Can You Still Allocate After the Rally? Three Managers Break Down Trend, Alpha and Performance-Chasing Risk
In four days in August, Bitcoin rose 21%. By the end of the month every directional strategy on 1ndex with a full month of data was up for August. A strong month is often the moment an allocator first looks at a strategy, and also the moment it is hardest to tell what the strategy actually does.
To see what the rally actually rewarded, 1ndex did two things. We split the NAV history of the strategies listed on the platform into the phases of the move. Then we put the same two questions to three managers who run very different books. What did your strategy actually capture? And what is an allocator really buying if they subscribe now?
The three managers sit at different points between market-neutral and directional. Positive Research runs a dollar-neutral, cross-sectional long-short book on perpetual futures. S1Quant, live since 2019, combines a multi-factor Bitcoin trend strategy with arbitrage. Algotoria, live since October 2023, runs a systematic trend-following CTA and keeps part of its collateral in Bitcoin. Each earns its return in a different way, which is the reason to read them side by side. An allocator's choice among them depends less on which did best in August than on what the rest of their portfolio already holds, and on how many distinct sources of return they want in it.
What the rally looked like on 1ndex
The move had four parts. Bitcoin ended June near $58,600 and spent seven weeks in a choppy range between $58,000 and $66,000. Between 18 and 21 August it broke from about $64,700 to above $78,000. It pushed above $80,000 in September, and then gave back about 7% between early October and early November.
Strategies on 1ndex responded very differently depending on what they are built to do. The table shows the median return of USD-denominated strategies in each 1ndex category over each phase. Each figure is the total return over that window, not an annualised rate, so a four-day column and a two-month column are not directly comparable.
| 1ndex category | 1 Jul – 17 Aug (7 weeks, range) | 18 – 21 Aug (4 days, breakout) | 1 Jul – 31 Aug (2 months) | October (1 month) |
|---|---|---|---|---|
| Delta Neutral | +1.0% | +0.1% | +1.6% | +0.7% |
| Market Neutral | +0.9% | +0.1% | +1.6% | +0.5% |
| Directional | −1.0% | +5.3% | +9.2% | −3.7% |
Delta-neutral and market-neutral strategies moved little with Bitcoin. The median one earned about 1% over the seven-week range, about 0.1% during the four-day breakout, and between 0.5% and 0.7% in October. All 15 delta-neutral strategies with full data were positive over July and August, and so were 15 of 19 market-neutral ones.
Directional strategies had a different two months. During the range the median directional strategy lost 1%, and fewer than half were positive, even though Bitcoin itself rose about 10%. Then came the four days. The median directional strategy made 5.3% between 18 and 21 August, and results ranged from a small loss to more than doubling. By the end of August, 13 of 14 were up for the two months, with a median gain of 9.2%.
October reversed part of it. The median directional strategy lost 3.7% and only 3 of 15 were positive, while the neutral categories continued to grind higher.
So the headline number an allocator saw at the end of August was real, but it came from a very small window. Most of the two-month return in directional strategies was earned in four trading days, after seven weeks in which the typical one lost money. A two-month figure for a directional strategy therefore says more about one week than about the strategy.
Three ways to make money in the same market
Algotoria: seven weeks of losses, then three days
Algotoria Diversified's NAV on 1ndex follows the shape of the category data closely. During the July range it lost 4.8% while Bitcoin rose 10%. The team calls this kind of market, where direction reverses daily with large swings, its hardest regime. By 16 August the strategy was in the deepest drawdown of its history, 31.2% below its peak, and had been under water for 175 days.
Then the trend arrived. Between 18 and 21 August the strategy returned 43.6% against Bitcoin's 21.4%, including 20.6% on 19 August alone. Algotoria is explicit that it did not see the move coming. Its systems built exposure on the breakout day itself, which, in the team's words, is how a reactive trend-follower should behave.
It also gave some back. In the first half of September the strategy fell 6.7% while Bitcoin fell 3.7%, and its NAV on 1ndex shows a further 10.6% decline in October. From 1 July to 31 October it was up 20.6%. Algotoria describes this, too, as normal: a trend system gives back part of a spike once the trend stalls.
