You're eldest hand (left of the dealer), the other team called trump, and the opening lead is yours. Pick the card that hurts the makers most.
Ace first — except 1 trump with scattered 2-1-1 going for a euchre. Then singleton. Dump the 9. Pull trump only when it's dead or you've got a fistful.
Answer the table in front of you and get the call. Same logic as the decision tree below.
The only place the two defensive goals genuinely disagree. Decide your priority, then lead.
Your lone low trump is probably getting dragged out next time trump is led, so saving it often buys nothing.
| Lead | Euchred | Kept ≤1 pt | Gave 2 pts |
|---|---|---|---|
| Lowest trump | 11.95% | 70.27% | 29.73% |
| Off-suit ace | 9.99% | 73.91% | 26.09% |
| Singleton / short low | 9.95% | 74.64% | 25.36% |
Best picks in red. Swipe a table sideways for the full row.
| You hold | Shape | Best for euchre | Euchred | Best to prevent 2 | Kept ≤1 |
|---|---|---|---|---|---|
| 0 trump | 2-2-1 | Off-suit ace | 10.6% | Off-suit ace | 72.6% |
| 0 trump | 3-1-1 | Off-suit ace | 9.6% | Off-suit ace | 70.7% |
| 0 trump | 3-2 | Off-suit ace | 9.3% | Off-suit ace | 69.4% |
| 1 trump | 2-1-1 | Lowest trump | 11.95% | Singleton / ace | 74.6% |
| 1 trump | 2-2 | Off-suit ace | 13.5% | Off-suit ace | 80.4% |
| 1 trump | 3-1 | Off-suit ace | 13.4% | Off-suit ace | 78.9% |
| 2 trump | 2-1 | Short side, low | 19.2% | Same | 87.5% |
| 2 trump | 1-1-1 | Lowest trump | 14.5% | Side-suit lead | 73.2% |
| 2 trump | 3-card | Off-suit ace | 26.6% | Off-suit ace | 91.2% |
| 3 trump | 1-1 / 2 | Lowest trump | 28–37% | Lowest trump | 93–96% |
| Doubleton | Lead | Avoid | Euchred (low / high) |
|---|---|---|---|
| K-9 | 9 | K | 10.98% / 10.83% |
| Q-9 | 9 | Q | 10.88% / 10.68% |
| J-9 | 9 | J | 10.27% / 9.78% |
| K-10 | K | 10 | 10.40% / 10.98% |
| Q-10 | 10 | Q | 10.63% / 10.38% |
| K-Q | K or Q | — | ~11.4% / 11.4% |
| Singleton | Best play | Euchred |
|---|---|---|
| 9 | Lead the 9 | 11.5% |
| 10 | Lead the 10 | 11.0% |
| K | Off-suit ace if held, else the K | 12.4% |
| Suit | Best lead | Worst lead |
|---|---|---|
| A-K-9 | Middle (K) or ace | 9 — only 9.85% |
| A-Q-10 | Ace (high) | Low |
| A-Q-9 | Ace (high) | Low |
| K-Q-10 | High or middle | Low |
| Lead | Euchred | Kept ≤1 | Gave 2 |
|---|---|---|---|
| Off-suit ace | 11.89% | 76.0% | 24.0% |
| Shortest suit, high | 11.61% | 76.5% | 23.5% |
| 3-card suit, high | 11.60% | 76.0% | 24.0% |
| Lowest trump | 11.50% | 70.0% | 30.0% |
| Singleton | 11.47% | 76.1% | 23.9% |
| Shortest suit, low | 11.47% | 76.1% | 23.9% |
Gold-ringed card is the lead. Trump is hearts in every example.
Trump ♥ — no hearts, no left bower (J♦) in hand.
Lead the A♠ — off-suit ace. No ace? Lead low from a doubleton (the 9♦) or a singleton.
Trump ♥.
No ace → lead the J♣ (singleton). From the K-9, you'd lead the 9♠, not the K.
Trump ♥. One trump, spades doubleton, two singletons.
Trump ♥.
No ace → lead the J♦ (singleton) or the 9♠ from the doubleton. Don't lead trump here.
Trump ♥.
From the 3-card club suit, lead the K♣ (middle/high) — not the 9♣.
The opening lead is one card of five. Here's what 375k hands say about the next four — opening lead fixed to the cheat-sheet pick, only your tricks 2–5 policy varies.
Most hands you lost trick 1, so your follow-up matters most there. Across every policy tested the gaps are small (~0.2–0.7 pts of euchre rate), with one clear loser: leading trump aggressively is consistently worst.
Two ideas that sound alike but aren't:
Win cheap, spend honours last
Spend garbage early. Spend honours only when the trick is yours to take. What you want left is your last winner or a trump — not "always save the highest card."
Differences between follow-up policies are small; baseline (generic play) holds up well and trump-aggressive is worst in every segment.
All hands
| Policy | Euchred | Kept ≤1 | Gave 2 |
|---|---|---|---|
| Baseline (generic) | 12.50% | 75.06% | 24.94% |
| Cash winners | 12.47% | 75.06% | 24.94% |
| Adaptive | 12.27% | 75.05% | 24.95% |
| Hoard honours | 12.26% | 75.05% | 24.95% |
| Strip suits | 12.14% | 74.95% | 25.05% |
| Trump aggressive | 11.76% | 74.95% | 25.05% |
They won trick 1 (~73%)
| Policy | Euchred | Kept ≤1 | Gave 2 |
|---|---|---|---|
| Cash winners | 7.36% | 65.71% | 34.29% |
| Baseline | 7.32% | 65.71% | 34.29% |
| Hoard honours | 7.20% | 65.70% | 34.30% |
| Strip suits | 7.11% | 65.56% | 34.44% |
| Trump aggressive | 6.99% | 65.56% | 34.44% |
You won trick 1 (~15%)
| Policy | Euchred |
|---|---|
| Baseline | 22.30% |
| Cash winners | 21.96% |
| Hoard honours | 21.56% |
| Strip suits | 21.48% |
| Trump aggressive | 20.11% |
Partner won trick 1 (~12%): baseline wins again (~31.2% euchre). Support partner — dump low when they're winning; don't over-trump or burn honours.
