A few weeks in, a lot of backtests later. The Arena has now run 7,299 backtests across 601 assets, with more than 6,000 different configurations. Enough to start seeing patterns — and enough to bump up against patterns we hadn't seen quite that way before.

We set the Strategy Insights filter to "Only B&H beaters" this morning, meaning only strategies that beat the Average Buy & Hold benchmark show up. The picture that comes out is not what an average trading guide would promise. When you look at a table of strategies × timeframes and ask which cells are green and which are red, the colours alone tell you a fairly clear story.

The Big Finding: Daily Is Whipsaw Hell

Look at the Daily column and it glows red. RSI/SMA Cross, Golden Cross, RSI Overbought/Oversold — all negative. Move one column to the right onto 2D or 3D, and the picture turns greener. Weekly is friendly. Monthly is, for many strategies, distinctly friendly.

This isn't a strategy problem. It's a noise problem.

Daily charts produce a constant stream of small movements that look like signals but aren't. An SMA-cross strategy walks into every one of these mini-trends, pays the fees, and waits for the next whipsaw. On weekly charts, those mini-moves average out — what remains are the real trend changes. This isn't a new insight from the trading literature, but it's one that shows up so brutally in our own data that we want to highlight it explicitly.

Concretely: RSI/SMA Cross loses an average 21% CAGR on Daily. On Weekly, only 4.5%. Same strategy, same asset universe, same logic. Just a different observation rhythm. Same pattern with Golden Cross: -13% Daily, -5% Weekly, even +1% Monthly. With RSI Overbought/Oversold: -17% Daily, +3% Weekly. It's not that these strategies suddenly print money on Weekly — most of them are still no money machine. But they stop being a systematic loss. That's the difference between "this strategy doesn't work" and "this strategy doesn't work on this timeframe."

Why That Matters More Than Strategy Choice

Most traders coming into our platform ask first: which strategy is the best? That's the wrong question. The right one is: which strategy works on which timeframe — and does that timeframe match how I actually want to trade?

Saying "I want to trade daily" implicitly chooses a sub-class of strategies that have any chance on daily at all. Those are not the classics. Saying "I want to trade weekly" gives you a much wider strategy universe to choose from, because the noise is averaged out.

That's exactly why the most important filter setting on the Strategy Insights page isn't asset type or strategy — it's timeframe. We didn't make that as explicit on the page as it deserved. If you've only ever looked at the page with default settings, you may have missed the effect. Go back, click through the timeframe columns — the change is dramatic.

Daily Trading Needs Different Tools

There are strategies that work on daily — they're just not the classics. The Daily winners in our data:

What ties them together: none of them reacts directly to the daily price movement of the traded asset. F&G Cadence uses sentiment, BTC Signal uses a market leader, EMA and WMA smooth so heavily that daily resolution effectively becomes weekly. OBV-MACD pulls volume into the calculation, adding a dimension that pure price-cross indicators don't have.

The lesson: if you want to trade on daily, don't take the daily price itself as your main signal. Take something that filters out daily noise — sentiment, cross-market correlation, volume indicators, or aggressive smoothing. That's a different strategy family than what most online tutorials teach.

There's a second, less obvious finding in the same column: the Daily losers tend to have high run counts (RSI/SMA Cross has 373 runs on Daily, Golden Cross 163, RSI OB/OS 270). Meaning: lots of traders have actually tried these strategies on daily. The Daily winners consistently have low run counts (F&G Cadence: 6 runs, OBV-MACD: 9 runs, BTC Signal: 50 runs). That's information too: the strategies that work on daily are the ones that get touched least often. Maybe because they're less intuitive, maybe because they're less hyped, maybe both.

What Paid Indicators Are Promising

A short side glance that became interesting to us while building.

Two of our best Daily performers in the data — EMA Trend Bias and WMA Trend Signal — belong to the same indicator family as two popular paid indicators in the crypto scene: the Larsson Line by CTO Larsson, and the MoneyLine from Ivan on Tech's Bullmania community. Both are invite-only on TradingView and accessible only to paying members.