S1Quant: trend exposure, added on confirmation
S1Quant's return also came mainly from Bitcoin's direction. Its NAV on 1ndex rose 8.8% during the July range, 20.2% in the breakout days, and another 9.9% from 22 August to the end of September. It peaked on 23 September and stood 8.3% below that peak on 24 October, the latest data point on the platform.
The team attributes about 90% of its return to its trend strategy and about 10% to arbitrage on funding rates and options term structure. It says its models began building long positions roughly four hours before the main August move, and that it does not rely on any single indicator. The strategy weights about 50 live factors, drawn from a library of over 250. Price and volume factors carry about 20% of model weight; the other 80% comes from alternative-data factors the team has built from data it has collected since 2016. Notional leverage is about 1.3x. The team describes the product as a Bitcoin-enhanced strategy: net exposure is kept within 0.3, and it says the strategy has outperformed Bitcoin by 30% to 40% a year on average.
S1Quant also makes a point that is easy to miss in a rally. It does not hold a fixed long position and does not try to capture a whole move. Exposure is added when signals confirm and cut when they weaken, so participation in any given rally is partial by design.
Positive Research: paid on both sides, in both directions
Positive Research's book is built to reduce its dependence on Bitcoin's direction. It forecasts which perpetual contracts will outperform and underperform each other, goes long the first group and short the second, and keeps net dollar exposure close to zero. Average end-of-day net exposure over the period was −0.13% of capital.
The team reports a net, non-compounded return of 42.1% from 1 July to 4 November, against Bitcoin's 32.1%. The split by phase shows the two halves of the book taking turns. While Bitcoin rose 42.6% between July and the start of October, the strategy made 31.2%, mostly from its longs. When Bitcoin then fell 7.3%, it made a further 10.9%, almost entirely from its shorts, which gained 16.3% while the long book lost 5.4%.
How much was skill, and how much was Bitcoin?
This is the question behind the first one, and each manager answers it in its own terms.
Algotoria can measure it, because it runs a twin. Algotoria Stable trades the same signals as Diversified but holds only USDT as collateral, so the difference between the two is the Bitcoin. Stable made 24.4% in the breakout. Scaled to Diversified's larger risk budget that is about 28%, which leaves 12 to 15 points of Diversified's gain explained by its Bitcoin collateral. The team's estimate is roughly two-thirds trend capture and one-third market beta. The gap persists in the platform data: from July to October, Diversified's NAV on 1ndex rose 20.6% and Stable's 0.3%. Over its full record Diversified has a daily correlation of 0.52 with Bitcoin. Stable's is 0.09.
S1Quant says it does not attribute all of the return from a rising Bitcoin price to alpha. Its argument is that the exposure is timed and bounded rather than held. The period it points to is the year before the rally: between 18 September 2025 and 21 October 2026, Bitcoin fell by about 27% by the team's measure, while S1Quant's NAV on 1ndex rose about 32% over the same dates.
Positive Research reports the smallest directional contribution. Of its 42.1% return, 1.1 points came from residual net exposure. In the most extreme 10-, 30- and 90-day windows for Bitcoin since March 2024, the strategy's hourly correlation with Bitcoin averaged between −0.04 and −0.15, in rallies and sell-offs alike. Against Ethereum it was slightly positive in sell-offs, between +0.07 and +0.10.
The team adds a warning that applies to every market-neutral claim in crypto. Equity managers and allocators share commercial risk models such as Barra. Crypto has no equivalent, so each manager defines its own risk factors, and "neutral" means whatever that manager constrains. Funding rates, liquidations and auto-deleveraging, mark prices and venue liquidity can also reinforce each other in a stress event in ways that factor limits estimated in calm markets do not capture. Allocators should ask which risks a manager actually limits, and whether those definitions fit the strategy.
None of this makes beta a bad thing. Algotoria holds Bitcoin collateral on purpose, to diversify away from stablecoin issuer risk and to add upside in bull markets. What matters for an allocator is knowing which part of a return was beta, because that part overlaps with Bitcoin they may already hold and behaves differently in a decline.
