DE|EN

AVALONPLUS — Strategy Report

By Frank Gärtner

Date: 2026-05-08  |  Starting Capital: $10,000 / $100,000  | Instrument: MES Futures ($5/Punkt)  |  Costs: always ON

Contents
0.Who is AVALONPLUS suitable for?1.Compounding Comparison (3 Curves)2.Source of ReturnsThe Three Signal Types3.Performance — Three Time Windows4.Live-Trading vs. Backtest5.Drawdowns in Practice6.Path Dependency — Entry Timing and Realized Return7.Risk Quartet8.Benchmark Comparison9.Glossary10.Methodological Limitations11.Next Steps
Backtest 2000–today at a glance · MES Futures · $10K starting capital · costs ON
31.5%
CAGR (Reinvest)
–17.2%
Max Drawdown (Reinvest)
0.42
Sharpe Ratio
74%
Win Rate (2000–today)
62
Trades 2000–today · 132 since 1963
2–5
Trades/year (approx.)
0 (start: 03/2026)
Live Trades

0. Who is AVALONPLUS suitable for?

AVALONPLUS is a rule-based EOD strategy on MES/ES Futures with underlying SP500 index (SPX). All examples refer to the MES Future ($5 point value) for simplicity. The system generates approx. 2–5 trades per year. In backtest 2000–today, $10,000 with reinvest achieved a net profit of $12,492,594, without reinvest $132,595. With $100K, reinvest achieved a net profit of $13,916,416, without reinvest $1,457,876. The difference between both scenarios arises solely from growing position size and the normalization of position size to realistic values — not from better signals.

Suitable if:
  • Risk capital that can be completely lost
  • Willingness to consistently reinvest profits
  • Rule-based trading even during drawdown phases
  • Trend-following approach with short component desired
Not suitable if:
  • Capital with a planned use within 5–10 years
  • Exit during drawdown >20% likely
  • Regular profit withdrawal planned
  • Daily trading signals expected

EOD strategy: signal check once daily after market close, approx. 2–5 minutes effort.

1. Compounding Comparison

Same trades, same period. Difference: starting capital and reinvest yes/no.

Logarithmic scale · 2000–today · MES Futures · — $10K · -- $100K

ScenarioStartBasePnL (Net Profit)Factor~CAGR
Buy & Hold S&P 500$38,9864.9×~7.5% p.a.
AVALONPLUS No Reinvest$10K$10K$132,59514.3×10.8% p.a.
AVALONPLUS Reinvest$10K$10K$12,492,5941250×31.5% p.a.
AVALONPLUS No Reinvest$100K$100K$1,457,87615.6×11.1% p.a.
AVALONPLUS Reinvest$100K$100K$13,916,416140×20.9% p.a.

Why is the difference between Reinvest and No Reinvest so large?

The system calculates position size from available capital. Concrete example:

L1 signal, SPX ~1,400 (year 2000) — Normalize Positions ON:
  Cost/contract: 1,500 × (11% + 7%) × $5 = $1,350
  $10K capital → floor($10,000 / $1,350) = 7 contracts

L1 signal, SPX ~5,000 (year 2024):
  Cost/contract: 5,000 × 18% × $5 = $4,500
  Without Reinvest ($10K base, unchanged capital):  floor($10,000 / $4,500) =   2 Kontrakte
  With Reinvest ($10K start, account now ~$500K):      floor($500,000 / $4,500) = 111 Kontrakte

The edge per trade (%) is identical.
The difference in absolute return arises solely from position size.

MES Futures · $5/point · Costs ON · Normalize Positions ON · No tax · $10,000 starting capital

2. Source of Returns

The Three Signal Types

L1 — Long Trend-Follow · 88% of Trades
Trend-follow long on SPX uptrend. Exit rule-based at trend end. Stop-Loss: –11%. Not triggered in 63 years of backtest (117 trades).
117 Trades·Win Rate 71%·Expected value $1,474/Trade
S1 — Short Counter-Trend · 6% of Trades
Profits in falling markets. Stop-Loss: –9%. Not triggered in 8 trades.
8 Trades·Win Rate 100%·Expected value $4,160/Trade
L2 — Long Follow Signal · 6% of Trades
Long after S1-exit when volatility retreats. Tighter stop (–1.5%), triggered in approx. 14% of L2 trades.
7 Trades·Win Rate 71%·Expected value $4,105/Trade
Signal Timing
Signals after market close, entry next day at open. Inputs: VIX proxy, SPX trend, momentum. Fully rule-based.

