Propose exactly SIX different daily trading rules for a prospective public research experiment. Use only a single market's past and current daily closing prices. Each rule must work unchanged across stock-index, gold, oil, FX and cryptocurrency price series. Positions must be exactly 0 (cash) or 1 (long), with no leverage and no shorting. Fix every parameter now. There will be no tuning, backtest feedback, tool use, web search, file access or follow-up optimization.

The experiment starts each rule in cash. Historical prices may warm up indicators, but the position state starts in cash on the first observed season date. A signal computed after close t is executed at close t+1 and first earns the return from close t+1 to close t+2. If an exit and entry are both true, exit wins. Otherwise enter when entry is true, exit when exit is true, and keep the previous target when neither is true. Missing/insufficient indicator history means no action. You are proposing research rules, not recommending trades or claiming future profits.

Return ONLY one JSON object with this exact structure:
{"protocol":"quantarena-proposals-v2","strategies":[{"name":"Short name","idea":"One sentence rationale","entry":CONDITION,"exit":CONDITION}, ... exactly six items]}

CONDITION is either:
- {"op":"gt"|"gte"|"lt"|"lte"|"cross_above"|"cross_below", "left":OPERAND, "right":OPERAND}
- {"op":"all"|"any", "conditions":[CONDITION,...]}

OPERAND is either {"value":NUMBER}, {"indicator":"close","lag":INTEGER}, or {"indicator":NAME,"period":INTEGER,"lag":INTEGER}.
Allowed NAME values: sma, ema, rsi, roc, rolling_max, rolling_min, volatility.
Periods must be integers from 2 through 252; lag must be an integer from 0 through 5. Use no other operators, fields, indicators or executable code. Keep each entry/exit tree at most 20 condition nodes total and nesting at most 4 levels. Constants must be finite numbers.

Indicator definitions:
- sma: simple average of period closes; available after period observations.
- ema: recursive EMA with alpha=2/(period+1), initialized at the first available close; unavailable until period observations.
- rsi: Wilder RSI; first average gain/loss is the simple mean of the first period close differences, subsequent values use alpha=1/period; both averages zero gives RSI=50, only loss zero gives RSI=100.
- roc: close[t]/close[t-period]-1 (a fraction, not a percentage).
- rolling_max / rolling_min: maximum/minimum of the latest period closes, INCLUDING the current close before applying lag.
- volatility: sample standard deviation (ddof=1) of the latest period daily close-to-close simple returns, multiplied by sqrt(252).
- lag shifts the completed indicator by that many observations. A lag of 1 is useful for prior-window breakouts.
- cross_above: left[t] > right[t] AND left[t-1] <= right[t-1]; cross_below is the symmetric comparison. All needed current/prior operands must exist.

Do not use tools or request market data. Give all six proposals in one answer and do not report backtest performance. All proposals, including weak or duplicate ones, will be retained without selecting by performance.