What the managers tell allocators who arrive after the rally
We asked each manager what an allocator is actually buying if they subscribe after the NAV has jumped, and what they should check first. Their answers reflect how each strategy is built, and they are worth reading as three descriptions rather than one recommendation.
Algotoria: most of the jump was recovery
Algotoria frames the question around what produced the recent return. An allocator looking at August's +35.4% is looking at a strategy that recovered its worst drawdown in three days. Its trailing 12-month return at the end of August was 17.3%, and its 6-month return was −1.2%. Most of the jump was recovery rather than new ground, and the team suggests judging an entry on the longer record instead.
What the allocator buys today is also not what produced August. By mid-September the trading book was lightly positioned, at about 0.2x exposure against a usual 0.8x to 1.0x, alongside a Bitcoin collateral position that moves with the market.
Its three checks follow from that. Ask what share of collateral is in Bitcoin today, because that explains about a third of the recent gain and most of the correlation. Look at the record over a year or more: 14 of the 32 months since January 2024 were negative, because the strategy loses small and often and wins large and rarely. And decide whether the drawdown budget fits. The High tier allows 35%, and the strategy used almost all of it this year. Lower tiers run the same signals at 25% and 15%. Algotoria recommends a horizon of at least one year.
S1Quant: a new account does not inherit old risk
S1Quant starts from the purpose of quantitative investing, which in its view is to reduce the role of timing. An allocator should decide on the strategy's long-run statistics, not on where the NAV stands this month. Its own are a historical win rate of about 74% and a ratio of total profits to total losses of about 6.7.
It also describes how new money is put to work. Without a fresh trend signal, a new account does not copy the directional exposure that older accounts have built up. Its risk budget starts small and grows only as the account accumulates its own profit cushion. The effect is that an allocator who arrives after a rally is not buying the positions that earned it.
The third check is whether risk rules are mechanical and followed. S1Quant begins cutting positions at a 3% loss on principal, goes to about half size at 5%, and stops out at 9%, with a target of keeping the maximum drawdown on principal within 10%.
Positive Research: past NAV does not carry forward
For a book that is rebuilt continuously, Positive Research argues that past NAV appreciation says little about whether an allocation is attractive today. New capital takes part in the opportunities available now, and in the team's continuing ability to research, build portfolios and execute.
It suggests three questions. Where did recent returns come from? Have current risk exposures, or the ones expected, changed? And are competition, crowding, trading costs and the manager's own assets compressing what is left to earn? Recent performance is useful evidence, the team says, but it should not be extrapolated into an expectation.
It also describes what can go wrong. The long-term risk is that realisable alpha decays as trades become crowded, so a long-term allocation is partly a judgement on the team's ability to keep producing research. The short-term risk is a market regime the models do not cover well.
Know the weak regime before you buy
The question after a rally has a mirror image: what to make of the next drawdown. All three managers draw the same line between a drawdown that is part of the strategy and one that means the thesis is broken, and each names its weak regime in advance.
For Algotoria the weak regime is a wide sideways market or a long stretch of low volatility. A drawdown that builds there, inside the agreed budget, is normal. In its history such drawdowns have usually begun to recover by the third or fourth month, and the recovery tends to come without warning, as it did on 19 August. The team's view is that a reassessment should be triggered by something else: losing money through a clean, sustained trend, which is the one regime the strategy exists to monetise, or a persistent gap between live and research results. Algotoria's Investment Committee reviews an account at two-thirds of its drawdown budget, and at the full budget the client decides whether to continue.
For S1Quant the hard markets are breakouts that reverse quickly and sharp rises followed by sharp falls. The models confirm, enter, and are then stopped out, and costs accumulate. Arbitrage income also shrinks when spreads narrow or become crowded. Losses of that kind, with exposure and stops behaving as designed, are the cost of the strategy. A reassessment is warranted if return sources decay over a long window, if costs run clearly above what the models assume, or if actual exposure and risk control drift from the rules.
