Why monthly (and why quarterly): how rebalance cadence shapes the STI index family

Most CS2 price trackers update their data continuously and call themselves indices. They're not. An index is a fixed methodology applied to a defined universe at fixed intervals — and the interval matters more than people think.

The S&P 500 rebalances quarterly. The Dow Jones makes changes event-driven, sometimes years apart. The Russell 3000 reconstitutes annually. Each cadence implies a different theory of what the index is measuring, and a different trade-off between signal and noise.

The STI family uses two cadences: monthly for STI 100, STI 500, STI 1000, and STI Cases; quarterly for STI 30. This post is the why.

What rebalance actually does

A market-cap weighted index has three moving parts: the universe (which assets are eligible), the constituents (which subset actually appears in the index), and the weights (price × supply). Rebalance is the moment those three are recomputed from current data.

Between rebalances, the universe stays frozen — even if a skin rallies past the inclusion threshold or another skin drops out of liquidity. The index continues to track the basket as it was set at the last rebalance. The chart between rebalances reflects pure price movement of a fixed basket.

At rebalance, the basket is reset. Items that no longer satisfy the inclusion criteria (price floor, listings minimum, history length) drop out. Items that newly satisfy the criteria are added. Weights are recomputed. The divisor adjustment ensures the chart line has no break at the rebalance moment — the index value is preserved across the methodology change.

Why a price tracker without rebalance isn't an index

Three problems show up in a no-rebalance system:

• Survivorship inflation — an item that crashed 85% and stopped trading remains in the "index" at its stale price, dragging the total return up artificially. With rebalance, the item drops out of the universe and the next constituent takes its place.
• Composition drift — an item that 5×'d since the last refresh now dominates the basket. Without rebalance, the index becomes a bet on that one outlier. With rebalance, weights snap back to current market cap and the outlier's influence is capped.
• Missed structural shifts — when a new asset type emerges (Genesis Terminal cases in 2025, AI-generated skin patterns, etc.), a no-rebalance basket misses it entirely until manually adjusted. Rebalance picks up structural shifts automatically.

A live price feed across 30,000 skins is a price tracker. An index is a price tracker plus a defined methodology applied at defined intervals. The interval is the methodology.

Monthly vs quarterly — the trade-off

Cadence is a trade-off between two competing goals:

For a fast-evolving universe (mid-cap skins where new operations add cases every quarter, where Valve drops new items, where existing items get retired), monthly cadence captures the structural change without lagging too long. For a slow-evolving universe (top-30 blue chips that rarely change composition), quarterly is enough — and the lower turnover means the backtest is less affected by methodology noise.

Why STI 30 is quarterly

STI 30 includes the top 30 skins by market cap with at least 36 months of history and below-median volatility. By construction, the universe is the most stable part of the skin market: rare knives (Karambit Doppler, M9 Bayonet Fade), Major Souvenirs, ultra-rare consensus blue chips. These don't come and go. When they do change, it's usually because of a structural break — a wave of supply hits a previously locked tier, or a collection retires.

Quarterly rebalance matches the actual rate of change. A monthly rebalance for STI 30 would, in most months, just be the same basket re-priced — adding methodology noise without adding signal. The S&P 500 uses the same logic for its quarterly schedule: large-cap US equities don't reshuffle fast enough to need a monthly review.

Practically: STI 30 had roughly one constituent change per quarter across the 6-year backtest, sometimes zero. Monthly rebalance would have generated ~12 events per year for a universe that actually changes ~4 times per year — pure overhead.

Why STI 100, 500, 1000, Cases are monthly

The broader indices have universes that change much faster. STI 500 covers around 265 skins post-filter; the marginal constituent (rank 250-300 by market cap) cycles in and out of the basket multiple times per year as prices and listings shift. STI Cases tracks 40 containers; Valve removes a case from the drop pool roughly once every quarter, and existing case inventories get consumed faster than that.

For these universes, quarterly rebalance would lag the structural shift. A skin that rallies into the top 100 in May would not appear in the index until the next quarterly review in July — three months of return missing from the basket. With monthly cadence, the lag is ≤ 30 days.

STI 100/500/1000 each had 2 to 8 constituent changes per month over the backtest. Quarterly rebalance would batch these into 6 to 24 events per quarter, producing visible jumps in the chart at each review. Monthly cadence smooths the same changes across the year.

Worked example — what one rebalance looks like

Simplified 5-skin universe for clarity. Real STI rebalances run across hundreds of items, but the mechanic is the same.