Stop-Distance-Based Sizing — the Compounding Anchor

Position size is budgeted against the planned stop distance — deliberately set wider than the historically observed MAE. Two consequences: the stop holds in the worst case, and even if it were to trigger, the per-trade loss is a fixed cap on the account.

qty = floor( capital ÷ ( entry_price × (stop_distance% + margin%) × point_value ) )

Example L1, $10K, SPX 5,000, stop 11% + margin 7% = 18%:
  qty = floor( 10,000 ÷ (5,000 × 18% × $5) ) = floor( 10,000 ÷ 4,500 ) = 2 contracts

Empirics: historical worst-case MAE for L1 is ~8%, stop at 11% — in 63 years not a single L1 or S1 stop has been triggered. L2 stop (–1.5%) is tighter; triggered once by design (2015). The effect: per-trade account drawdown is hard-capped, which keeps the reinvest path stable across small losing trades — the long L1 trades can build their capital buffer without the compounding being knocked off course by an intermediate string of losses.

Expected Value per Signal Type · 1963 to today · MES · $10K · no reinvestment

SignalTradesWin RateAvg WinAvg LossExpected value / Trade
L111771%$2,284$-505$1,474
S18100%$4,160$4,160
L2771%$6,376$-1,572$4,105
Total13273%$2,654$-564$1,776

EV per trade is the average PnL after costs ($/trade). The values scale with starting capital and reinvest — profit factor and win rate stay stable across instruments. Caveat: S1 (8) and L2 (7) are small samples, confidence intervals are wide — see Section 7 (Risk Quartet).

3. Performance — Three Time Windows

PeriodCapitalReinvestCAGRMax DDSharpeSortinoWin Rate
Last 3 Years$10KYes59.4%32.4%1.041.8070.0%
Last 3 Years$100KYes69.0%32.8%1.051.9270.0%
Last 3 Years$10KNo41.6%11.0%0.964.2670.0%
Last 3 Years$100KNo46.5%19.9%0.932.9570.0%
2000–today$10KYes31.1%17.2%0.423.5674.2%
2000–today$100KYes20.7%14.7%0.394.0474.2%
2000–today$10KNo10.6%16.0%0.293.4974.2%
2000–today$100KNo11.0%19.9%0.303.4774.2%
1963–today$10KYes12.6%27.6%0.390.6672.7%
1963–today$100KYes8.8%27.4%0.310.3472.7%
1963–today$10KNo5.2%17.0%0.110.1572.7%
1963–today$100KNo5.3%19.9%0.110.1772.7%
Buy & Hold S&P 500~7.5%–56.8% (2009, all-time)~0.50~0.30n/a
Sharpe note: The system trades only ~2–5 times per year. In months without trades the return is 0%. This compresses the Sharpe ratio — the Sortino (which only penalizes downside volatility) is more informative. Sortino >2 is considered very good.

Sortino $10K vs. $100K (no reinvestment): The higher Sortino at $10K no reinvestment (~4.3 vs. ~3.0 at $100K) is not an error — it arises from the position-cap effect: At $10K, MES contracts (minimum size ~$50) represent a larger share of capital than at $100K. Return dispersion relative to the capital base is therefore higher, which disproportionately reduces downside volatility and lifts the Sortino ratio.
Max DD of short periods: The Max-DD-% for short time windows can exceed the 26-year value — this is mathematically correct and not a contradiction. Reason: DD-% is measured against the peakCumProfit of the respective time window. Example: a $3,000 decline after $9,200 profit in the 3-year window = 32% DD (basis: $9,200 profit peak). The same $3,000 decline against a 26-year peak of $12.5M would be <0.03% — the relevant basis is the historical profit peak. Details on Max DD 2000-today are in Section 5.

MES Futures · No tax · Sharpe/Sortino annualized · Rf = 4% p.a.