For Positive Research, the difference between a normal drawdown and a thesis break is not the size of the loss but whether the logic that generates returns is intact. Signals can lose effectiveness for a period while overall signal quality and risk stay within expected ranges. A break looks different: predictive power that keeps deteriorating, risk exposures that drift systematically from plan, returns that depend more and more on common exposures the strategy was not meant to carry, or costs and capacity that stop alpha from reaching net returns. The decay of a single factor is not a break. The test is whether the research and trading system as a whole still produces realisable excess return. The team also cautions that market direction and trading volume alone do not tell you whether its environment is favourable. What matters is how many predictable relative-value opportunities exist across instruments, and whether they can be traded at acceptable cost.
Capacity belongs in the same conversation, because it decides whether a latecomer gets the same strategy as an early investor. Positive Research estimates capacity at $80 to $120 million across exchanges for its faster book and about $300 million for its slower one, based on the thin liquidity that followed the FTX collapse. It caps each position at 2% of the portfolio, 1% of 30-day average daily volume and 10% of open interest. Algotoria estimates about $150 million across OKX, Binance and Bybit and says it would close to new capital before reaching it. S1Quant gives no overall figure. It currently manages more than 150 sub-accounts and places orders across them in rotation, with time-weighted execution, which it says keeps the slippage difference between accounts within 0.2%. Because it runs a single strategy, every client's return curve is essentially the same. The $5 million it allows in each sub-account is a limit on a single account, not on the strategy. The team argues that trading the most liquid asset and splitting orders across time and venues keeps its market impact below one basis point, and that capacity remains large. All three say they are currently below capacity.
Allocator takeaway
The rally paid for three different things. Positive Research earned the spread between winners and losers across perpetuals, with little help from the market. S1Quant earned Bitcoin's trend, entered on confirmation and held under a defined loss limit. Algotoria earned a multi-day trend in three days, plus the return on the Bitcoin it holds as collateral. Each is a distinct source of return, and the three are far from interchangeable.
Which one fits depends on the portfolio it joins. An allocator who already holds Bitcoin or directional strategies gains the most diversification from a market-neutral book such as Positive Research's, where the return comes from relative value rather than direction. An allocator who wants Bitcoin exposure but with managed entry, exit and a defined maximum loss is looking at something closer to S1Quant. An allocator with a horizon of a year or more, able to sit through a deep drawdown, and comfortable that part of the return is Bitcoin beta, is looking at something closer to Algotoria Diversified. A portfolio can also hold more than one: trend and relative value tend to earn in different phases of the same market, as July to October showed.
The platform data and the managers suggest the same way of reading performance. A strong recent month is best read alongside the phases that produced it; in directional strategies much of it may come from a few days and from positions that no longer exist. A drawdown is best read against the regime the manager identified in advance as its weakest. And the managers agree on what should prompt a reassessment: a source of return that decays over a long window, costs or exposures that drift from what was described, risk rules not applied as written, or, for a trend strategy, losses in a clean trend.
Asked whether an allocator can still allocate after the rally, none of the three managers answered yes or no. Each described what its strategy is exposed to today and what to check. That leaves the decision where it belongs, with the allocator and the shape of their own portfolio.
About the data
Category figures are medians of USD-denominated strategies and funds listed on 1ndex whose NAV history covers each window, using the category shown on the platform. All returns in this article are simple returns over the stated window and are not annualised. Between 13 and 19 strategies per category had data for each window. NAVs are supplied by the managers through API connection or upload, and are not audited by 1Token. Bitcoin prices and returns are those quoted by the managers and differ slightly with their measurement dates.
S1Quant and Algotoria returns by phase are calculated from their NAV on 1ndex unless attributed to the manager. Algotoria's figures are gross of its success fee, and its split between trend capture and collateral beta is its own estimate. Positive Research's figures are self-reported, net of costs and non-compounded.
Positive Research's performance can also be checked through three channels beyond its own reporting. The team took part in the LTP Trading Competition (Track B – Liquidity Pro) from 9 September to 4 November 2026, where it says its return was among the highest of the long-short strategies entered. It has publicly accessible performance records through a Binance Portfolio Account going back to late September 2026. For institutional allocators who have signed an NDA, it can also provide read-only API access to verify recent live trading performance independently.
This article is for information only. It is not an offer, a solicitation or investment advice. Digital-asset trading involves substantial risk of loss, and past performance does not indicate future results.

Comments ()