Recommended Free Tools
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
There is no fixed Elo value for one extra ply. In LK Forge’s 2026 self-play benchmark, the gain depended sharply on the game and depth: chess gained +321 Elo from depth 1 to 2, +315 from 2 to 3, and +225 from 3 to 4. Those are differences between the same engine playing at adjacent search depths—not ratings against people or other engines.
What the benchmark found
LK Forge tested four fixed-depth negamax engines—Connect 4, Checkers, Othello, and Chess—and found that an extra ply did not have a consistent value across games. The largest first-step gains came in the games other than chess:
| Game | First extra ply | Depths tested |
|---|---|---|
| Connect 4 | +953 Elo, depth 1 to 2 | 1–7 |
| Checkers | +546 Elo, depth 1 to 2 | 1–7 |
| Othello | +610 Elo, depth 1 to 2 | 1–6 |
| Chess | +321 Elo, depth 1 to 2 | 1–4 |
Chess’s adjacent-depth gains were +321 Elo from 1 to 2, +315 from 2 to 3, and +225 from 3 to 4. The first two steps were nearly equal, while the third was smaller. These figures are the benchmark’s colour-balanced self-play Elo steps; they should not be read as a universal exchange rate between depth and strength. LK Forge’s benchmark and methodology describe one implementation, one set of seeded openings, and a limited range of depths.
What “one ply” means
A ply is one half-move: one player makes a move, then the other player’s reply is another ply. Search depth counts these half-moves from the current position. A depth of four therefore looks ahead four individual moves, rather than four complete turns.
#1 Best Overall
- Bantam, A great option for a Book Lover
- Must try for a book lover
- Compact for travelling
That count is a practical search label, not necessarily a uniform examination of every possible line. As the Stockfish FAQ explains, “In practice, a reported ‘depth’ is a useful guide but not a strict guarantee that you’ve examined every move to that ply.” Real engines use techniques such as pruning, reductions, and extensions to focus computation on promising or important branches. A simple full minimax search would inspect every line to its stated depth; modern engines do not necessarily do that.
How LK Forge measured the Elo steps
The benchmark reports 2,400 headless self-play games in total. For each adjacent pair of depths, it played 120 colour-balanced games. It used random legal opening moves, replayed each opening with the sides swapped, and counted a draw as half a point. The author specified deterministic master seed 20260819 and calculated each Elo step from the colour-balanced score using 400 × log10(p/(1−p)).
Colour balancing helps reduce the effect of moving first, but 120 games per depth pair is still a finite sample. LK Forge also notes that draws are frequent in chess and checkers, so the Elo gaps in those games are approximate. The values are best understood as results of this test setup, not precise predictions of how much strength any engine gains from an extra ply.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Why the first extra ply can matter so much
At depth one, an engine can miss an opponent’s immediate reply: it evaluates a move without seeing what follows. Looking one half-move farther can expose a simple tactical consequence and prevent many one-move errors. That helps explain why LK Forge’s depth-1-to-2 step was especially large in Connect 4, Checkers, and Othello.
Chess behaved differently across its tested range. Its large branching factor means the position tree expands rapidly, so even an additional ply leaves much of the tree unexplored. LK Forge’s chess results show substantial gains at every measured step, but they do not establish a general curve for chess engines or for deeper searches.
What these numbers do—and do not—compare
LK Forge changed search depth while keeping the engine implementation fixed. That makes its question narrower than “Which engine is strongest?” and less dependent on processor speed than a comparison that gives each move the same number of milliseconds. It still does not isolate a universal value for depth: results can change with the evaluation function, search method, hardware, positions, time control, and draw rate.
Rank #4
- Used Book in Good Condition
Search speed is not strength by itself. Stockfish’s advanced documentation describes nodes per second (NPS) as useful for comparing speed, while cautioning that NPS differences between engine versions do not necessarily correspond to playing-strength differences.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesDifferent search approaches also make raw depth figures difficult to compare. Leela Chess Zero’s FAQ describes its neural-network engine and Monte Carlo Tree Search, and notes that it can run on a CPU but is significantly slower than on a modern GPU. Its search counts or depth should not be directly equated with LK Forge’s fixed-depth negamax results; hardware and testing conditions matter in strength comparisons.
For a separate consumer-analysis context, Chess.com’s August 3, 2026 support article says its server-side analysis uses Stockfish 18 with NNUE and describes diminishing returns in engine quality past the low 20s for broadcast analysis. That is Chess.com’s account of its own analysis use, not a universal depth threshold for engines or positions. Read Chess.com’s explanation.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to judge an “extra ply” claim
When comparing search-strength claims, check that the test holds the important variables steady. A meaningful result should make clear:
- Whether the engines were given fixed depth, fixed time, or a fixed node budget.
- Which search method and evaluation model were used.
- What hardware and thread count ran the tests.
- Which positions or openings were played and whether colours were balanced.
- How many games were played, how draws were scored, and how uncertainty was handled.
- Whether “Elo” means an internal self-play difference or a rating measured against an outside pool.
LK Forge provides a useful, controlled illustration of depth scaling within four implementations. Its results demonstrate why “one ply is worth X Elo” has no general answer: the value depends on the game, the depth interval, and the way strength was measured.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