4. Live Trading vs. Backtest

ParameterStatus
Live-Trade StartMarch 2026
Live Trades so far0
Out-of-Sample Performancenot yet available
Backtest period1963–2026 · 63 years · 132 Trades
Backtest CAGR (Reinvest, 2000–today)$10K: 31.5% p.a. · $100K: 20.9% p.a.
Backtest Max Drawdown (Reinvest, 2000–today)$10K: 17.2% · $100K: 14.7%

AVALONPLUS has been evaluated exclusively in backtest to date. Frank Gärtner started live trading in March 2026. Once 12+ live trades are available, a comparison table Backtest vs. Live will be added here. This is the actual robustness test of the strategy.

What backtest results do not capture: Slippage in real crash phases can be 5–20 ticks instead of the assumed 2 ticks. Gap-opens on volatile days can degrade entry quality. Psychological factors (holding through drawdown, executing signals consistently) are only measurable in live trades.

5. Drawdowns in Practice

Underwater Curve — reinvestment/no reinvestment $10K · reinvestment/no reinvestment $100K · Buy&Hold S&P 500 · 2000–today

Max Drawdown — $10,000 Starting Capital

Reinvest · 2000–today
17.2%
No Reinvest · 2000–today
16.0%
Reinvest · 1963–today
27.6%

Reinvest basis: peak of cumulated profit (= peakEquity − starting capital). A drawdown of e.g. 17% means: 17% of the profit earned up to that point was given back from the peak — not 17% of the starting capital.

Max Drawdown — $100,000 Starting Capital

Reinvest · 2000–today
14.7%
No Reinvest · 2000–today
19.9%
Reinvest · 1963–today
27.4%

Max Drawdown — Buy & Hold S&P 500 (Reference)

2000–today
50.8%

Measured at trade exit dates (not continuous) — the historical all-time max DD of the S&P 500 is –56.8% (March 2009).

Top-5 DD Periods — Reinvest $10K · 2000–today

StartTroughMax DD
2010-05-252010-05-25-17.2%
2000-09-202000-09-20-14.3%
2011-03-152011-06-24-11.6%
2015-09-242015-09-24-6.2%
2012-04-172012-05-30-6.0%

Top-5 DD Periods — No Reinvest $10K · 2000–today

StartTroughMax DD
2010-05-252010-05-25-16.0%
2015-09-242015-09-24-15.4%
2022-02-072023-09-07-11.9%
2025-03-212025-03-21-11.0%
2011-03-152011-06-24-10.6%

Top-5 DD Periods — Buy & Hold S&P 500 · 2000–today

StartTroughMax DD
2008-10-102008-10-10-40.5%
2000-09-202003-08-26-32.8%
2020-04-032020-04-03-24.9%
2025-03-212025-04-22-13.0%
2022-02-072023-03-08-12.6%

Additional Metrics · MES · 2000–today

L1 Stop triggered
0/117
S1 Stop triggered
0/8
L2 Stop triggered
1/7

An L1 stop (–11%) has not been triggered in 63 years (117 trades). A first trigger would cost approx. –61% of the trade capital deployed, not of the total account. The L2 stop (–1.5%) was triggered once (2015). For SPX/SPX-CFD the stop price is slightly different — 2 L2 stops were registered there.

6. Path Dependency — Entry Timing and Realized Return

Reinvest · X-axis: years since entry · MES Futures · — $10K · -- $100K

VariantPnL $10KFactorCAGR $10KPnL $100KFactorCAGR $100K
2000 (Dot-Com Crash)  (26 years)
Reinvest$12,492,5941250×31.5%$13,916,416140×20.9%
No Reinvest$132,59514×10.8%$1,457,87616×11.1%
Buy&Hold$38,9866.3%$389,8586.3%
2010 (after Financial Crisis)  (16 years)
Reinvest$822,13683×32.4%$1,352,43015×18.5%
No Reinvest$96,06811×16.2%$1,075,54412×16.9%
Buy&Hold$52,75612.4%$527,55912.4%
2016 (Bull Market)  (10 years)
Reinvest$97,29411×27.6%$423,17818.5%
No Reinvest$71,35524.0%$807,76025.4%
Buy&Hold$23,63413.3%$236,33813.3%

The achievable result depends significantly on the entry timing. An entry before the Dot-Com Crash (2000) delivers over 26 years a CAGR of ~31% ($10K) or ~21% ($100K), because the system navigated crash phases early and early SPX price levels allow proportionally more contracts. An entry in 2016 in the middle of the bull market shows ~28% ($10K) / ~19% ($100K) CAGR — entry timing and holding period length significantly influence the result. The 26-year period from 2000 is the most conservative and methodologically sound reference, as it contains all relevant market regimes (Dot-Com, financial crisis, bull market, COVID).

