When Does Less Information Pay? An Exact Finite Atlas of Representation, Payoff, and Cost
Organizations: Independent researchers, Moscow, Russia
Abstract
A representation that retains one bit about a four-state world can solve a task perfectly, while another that retains 1.5 bits on average reaches only 75% accuracy. The difference lies in which distinctions each representation preserves. We examine this relationship in a uniform four-state world by enumerating all admissible representations and binary action rules. Constructions are compared both by replacing one component while holding the others fixed and by choosing the best action rule separately for each module. We consider three tasks, their mixtures, and four separate hypothetical cost tariffs. In some comparisons, the representation retaining less world information has the higher payoff; in others, the more informative representation does. There are also pairs with different information amounts but equal payoff--that is, equal task accuracy. If a representation preserves all existing distinctions and adds new ones, its best zero-cost payoff cannot decrease. Lower-information advantages therefore depend on which distinctions are retained. Equal counts of opposing advantages under representation replacement follow from coding symmetry. In broader comparisons, the shares depend on how repeated realizations are counted. A different task set or resource price can change which member of a fixed pair performs better. These results are exact, complete censuses of the specified finite sets of constructions and comparisons. The atlas provides a reproducible worked model for examining how action constraints, counting rules, and cost assumptions affect comparisons. It illustrates established distinctions between information quantity and decision value rather than proposing a new general law of their relation. measures retained world information, and payoff measures task accuracy.
Figures & tables
| Census | Number of pairs | Number of rows |
|---|---|---|
| One-component replacements, all five types | 9,568 | 66,976 |
| Of these: representation replacements only | 1,416 | 9,912 |
| All pairs of the 181 modules | 16,290 | 114,030 |
| All pairs of the 15 representations, without repetitions | 105 | 735 |
| Resource priced | Notation | What the quantity measures |
|---|---|---|
| Sensory access | The number of distinguishable inputs beyond one | |
| Provisioned capacity | The number of provisioned categories beyond one, even if not all are used | |
| Representation categories | The number of categories actually used beyond one | |
| Retained world information | The amount of retained world information in bits |
| Representation | Categories | World information , bits | Task information , bits | Information in the action , bits | Best task payoff |
|---|---|---|---|---|---|
| 0011 | 2 | 1.000 | 1.000 | 1.000 | 1 |
| 0102 | 3 | 1.500 | 0.500 | 0.311 | 3/4 |
| Information and payoff-ceiling relationship | All module pairs: rows | Distinct representation pairs: rows |
|---|---|---|
| Equal information, regardless of payoff | 31,122 | 168 |
| Lower information, higher payoff ceiling | 7,344 | 60 |
| Higher information, higher payoff ceiling | 51,891 | 296 |
| Unequal information, equal payoff ceilings | 23,673 | 211 |
| Total | 114,030 | 735 |
| Tasks evaluated | Left construction’s payoff | Right construction’s payoff |
|---|---|---|
| Right-bit task | 1/4 | 3/4 |
| Equal mixture of the left-bit and XOR tasks | 1/2 | 1/4 |
| Resource priced | Resource amount | Rows with a positive threshold |
|---|---|---|
| Sensory distinctions | 31,415 | |
| Provisioned capacity | 30,414 | |
| Representation categories | 45,951 | |
| Retained world information | 51,891 |
| Stage | Decision | Assumption | Claim | Exact config SHA-256 |
|---|---|---|---|---|
| 7.2A | RDR-0028 | ASM-INT-0002 | CLM-INT-BOTTLENECK-001 | a1feca94512b19023e518a6000de7901cebb17e72d387873cc0d397ad94e6622 |
| 7.2B | RDR-0029 | ASM-INT-0003 | CLM-INT-GENOTYPE-001 | 0f2fe87d1347fcfd648bcdceb39395df91617ef2ad0926227ca86f6f5dd9298e |
| 7.2C | RDR-0030 | ASM-INT-0004 | CLM-INT-SELECTION-BRIDGE-001 | 5d3f2e25a80b60d9e89d4abbfaed5c4c295f44eb9f55f2bb00a3558cef40c8ef |
| Later analysis | Reason added | Scientific effect |
|---|---|---|
| representation-once quotient | expose multiplicity weighting in the 181-module census | alternative counting measure over the same realized representations; no new model rows |