Re-run: python main.py --phase-d --hands 375000
Everything above assumes they called and you defend. This is the other chair: when should you take the bid, and when should you drop your partner and go alone?
Default is keep your partner — 0.78 vs 0.53 avg maker points, and a loner gets euchred ~60% of the time vs ~34% with help. Only drop them when the hand is genuinely big:
| Your hand | Verdict |
|---|---|
| Both bowers | Alone — slightly better |
| 4+ trump | Alone — much better |
| Monster hand (eval 9+) | Alone — much better |
| Strong hand (eval 7–8.9) | Alone — slightly better |
| 2 trump · 1 trump · no bower | Keep partner — not close |
You never need to add up a score at the table — just recognize the shape. Trump is hearts here, so J♥ is the right bower and J♦ is the left bower (both play as hearts).
Right bower plus trump A, K, 9 — four trump, mostly high.
Very strong call. With bowers and real trump length like this, going alone is slightly ahead in the sim.
Both bowers, trump ace and king, plus an off-suit ace.
This is loner territory — alone beat keeping partner clearly in Phase E.
Eval is the sim's hand-strength number for a chosen trump suit — not a rulebook rule. It adds up roughly how many winners you'd hold if that suit were trump, then buckets the total. You don't compute it at the table; the buckets do the work.
| What you hold (in trump) | Points |
|---|---|
| Right bower (J of trump) | 3.0 |
| Left bower (same-colour J) | 2.5 |
| Trump ace | 2.0 |
| Trump king | 1.5 |
| Trump queen | 1.0 |
| Trump 9 or 10 | 0.5 ea |
| Off-suit ace | +1.0 ea |
| Void in a side suit (with trump to ruff) | +0.5 ea |
| Bucket | Score | What it means |
|---|---|---|
| Below call | < 3.5 | Usually pass |
| Marginal | 3.5–4.9 | Thin call |
| Solid | 5.0–6.9 | Good call — keep partner |
| Strong | 7.0–8.9 | Alone candidate — the floor for going it alone |
| Monster | 9.0+ | Alone often right |
In plain terms: you want to clear 7.0 before dropping your partner — roughly both bowers, or a bower plus real trump length and some off-suit help. Below that, keep partner even on a confident call.
Re-run: python main.py --phase-e --hands 375000
375,000-hand Monte Carlo simulation (seed 42), stick-the-dealer on, realistic calling hands. The same deals are replayed under each strategy so comparisons are fair.
You're eldest; the dealer's team called on a realistic hand. After your lead, all players use the same basic heuristic — not expert human play. Phase D layers your own tricks 2–5 policy on top of that. Bidding is a simplified model.
Best vs worst strategies differ by ~1–2 points on euchre rate: strong guidance, not gospel.
Re-run anytime: python main.py --segmented --hands 375000, --phase-c, or --phase-d
182,951 hands. Verdict: when you have an ace, lead it. Don't lead small from A-x instead.
| Strategy | Euchred | Kept ≤1 | Gave 2 |
|---|---|---|---|
| Lead ace always | 12.68% | 78.24% | 21.76% |
| Ace only if singleton | 11.94% | 78.10% | 21.90% |
| Shortest side suit low | 11.83% | 77.83% | 22.17% |
| From A-x, lead small | 11.09% | 73.94% | 26.06% |
57,422 hands. With no ace, lead the singleton — not a doubleton. Which doubleton you'd pick made no difference.
| Strategy | Euchred | Kept ≤1 | Gave 2 |
|---|---|---|---|
| Off-suit ace | 10.63% | 72.56% | 27.44% |
| Singleton | 9.96% | 72.22% | 27.78% |
| Doubleton low | 8.88% | 68.61% | 31.39% |
374,282 hands. Leading the suit the dealer threw away on pickup is not better than ace or singleton. Interesting idea, but the ace still wins.
| Strategy | Euchred | Kept ≤1 | Gave 2 |
|---|---|---|---|
| Off-suit ace | 11.87% | 76.01% | 23.99% |
| Singleton / short low | 11.46% | 76.11% | 23.89% |
| Dealer's discard suit (low) | 11.21% | 74.74% | 25.26% |
Don't always pull trump. Use the conditional rules: 1 trump + 2-1-1 → low trump if chasing a euchre; 2+ trump with no ace → pulling is reasonable; always pulling is the worst on giving up 2 points.
| Strategy | Euchred | Kept ≤1 | Gave 2 |
|---|---|---|---|
| 1 trump + 2-1-1 (smart pull) | 12.49% | 74.92% | 25.08% |
| Off-suit ace always | 11.89% | 76.03% | 23.97% |
| 2+ trump, no ace | 11.80% | 75.21% | 24.79% |
| Always lowest trump | 11.50% | 70.00% | 30.00% |
Who called? Almost all hands are dealer pickup (374,282 of 375,000), so results match the overall table. The dealer's-partner-called bucket is only 703 hands — too small to trust fully.
Stronger partner AI: teaching the partner to trump when void did not improve results — euchre rates dropped slightly. The opening-lead rankings were unchanged, so the conclusions don't depend on a super-partner.