MES Futures · Reinvest · Costs ON · No tax · All trades from the respective start date

7. Risk Quartet

L1 Stop Untested
0 of 117 L1 trades in 63 years ended by stop. First trigger: approx. –61% of the trade capital deployed (not the total account). The scenario is parameterized but empirically untested.
S1/L2: Small Sample Size
S1: 8 trades, L2: 7 trades. Confidence intervals for win rate and expected value are wide. L2 stop triggered once (2015, –1.5%) — by design as a tighter protective stop. Statistical significance is insufficient for reliable forecasts.
Slippage Assumption
Backtest: 2 ticks slippage. In stress situations (flash crash, gap-open) 5–20 ticks may occur. MES is liquid enough for moderate deviations but not infinitely tolerant.
Compounding Path Dependency
The reinvest end values depend heavily on the historical trade sequence. An early large drawdown would have substantially reduced the 26-year result, even with an unchanged edge.

Summary of the four open points — detailed analysis in Section 10.

8. Benchmark Comparison

MetricS&P 500 Buy & HoldSG Trend Index (CTA)AVALONPLUS Reinvest
CAGR (2000–today)~7.5%~5.1%31.5%
Max Drawdown–56.8% (2009, all-time)~–21%–17.2%
Sharpe Ratio~0.5~0.40.42
Sortino Ratio~0.3~0.53.56
Calmar Ratio (10Y)~1.0~0.241.83
Sharpe interpretation: AVALONPLUS and S&P 500 have similar Sharpe values (~0.4–0.5) because the strategy makes only 2–5 trades/year and has many 0%-return months that inflate the standard deviation.
The Sortino (penalizes only downside volatility) at >3 is far above the market — the downside is structurally limited by the sizing system.

Calmar note: 1.83 refers to 2000–today (26 years, simulation $10K, Reinvest). For fair comparisons with benchmarks like the SG Trend Index the 26-year value is used.

SG Trend Index: approximation based on publicly available data. AVALONPLUS: MES · Reinvest · $10,000 starting capital · Costs ON · No tax.

9. Glossary

Sorted alphabetically — each term with definition and AVALONPLUS-specific example.