We don't want to badmouth pricing or claim our indicators are "better." But three observations from the data and from publicly available information are clear enough to write down:

1. The underlying maths is public. Larsson himself describes his indicator as a system of four smoothed moving averages (SMMA) on hl2 with periods 15/19/25/29 — the logic is documented online, open-source implementations are publicly findable on GitHub and in the TradingView Public Library. The MoneyLine isn't openly documented, but per its description belongs to the family of trended moving averages. Neither is secret science — both are variations on a known indicator type.

2. Similar logic performs similarly. Our EMA Trend Bias and WMA Trend Signal aren't direct clones — they're independent implementations using different smoothing methods (EMA and WMA instead of SMMA) and configurable periods. But they belong to the same family: moving averages of various lengths, cross logic for entry/exit, with a smoothing component to reduce whipsaws. On Daily they deliver +10.6% and +11.3% CAGR on average across all B&H-beating runs. That's the same league the paid indicators claim to play in within their own marketing charts.

3. A trading strategy is not the same as an indicator. What the paid communities sell isn't just the line on the chart — it's the surrounding package: Discord, livestreams, education, coaches, alpha streams. That's a different product than what we make. We don't sell coaching, no Discord, no private calls. We sell the ability to test the underlying strategies yourself, with your parameters, on your pairs, over time periods up to 25 years. If you want the package, the communities are the right address. If you want to know whether the underlying indicator logic actually makes money, you're at the right place here — and the answer is: often yes, sometimes no, depends on the pair and the timeframe.

What we noticed while building: many indicators that get passed around in the scene as insider tips are simpler in their core mechanics than the marketing suggests. A Smoothed Moving Average is a variation on the Exponential Moving Average. A "trend color switch" is, at its core, cross detection between two lines. That's legitimate trading logic, but it's not rocket-grade maths. The question of whether such indicators are profitable on your pair and timeframe is not a question of belief — it's a question of backtest.

Two Newcomers in the Spotlight

Two strategies stand out — both still young in the Arena, but with striking early results.

Keltner Channel Breakout went live a few days ago. ATR-based volatility bands, trend-following logic, we covered it in detail last week. On 2D and 3D it currently sits ahead of every other strategy — +31.9% and +14.6% CAGR with scores of 80 and 55. Mid-line exit, EMA(20), ATR(10), multiplier 2.0. On weekly it's still solidly in the green with the highest score (61) in that column. On monthly it falls back to neutral, which fits the logic: Keltner is a volatility-breakout system, and on monthly charts there are simply too few volatility movements to trigger breakout signals. It needs the mid-range timeframes to breathe.

BTC Signal — RSI/SMA is the strategy with the most unusual logic in the roster: the signal comes from BTC, the trade runs on an altcoin. The idea: BTC leads the market, and when BTC makes an RSI/SMA crossover, alts follow with a lag. On weekly it delivers 35% CAGR with a score of 90 — the highest single value in the entire table. On 3D and monthly it's also clearly positive (+22% and +27% CAGR). That's consistently positive across multiple timeframes, which raises confidence in the pattern.

What both strategies share is that they don't query the asset they trade on directly. Keltner converts price into volatility. BTC Signal takes a different market as the indicator. Maybe that's the real common thread running through the data: strategies that take one step back from the direct price are currently outperforming those that hug the price.

That's a hypothesis, not a hard conclusion. We want to keep watching it as the data thickens. It would explain why F&G Cadence (sentiment instead of price) does so well on daily, why EMA/WMA Trend (smoothed price representation) still works on daily, and why the classics that react directly to price crosses get ground down on daily. If the hypothesis holds, it would also have a practical consequence: in a world where everyone sees the same price data in real time, the "edge" lives more in how you transform the data than in the data itself.

