DFS and pick'em APIs: what to check

DFS and pick'em are different data

People search for a DFS API when they mean two different products. Salary-cap DFS has contests, slates and player salaries. Pick'em has a player, a stat, a target and a choice such as More or Less. The two have different records and different questions, so a provider that covers one may not cover the other.

A multiplier is neither a salary nor a sportsbook price. The three-way guide goes through what each number can support.

What this API supplies

We supply DraftKings data only: selected Pick6 categories and selected DFS slates. We do not supply Underdog, Sleeper, Betr, Dabble or any other pick'em feed. If you need several pick'em operators in one response, this API cannot do that.

curl --fail-with-body "https://draftkingsapi.com/api/v1/pick6/projections?sport=NFL&stat=Receiving%20Yards&limit=5" \
  -H "Authorization: Bearer $DRAFTKINGS_API_KEY"

curl --fail-with-body "https://draftkingsapi.com/api/v1/dfs/slates?sport=NFL&limit=5" \
  -H "Authorization: Bearer $DRAFTKINGS_API_KEY"

For the other operators we publish research rather than data: Underdog provider research and Sleeper provider research.

Who claims Pick6 coverage

Several multi-book providers publicly list Pick6 as a source. These are their own statements on the date below. We have not tested them, and a listed key does not tell you depth, freshness or how selections are represented.

• OpticOdds lists DraftKings (Pick 6) and DraftKings (Pick 6 Multipliers) in its sportsbook list. It also lists Sleeper and Underdog Fantasy entries.
• ParlayAPI lists a pick6 bookmaker key classed as DFS pick'em, alongside underdog and sleeper. Its docs describe native directions, selection IDs and a payout_multiplier_scope field.
• The Odds API lists pick6 in its us_dfs region and says selections with non-default multipliers are included in alternate markets.

A provider that spans operators is the stronger choice when breadth matters. The provider evaluation compares them in more detail.

What to test in a pick'em API

Pick'em data has failure modes that a sportsbook feed does not. Check these with real rows rather than from a feature grid.

1. One-sided markets. Does a More-only row stay one-sided, or does the feed fill in a side?
2. Multipliers. Is the value labelled as a standings multiplier, payout multiplier or odds? Which scope does it apply to, the selection or the entry?
3. Alternate targets. Can you see several targets for one player and stat, and can you tell them apart by ID?
4. Paused and live state. Are paused rows marked, and is a live stat clock kept separate from the offer clock?
5. Clocks. Is there a capture time? Is an unknown source time reported as unknown, or replaced with something else?
6. Coverage honesty. Does the feed say a category was not captured, or do rows just disappear?
7. Disappearance. Does a row missing from a response mean settlement, removal or a partial capture?

What to test in a DFS API

Salary data is simpler to read and easier to misuse. Most errors come from identity and from assuming complete slates.

• Slot records. One athlete can hold several draftables in one slate. Deduplicating by athlete alone drops valid salaries.
• Slate selection. Does the feed have every draft group or a selected set? On this API it is a selected set.
• Contest state. A contest in a lobby capture is an observation, not proof that it is open.
• Schema by sport. Fields under data vary, so validate before reading rather than assuming one schema.
• Rights. DFS records are public-source observations, and redistribution rights are unverified here. Ask any provider how this applies to your product.

curl --fail-with-body "https://draftkingsapi.com/api/v1/dfs/players?sport=NFL&slate_id=YOUR_SLATE_ID&limit=25" \
  -H "Authorization: Bearer $DRAFTKINGS_API_KEY"

Joining the two datasets

In a 2026-11-03 sample of same-named NFL players, Pick6 entity_id matched DFS playerDkId for 128 of 129. That is an observation to verify, not a guarantee, and it does not hold for playerId. Check it for your sport and treat unmatched rows as unmatched.

python3 comparison-feed.py --sport NFL --stat "Receiving Yards" --max-requests 12

The example reports what it joined and what it could not. It does not compute a mismatch or a pick, because a target and a salary are different kinds of number.

Check the recorded data

• Sources and status
• Measured coverage report
• Recorded Pick6 props

Sources

1. OpticOdds docs: sportsbook list checked 2026-11-03
2. ParlayAPI: documentation checked 2026-11-03
3. The Odds API: bookmaker list checked 2026-11-03

Related resources