Bear Market / Bull Market
Bear Market: price decline >20% from the last high. Bull Market: uptrend >20% from the last low.
Example: AVALONPLUS profits in both bull markets (L1 long) and bear markets (S1 short). S1 was active in both major bear markets: Dot-Com 2000–2002 and Financial Crisis 2007–2009.
CAGR (Compound Annual Growth Rate)
The geometrically averaged annual return an investment would need to achieve to grow from starting to ending capital.
Example: $10,000 grow over 26 years to $12.5M. This equals a CAGR of ~31% — as if the value rose by that percentage every year, even though actual annual returns varied.
Calmar Ratio
CAGR divided by Max Drawdown. Measures return per unit of drawdown risk.
Example: AVALONPLUS 2000–today (Reinvest, $10K): CAGR ~31% / DD ~17.2% ≈ Calmar 1.8. S&P 500: CAGR ~7.5% / DD ~56.8% ≈ Calmar 0.13. The higher, the more efficient.
Drawdown
The maximum decline from the last peak to the trough before a new high.
Example: Account at $20K, falls to $16K, rises again to $25K. Drawdown = ($20K – $16K) / $20K = 20%. AVALONPLUS: Max DD ~17.2% (Reinvest) or ~19.9% (no reinvest) over 26 years (2000–today, $10K).
EOD (End of Day)
End-of-day trading: signals are checked once daily after market close, the trade is executed the next day at market open.
Example: Sunday evening you check the system. If an L1 signal was generated, the order is placed Monday at open. AVALONPLUS: approx. 2–5 minutes effort per day — no intraday monitoring needed.
Expected Value (EV)
Average per-trade outcome after costs. Formula: winRate × Avg Win + lossRate × Avg Loss. Says what each trade brings on average — regardless of any individual outcome. Positive = the strategy has an edge.
Example: AVALONPLUS: ~$1,776 EV/trade over 132 trades ($10K no reinvestment, MES). With ~73% win rate, avg win ~$2,654, avg loss ~–$564. EV is the central pro-trader metric: positive EV × sufficient trade count × risk management = sustainable return.
Leverage
Ratio between nominal value of the position traded and the capital deployed.
Example: $10K capital, 2 MES at SPX 5,000 = $50K nominal → leverage 5×. A 1% move in SPX moves your account by 5%. AVALONPLUS: L1 ~5.6×, S1 ~6.3×, L2 ~10×.
Kelly Criterion
Mathematically optimal bet size that maximizes geometric capital growth. Full Kelly typically produces ~50% Max DD.
Example: AVALONPLUS uses no direct Kelly, but setup-specific MAE-based sizing. Result: Max DD ~20% instead of ~50% with full Kelly — with competitive CAGR.
MAE (Maximum Adverse Excursion)
How far a trade moved against the position during its lifetime, even if it was profitable at the end.
Example: Long at SPX 5,000, falls to 4,800, rises to 5,300. MAE = –4%. Profit at end: +6%. AVALONPLUS: historically worst MAE 8% — well below the L1 stop at 11%.
Margin Call
The broker requests more capital because the open position has exceeded a defined loss threshold.
Example: AVALONPLUS: the MAE-based sizing keeps positions small enough that a margin call is not triggered even at the historically worst MAE (8%) — as long as the account does not shrink significantly from consecutive losses.
Margin Requirements (Initial & Maintenance Margin)
Every futures broker requires two capital amounts per contract: the initial margin (deposit when opening the position) and the maintenance margin (minimum amount to be held during the lifetime of the position). If the account falls below the maintenance margin, a margin call is issued. For the MES Future the CME minimum margin is approx. $1,200–$1,500 per contract (varies with volatility). Brokers may require higher amounts.
Example: L1 signal, SPX 5,000: initial margin $1,400/contract. With $10K capital up to 7 contracts would be mathematically possible — AVALONPLUS limits to 2 (MAE sizing). Even after an 8% MAE trade the account remains well above the maintenance margin. The sizing is designed so that margin calls do not occur under normal market conditions.
MES (Micro E-mini S&P 500)
Smallest futures contract on the S&P 500. Multiplier: $5 per index point.
Example: SPX at 5,000. One MES contract = 5,000 × $5 = $25,000 nominal value. With $10K capital you trade approx. 2 contracts at an L1 signal ($50K nominal, leverage ~5×).
Monte Carlo Simulation
The historical trades are simulated 1,000× in random order to measure how strongly the result depends on trade sequence.
Example: Actual Max DD: ~20%. Monte Carlo shows: in 5% of simulations the DD would be >43%. This means: the historical sequence was rather favorable — not luck, but not worst case either.
Out-of-Sample
Backtest data that was NOT used for strategy development. The actual robustness criterion.
Example: 1963–1999: Out-of-sample (strategy parameters were not fitted to this data). The strategy delivers weaker but positive results there — no overfitting signal.
Out-of-Sample Length
The number of trades or the time period in which the system was tested with data that was not used for strategy development.
Example: AVALONPLUS: live trading from April 2026. Rule of thumb for statistical significance: 20–30 live trades. Until this threshold is reached, backtest results are the only reference — they are hypothetical, but spanning 63 years and 132 trades.
Position Sizing
The method determining how many contracts to trade per signal.
Example: AVALONPLUS: qty = floor(capital / (EP × (DD% + Margin%) / 100 × 5)). At $10K, EP=5,000, L1 (11%+7%=18%): qty = floor(10,000 / 4,500) = 2 contracts.
Position-Cap (Normalize Positions)
An upper limit for the maximum number of contracts per trade — regardless of how much capital is available. Simulates realistic liquidity constraints.
Example: AVALONPLUS: cap at 500 MES contracts. With a $2.5M account approx. 556 contracts would fit — the cap limits to 500. With $10K starting capital the cap is practically never relevant (max. 7 contracts early on).
Profit Factor (PF)
Sum of all gains divided by sum of all losses. Rule of thumb: PF >2 = good, >3 = excellent.
Example: AVALONPLUS: PF ~12–17 (depending on preset) over 62–132 trades. The magnitude is plausible: the system is very selective (~2–5 trades/year) and losses are rare.
Reinvest / Compounding
Profits are not withdrawn but increase the capital available for subsequent trades. This leads to exponential rather than linear growth.
Example: $10K doubles to $20K → $40K → $80K — each doubling is larger in absolute terms. AVALONPLUS: factor ~1,250× (Reinvest) vs. ~14× (no reinvest) over 26 years.
Sharpe Ratio
Risk-adjusted return. Measures how much return is achieved per unit of volatility. Rule of thumb: >1.0 = good, >2.0 = very good.
Example: AVALONPLUS Sharpe ~0.42 — similar to S&P 500 (~0.5). Misleading: the system makes only 2–5 trades/year. Months without a trade count as 0% return and artificially inflate the standard deviation.
Slippage
Difference between expected and actually executed price. In liquid markets typically 1–2 ticks, in stress phases 5–20 ticks.
Example: AVALONPLUS backtest assumption: 2 ticks ($10 per MES contract). With 2 contracts per trade: $40 slippage. Even at 10 ticks in crashes MES remains comfortably profitable — individual gains are typically four-digit.
Sortino Ratio
Like Sharpe, but only downside volatility (negative returns) counts in the denominator. Does not penalize strategies for rising strongly.
Example: AVALONPLUS Sortino >3 — far above the market. The downside is structurally limited: L1/S1 stops were never triggered, L2 stops rarely. Therefore Sortino is the more relevant metric.
Sortino vs. Sharpe
Both measure risk-adjusted return, but Sortino only penalizes downside volatility.
Example: AVALONPLUS: Sharpe ~0.42 (like S&P 500), Sortino >3 (far above S&P 500 ~0.3). The discrepancy shows: the strategy has many 0%-months (compresses Sharpe) but hardly any real loss fluctuations (lifts Sortino).
Stop-Loss
A preset price at which the position is automatically closed.
Example: Long position at SPX 5,000, stop at 4,450 (–11%). AVALONPLUS: L1 stop at –11% (63 years, never triggered), L2 stop at –1.5% (triggered in ~14% of L2 trades — by design).
Volatility (VIX Proxy)
The VIX measures the implied volatility of S&P 500 options — a fear barometer of the market.
Example: AVALONPLUS uses a VIX proxy for L1 delays (L1V): when VIX >21 the system waits for entry until volatility retreats. This improves entry quality in volatile markets.
Win Rate
Percentage of profitable trades.
Example: AVALONPLUS: ~73% win rate over 132 trades. Important: not meaningful on its own. What matters is the ratio of avg. win to avg. loss. A system with 40% win rate and 5× payout is more profitable than one with 80% win rate and 1.1× payout.