What "Insufficient Data" Cells Mean

The table contains many "Insufficient data" cells. That's not "we have no data for this" — it's "this combination doesn't beat Average B&H or doesn't have enough runs to show up in the B&H-beaters view." That's information.

Example: Golden Cross on 3D, empty. Meaning: the strategy doesn't reliably beat buy-and-hold on 3D crypto charts. Not surprising — Golden Cross is a very slow strategy that on a short timeframe like 3D virtually never produces enough trades to outperform B&H. Example: OBV-MACD on 3D and Monthly, empty. A volume indicator on long crypto timeframes behaves differently than on mid timeframes — possibly the volume signal gets dominated by the price trend over longer windows.

We could have shown these cells by default (with red numbers). But we picked the B&H-beaters filter because most people landing on the Insights page don't want to know "where does which strategy lose how much" but "where does which strategy actually win." An empty cell then implicitly says: not here.

A Note on Sample Size

Some of the greenest numbers in the table sit next to very few runs. Fear & Greed Cadence Daily at 39.6% is 6 runs on 6 assets — that's anecdote, not evidence. OBV-MACD (v2) on Daily at +23.7% is 9 runs. Stoch-RSI / SMA Cross on 2D at +5.9% is 19 runs. All interesting hints, none of them arguments.

Our house rule: under 30 runs a number is interesting but not an argument. Above 100 runs it becomes solid. Above 500 you can take it seriously.

By that rule, the truly solid cells right now are: RSI/SMA Cross on Weekly (1647 runs), DCA Daily (455 runs), Buy & Hold Monthly (476 runs), and a handful of others. Those are exactly the cells where the classics are getting ground down. The newcomer cells — Keltner with 51-151 runs, BTC Signal with 49-51 runs — sit in the "interesting but too early" bracket. We're calling that out explicitly here because we like the newcomers but don't want anyone moving money on this data that they can't afford to lose.

A Word on the B&H Benchmark Itself

Something easily overlooked in the table: the Buy & Hold (Fixed) cell on Monthly sits at -9.7% CAGR with 476 runs across 157 assets. And Average B&H — the actual benchmark the strategies are measured against — shows up in the "Ø B&H" numbers under the CAGR values, often in the negative range (-22.5%, -26.7%, etc.).

Meaning: the sample these 7,299 runs are based on has a substantial bear-market component. That's an inherent property of an asset universe of 601 coins/stocks/ETFs/etc. that's very broad and includes lots of long-tail assets sitting deep below their all-time highs. The "B&H beaters" view therefore effectively also filters for strategies that can step out of the market in bear phases — which is a different property than pure outperformance in bull phases.

What This Means for You

If you're stepping into the Arena and don't know where to start, here are the heuristics we'd derive from this data:

• Start on weekly. The data clearly shows that's the timeframe where most strategies can breathe. Drop down to 2D/3D only after you've found a strategy you understand and that works on weekly. Daily is a trap as an entry-level timeframe.
• If you want to trade daily, look at the strategies that don't hang on the daily price directly — F&G Cadence, BTC Signal, EMA/WMA Trend, OBV-MACD. The classics are loss-makers here.
• For newcomers like Keltner and BTC Signal: try them on your favourite pairs. The results are promising, but the data is young.
• Before paying for an indicator, test the indicator family here. EMA Trend Bias and WMA Trend Signal deliver Daily returns in our data similar to those advertised by paid indicators of the same family. Your money stays in your pocket, your backtests stay in your account.
• Compare multiple timeframes for the same strategy, instead of multiple strategies on the same timeframe. The spread across timeframes in our data is bigger than the spread across strategies.
• Don't get impressed by small samples. A 40%-CAGR cell with 6 runs is pretty, but too early to bet on. Look for cells that are both green and densely populated.

The insights get more honest with every run. 7,000 is a start. We'll publish updates as the data thickens and patterns either solidify or shift.

---

Look for yourself:

→ tradingstrategies.work/dashboard/strategy-insights

Backtesting Arena · Your own backtests, your own parameters — across 2,000+ assets in every major market.