Before rebalance (start of month):

Total market cap: $40,000. Index value at this snapshot: arbitrarily set to 100 at inception, divisor calibrated accordingly.

During the month:

• Skin A rallies from $50 to $80 (+60%). Listings drop slightly to 180.
• Skin D crashes — a Major reveal made it less rare. Price drops from $200 to $90, listings stay at 30.
• Skin F enters the eligible universe — a previously illiquid skin now has 80 listings at $90.

At rebalance (end of month):

Recompute market caps with current prices. Skin D's market cap fell from $6,000 to $2,700 — it's now too small for the 5-skin top tier. Skin F's market cap is $7,200 — it enters at rank 4. Weights snap to current values. The divisor is adjusted so the index value is continuous: the chart line at rebalance moment matches the value the moment before, even though the basket and weights changed.

From the next moment forward, the index tracks the new basket. Skin D no longer drags the index down. Skin F captures the new flow. Weights reflect current reality, not the start-of-month snapshot.

When rebalance hurts — turnover and look-ahead

Rebalance isn't free. Two costs are real even if the index itself is just a measurement:

Turnover cost (for anyone replicating)

A retail investor trying to replicate STI 500 would have to actually trade 2-8 items per month — and the friction tax covered in the liquidity penalty post applies to each trade. STI 30's quarterly cadence is partly an admission that anything more frequent would be impractical to replicate. SPDR's S&P 500 ETF (SPY) has the same trade-off internally — they batch rebalance trades to minimize market impact.

Look-ahead bias risk

More rebalance events mean more opportunities to data-mine the methodology. If we tweaked inclusion criteria after seeing a rebalance result, the backtest would be biased upward. Both PT and EN methodology pages document the inclusion rules explicitly so the reviewer can verify the methodology was fixed before the data, not adjusted after.

Methodology stability over the backtest

We made two methodology corrections during the 2026-04 audit that materially affected STI 30's historical return — one (commit 2940ad0) added missing ultra-rare blue chips, another (commit a2f4f84) excluded skins above the Steam Market $1,800 cap. Both are documented in the public audit (data-quality-2026-04-30.md). Neither was a backtest tune; both fixed real universe-coverage issues. STI 30's published return moved from +78.9% to +37.5% as a result — and the disclosure stays public to make this honest.

The honest take

Cadence is a methodology choice, and methodology choices always have trade-offs. Monthly cadence catches structural shifts faster but introduces more rebalance noise. Quarterly cadence smooths the chart but lags the universe. The STI family uses both because the underlying universes have different rates of change.

What rebalance does not do: predict the future. Past rebalances reflect past decisions on the inclusion rules. The next rebalance will reflect those same rules applied to whatever the market looks like in 30 days. If the rules are wrong, the index is wrong — and the only defense is to publish the rules clearly enough that a reviewer can spot the error and challenge it.

Same disclaimer as elsewhere: this is a description of how the STI is constructed, not a forecast of how it will perform. The 2020-2026 backtest is one specific window. The 2026-2032 window will look different — possibly very different — and the methodology survives only if the rules survive the test.

Bottom line

A price tracker becomes an index when it has fixed inclusion rules applied at fixed intervals. The interval is the methodology. STI 30's quarterly cadence and STI 100/500/1000/Cases monthly cadence reflect different rates of change in their underlying universes, not arbitrary choices.

The S&P 500 has been doing this since 1957. The Dow Jones since 1928. The STI is the same logic applied to CS2 skins, with the same trade-offs and the same honest disclosures.

Methodology is a choice. Cadence is part of the methodology. The chart you see is the consequence.

Fernandes, J. (2026). Why monthly (and why quarterly): how rebalance cadence shapes the STI index family. Skin Trackers. https://skintrackers.com/en/blog/rebalance-cadence-skin-index

Jorgin Fernandes, "Why monthly (and why quarterly): how rebalance cadence shapes the STI index family," Skin Trackers, 2026, https://skintrackers.com/en/blog/rebalance-cadence-skin-index.

@misc{skintrackers_rebalance-cadence-skin-index,
  author = {Jorgin Fernandes},
  title  = {Why monthly (and why quarterly): how rebalance cadence shapes the STI index family},
  year   = {2026},
  url    = {https://skintrackers.com/en/blog/rebalance-cadence-skin-index},
  publisher = {Skin Trackers}
}

https://skintrackers.com/en/blog/rebalance-cadence-skin-index

Disclosure: The author maintains a personal position in CS2 skins. No conflict of interest with any marketplace. Methodology choices documented at /en/methodology and the data API at /api/indices. Analysis is editorial, not buy or sell recommendation.