10. Methodological Limitations

As outlined in Section 7 as an overview — here the detailed analysis of the four open points.

1. Statistical Limits for S1 and L2

S1 has only 8 trades, L2 only 7. Confidence intervals for win rate are wide. The strong crash results (S1: 100% win rate) are anecdotally compelling but not statistically guaranteed to repeat. A different market sequence would have produced different S1/L2 results.

2. Tail Risk Untested

L1 and S1 stops have never been triggered in 63 years. This is simultaneously the strongest authenticity argument and an untested worst-case scenario. A first trigger would cost approx. 56–61% of the trade capital deployed. The system was parameterized for this case, but not tested against it.

3. Compounding Path Dependency

The reinvest figures depend heavily on the historical sequence of trades. An early large drawdown would substantially reduce end capital after 26 years — even with an identical edge. The DD% parameters are derived from the historical MAE distribution and fixed — not optimized in backtest. This protects against overfitting of sizing parameters, but does not protect against a less favorable trade sequence in the future.

11. Next Steps

All scenarios shown in this report can be reproduced directly in the AVALONPLUS Android app backtester — with your own capital and reinvest settings. After market close the app sends a daily EOD push with the current signal status (L1 / S1 / L2 active, position yes/no).

The first 30 days are free — EOD push and backtester included. A subscription is required to continue.

Android App
Data through 2026-05-08  ·  Live Trades: 0  ·  All results hypothetical (backtest) — no guarantee of future performance.