credit risk, and both numbers move.\n The sets: 42 high-yield, bank-loan and multisector bond ETFs with a short duration (scores 1.2 to 98.8); 32 investment-grade bond ETFs with a short duration (scores 1.6 to 98.4); 32 corporate bond ETFs with an intermediate duration (scores 1.6 to 98.4); 24 investment-grade bond ETFs with an intermediate duration (scores 2.1 to 97.9); 23 national municipal bond ETFs with an intermediate duration (scores 4.3 to 97.8); 21 international bond ETFs with an intermediate duration (scores 2.4 to 97.6); 20 government bond ETFs with an intermediate duration (scores 2.5 to 97.5); 16 national municipal bond ETFs with a short duration (scores 3.1 to 96.9); 14 national municipal bond ETFs with a long duration (scores 3.6 to 96.4); 13 government bond ETFs with a long duration (scores 3.8 to 96.2). They are ranked separately. A fund outside all of them has no score.\n What it is computed from: \n rate cushion \n What it does not say: It counts a year of income against today's duration, and a fund's income and duration both move: a cushion is where a fund stood on the dates beside it, not a promise about next year. It says nothing about credit: a high-yield fund's wide cushion is paid for with default risk, which is why each holdings group is ranked only among itself. Capital-gain cash is not income and is left out, and a fund whose issuer does not separate the two is left out by name rather than counted whole. Where an issuer has not yet classified this year's payments and prints them as income, they are counted as printed and the score says provisional until year-end.\n The two numbers: income paid over the 365 days to the date of the fund's NAV, divided by that NAV; and duration, the issuer's own number when it is the freshest dated, else the duration in its latest N-PORT. The computation is the same for every issuer. Capital-gain distributions are not income and are not counted. A fund whose issuer does not separate them is left out. If an issuer has not yet classified this year's payments but prints each one's parts, the parts are counted as printed and the score says provisional until year-end. A fund whose issuer lists less than the year of payments is left out. The cushion is the yield divided by the duration, in percentage points. It shows how far rates can rise in a year before the year's income is gone.\n The sets: each fund is ranked among bond funds in its own holdings group and duration band, as the Bonds category places it. A fund whose latest N-PORT holds bonds worth more than 105% of net assets is ranked the same way, and says how much beside its score.\n The bond funds it does not rank, and why: \n 152 its duration band is ultrashort, under a quarter of the core bond market's, where the ratio measures the duration rather than the fund\n 108 its issuer publishes no split of income from capital gains that ETFIQ reads, and the cushion counts income only\n 48 it has traded for less than the year a 12-month yield needs\n 44 its N-PORT gives no duration, so it has no maturity bucket (preferred stock and perpetual debt have no maturity)\n 40 its issuer refuses ETFIQ's requests, so its split of income from capital gains cannot be read\n 38 its issuer's page could not be read tonight\n 30 the issuer lists too few distributions to cover the year\n 29 one of its distributions in the year is not split into income and capital gains on its issuer's page\n 22 its issuer's distributions could not be read, so the year's income cannot be summed\n 11 its issuer's NAV could not be read\n 10 its issuer shows its numbers only to a visitor who accepts its terms, which ETFIQ has not done\n 3 its duration is under a quarter of the core bond market's, the ultrashort line, where the ratio measures the duration rather than the fund\n 2 it has no Bonds segment, so it has no holdings group\n 60 in 13 sets with fewer than 12 funds: corporate bond ETFs with a long duration (2); government bond ETFs with a short duration (11); high-yield, bank-loan and multisector bond ETFs with an intermediate duration (6); inflation-protected bond ETFs with a long duration (1); inflation-protected bond ETFs with a short duration (6); inflation-protected bond ETFs with an intermediate duration (9); international bond ETFs with a long duration (3); international bond ETFs with a short duration (3); investment-grade bond ETFs with a long duration (7); securitized bond ETFs with a short duration (1); securitized bond ETFs with an intermediate duration (1); single-state municipal bond ETFs with a long duration (2); single-state municipal bond ETFs with an intermediate duration (8)\n Every fund in this category, ranked by it \u00b7 Method \n What is not computed\n ETFIQ does not forecast what a fund will do or say whether it is worth owning. Each score ranks funds on one published number, which prints beside it.\n Where to next\n Downside Cover Score Distribution Coverage Score S&P 500 Difference Score 3-Month Tracking Score Coin Capture Score Stock Diversification Score Rate Cushion Score All seven The file Every method \n Cite this page. ETFIQ, How the ETFIQ scores are computed, data as of Oct 27, 2026. https://etfiq.com/methodology/scores Free to use with attribution for data ETFIQ computes; issuer data stays under its owner's terms. The underlying files are at Open data .\n Other forms Plain ETFIQ, How the ETFIQ scores are computed, data as of Oct 27, 2026. https://etfiq.com/methodology/scores APA ETFIQ. (Oct 27, 2026). How the ETFIQ scores are computed. Retrieved from https://etfiq.com/methodology/scores Markdown [How the ETFIQ scores are computed (ETFIQ, Oct 27, 2026)](https://etfiq.com/methodology/scores)","source":"ETFIQ, independent ETF data. Recomputed nightly from public filings and prices.","publisher":{"@id":"https://etfiq.com/#org","name":"ETFIQ","url":"https://etfiq.com"},"license":"https://etfiq.com/license/","license_text":"Data ETFIQ computes is free to use with attribution: ETFIQ (etfiq.com). Issuer data stays under its owner's terms.","citation":"ETFIQ, How the ETFIQ scores are computed, data as of Oct 27, 2026. https://etfiq.com/methodology/scores"}