Cycle 01

第一轮

29 May – 31 Jul 2026 · closed · paper account

2026 年 5 月 29 日 – 7 月 31 日 · 已了结 · 模拟盘

Put in
投入: $18,734
Took out
取回: $19,774
Made
净赚: +$1,040 +5.55%
Held for
持有: 63 days63 天
S&P 500, same days
同期标普 500: −1.00%

Cumulative profit and loss in dollars. Both lines start at $0 on the close before I entered and run to the close I exited on. Every point is the mark the account itself recorded that day, plotted as recorded, including 29 Jul, where the account repeated the previous session's figure to the cent and the line runs flat through a day the positions moved. The ledger note below says what that costs the statistics. SPY's line is the same sessions. The shaded stretch is the worst run the book had, $765 down from a high on 26 Jun to 8 Jul. It sits above the zero line because the fall started from a profit, not because it was painless. 累计盈亏（美元）。两条线都以建仓前一个交易日收盘为 $0 起点，画到清仓当天收盘。每个点都是账户当天自己记下的净值，原样画上去，包括 7 月 29 日：那天账户把前一交易日的数字一分不差地重复了一遍，线在那里是平的，而持仓其实在动。这对下面那些统计量的影响，写在成交明细后面的那段小字里。SPY 那条线取自同样的交易日。阴影区间是这本账最难看的一段：从 6 月 26 日的高点一路跌到 7 月 8 日，回撤 $765。它在零线之上，只是因为这一跌是从盈利里跌下来的，不代表它不难受。

What it cost to get there

这些收益的代价

Deepest drawdown
最大回撤: −$765 −4.08%
That hole ran
回撤区间: 26 Jun → 8 Jul6 月 26 – 7 月 8 日
Underwater from that high
跌破高点后未收复: 20 days, back above on 16 Jul20 天，7 月 16 日收复
Worst day
最差单日: −$399
Best day
最佳单日: +$581
Days up / down / flat
上涨 / 下跌 / 持平天数: 24 / 19 / 1
Volatility
波动率: 19.4% S&P 500: 14.8%标普 500：14.8%
Sharpe
夏普比率: 1.63
Beta vs S&P 500
相对标普 500 的贝塔: 0.13

All of these are measured on the capital actually deployed, from 44 daily returns, with Sharpe assuming a zero risk-free rate. Forty-four is a very short sample and the ratios inherit that: Sharpe's 1.63 carries a confidence interval running from below zero to above six, and beta's 0.13 one that comfortably contains 0.5. Read them as a description of this cycle, not as numbers that stand on their own. The one thing I will claim outright is descriptive rather than statistical: the book's daily moves tracked the market about as weakly as they could, correlation 0.10, with the index explaining roughly 1% of the daily variation. The +5.55% was not the market carrying me. Drawdown is shown against capital deployed; measured the conventional way, against peak equity, the same fall is 3.89%.

以上都按实际投入的资金计算，取 44 个交易日的日收益，夏普比率按无风险利率为 0 计算。 44 个样本非常少，这些比率也就继承了这个毛病：夏普 1.63 的置信区间从负数一直延伸到 6 以上，贝塔 0.13 的置信区间也轻松包含 0.5。把它们当作对这一轮的描述就好，别当成能独立成立的数字。我唯一愿意直接下的判断是描述性的，不是统计性的：这本账的日波动和大盘几乎不同步，相关系数 0.10，指数只能解释大约 1% 的日内变化。这 +5.55% 不是大盘带上去的。回撤是相对投入资金计算的；若按更通行的做法，相对峰值净值计算，同一段跌幅是 3.89%。

What I was holding

仓位构成

RUSHA 25.4% Trucking & commercial vehicles 卡车与商用车 +3.01pp
CENTA 24.8% Lawn, garden & pet 园艺与宠物用品 +1.72pp
NDSN 24.7% Industrial precision tech 工业精密技术 +0.50pp
CHD 25.1% Consumer staples 日用消费品 +0.32pp

Four names, four unrelated industries, sized to roughly a quarter each. That part was deliberate; what came back out of them wasn't. The bars are each position's contribution to the +5.55%, and RUSHA alone accounts for more than half of it. 四只股票、四个互不相干的行业，每只大约四分之一仓位，这部分是我刻意安排的；至于每只最后吐回来多少，就由不得我了。条形表示每只对 +5.55% 的贡献，光 RUSHA 一只就占了一半以上。

Every position

全部持仓明细

RUSHA Rush Enterprises, Class A Rush Enterprises（A 类） 66.5 sh · $71.67 → $80.15

+$564+11.83%
CENTA Central Garden & Pet, Class A Central Garden & Pet（A 类） 132.31 sh · $35.10 → $37.54

+$323+6.95%
NDSN Nordson Nordson（诺信） 16.02 sh · $288.30 → $294.15

+$94+2.03%
CHD Church & Dwight Church & Dwight 48.35 sh · $97.32 → $98.55

+$59+1.26%
Four positions 四笔仓位 all closed 31 Jul 2026 均于 2026 年 7 月 31 日了结

+$1,040+5.55%

Prices are average fill, fractional shares included; percentages are on capital actually deployed, not account size. The two ends of this page agree: adding up the fills gives $1,039.94, the account's own equity gives $1,040.06, a twelve-cent gap. That is consistent with no dividends landing over the 63 days, and it is not proof of it, because the fills shown here are rounded to the cent and the rounding across four positions is about the size of the gap being used as evidence. Fees were never going to appear in a paper account either way. That check is on the endpoints. In between, the curve plots the account's daily marks exactly as it recorded them, and on 29 Jul it repeated the previous session's equity to the cent, so that one point sits flat where the underlying positions moved. Two of the 44 daily returns are affected by that, not one: 29 Jul reads as a flat day, and 30 Jul absorbs two sessions of movement in a single step. Both sit inside the series the volatility, Sharpe and beta above are computed from, and inside the 24 / 19 / 1 count. One cycle of four positions is long enough to be honest about and too short to be a track record; the next one goes up here the same way, win or lose.

价格为平均成交价，含零碎股；百分比按实际投入的资金计算，而非账户总额。这一页的两端是对得上的：把每笔成交加起来是 $1,039.94，账户自己的净值是 $1,040.06，差 12 美分。这与“这 63 天里没有分红进来”是相符的，但还算不上证明：页面上的成交价都四舍五入到分，四只股票摊下来的舍入误差，本身就和这 12 美分差不多大。手续费则本来就不会出现，这是一个模拟盘。这项核对只针对首尾两端。中间那段，曲线是把账户每天记下的净值原样画上去的；7 月 29 日那天，账户把前一交易日的净值一分不差地重复了一遍，所以那一个点是平的，而底下的持仓其实在动。受影响的是 44 个日收益里的两个，不是一个： 7 月 29 日被记成了持平的一天，7 月 30 日则把两个交易日的涨跌压进了一步。这两个都在上面算波动率、夏普和贝塔所用的那串数字里，也在 24 / 19 / 1 的计数里。一轮四只股票，够诚实地讲清楚，但远不足以称为长期业绩；下一轮不管输赢，也照这样贴上来。