| strict nested-refinement census | test whether lower-information witnesses arise by merely deleting distinctions | narrows the mechanism claim; no new model rows |
| active/inert denominator decomposition | distinguish all categorical replacements from those that ever change payoff | clarifies denominators only |
| information-decomposition and order examples | make the information and reversal mechanisms readable | selected mechanically, never by effect magnitude |
| exact threshold canonicalization | collapse algebraically equivalent symbolic quotients | changes the number of distinct spellings/classes, not any saved threshold row |
| Quantity | Representation 0011 | Representation 0102 |
|---|---|---|
| Used categories | 2 | 3 |
| Category labels in world order | [0,0,1,1] | [0,1,0,2] |
| Retained world information, bits | 1 | 3/2 |
| Task-relevant information, bits | 1 | 1/2 |
| All optimal policies | [[0,1]] | [[0,0,1],[1,0,1]] |
| Information in an optimal action, bits | 1 | (3/2)+(-3/4)*log2(3) |
| Table | Rows |
|---|---|
| representations | 15 |
| representation-policy-task | 282 |
| representation-task ceilings | 45 |
| frozen transfer | 90 |
| active-category thresholds | 315 |
| information-rate thresholds | 315 |
| Table | Rows |
|---|---|
| compatible modules | 181 |
| genotypes | 1,424 |
| genotype-task | 4,272 |
| module frozen-transfer envelopes | 1,086 |
| contextual one-locus contrasts | 28,704 |
| epistasis rectangles | 9,108 |
| Count | Exact derivation |
|---|---|
| 181 compatible modules | capacity subtotals 15 + 47 + 59 + 60 = 181 for K=1,2,3,4 |
| 1,424 coded constructions | the 181 modules have realized- |R| distribution 60, 96, 24, 1 ; therefore 602^2 + 962^3 + 242^4 + 12^5 = 1,424 , where the extra factor 2 is the action code |
| 9,568 one-coordinate identities | unordered alternative pairs within every fixed four-component background: 3,264 + 1,320 + 1,416 + 2,856 + 712 = 9,568 for F, K, R, D, alpha |
| Table | Rows |
|---|---|
| genotype-environment profiles | 9,968 |
| module-environment envelopes | 1,267 |
| registered one-coordinate/environment comparisons | 66,976 |
| module pair-environment comparisons | 114,030 |
| registered one-coordinate threshold rows | 267,904 |
| module threshold rows | 456,120 |
| Comparison set | Pairs | Objectives | Comparison rows |
|---|---|---|---|
| All registered one-component replacements | 9,568 | 7 | 66,976 |
| Representation, decision-code, and action-code family subtotal | 4,984 | 7 | 34,888 |
| Actually payoff-active registered replacements | 3,592 | 7 | 25,144 |
| Registered representation-component replacements | 1,416 | 7 | 9,912 |
| Unrestricted module-ceiling pairs | 16,290 | 7 | 114,030 |
| Realized representations counted once | 105 | 7 | 735 |
| Class | Count |
|---|---|
| equal world information | 5,880 |
| lower world information, higher payoff | 792 |
| higher world information, higher payoff | 792 |
| unequal world information, payoff tie | 2,448 |
| Class | Count |
|---|---|
| equal world information | 31,122 |
| lower world information, higher payoff | 7,344 |
| higher world information, higher payoff | 51,891 |
| unequal world information, payoff tie | 23,673 |
| Class | Count |
|---|---|
| equal world information | 168 |
| lower world information, higher payoff | 60 |
| higher world information, higher payoff | 296 |
| unequal world information, payoff tie | 211 |
| Relation of reoptimized ceilings | Count |
|---|---|
| finer representation higher | 90 |
| tie | 45 |
| coarser representation higher | 0 |
| Universe | Unique identities | Single-task-vs-multi-task reversals |
|---|---|---|
| registered one-coordinate comparisons | 9,568 | 2,116 |
| module pairs | 16,290 | 3,672 |
| Component family | Identities | Active in ≥1 objective | Inert in all 7 | Reversals |
|---|---|---|---|---|
| sensory filter | 3,264 | 0 | 3,264 | 0 |
| provisioned capacity | 1,320 | 0 | 1,320 | 0 |
| representation | 1,416 | 748 | 668 | 676 |
| decision code | 2,856 | 2,494 | 362 | 1,272 |
| action code | 712 | 350 | 362 | 168 |
| representation/decision/action family subtotal | 4,984 | 3,592 | 1,392 | 2,116 |
| Comparison | Objective | Left payoff | Right payoff | Left minus right |
|---|---|---|---|---|
| Fixed construction pair | Second-coordinate task | 1/4 | 3/4 | -1/2 |
| Same construction pair | Equal first-coordinate/XOR mixture | 1/2 | 1/4 | 1/4 |
| Module pair | First-coordinate task; information tariff | 1/2 | 3/4 | -1/4 |
| Tariff | Positive crossings | Payoff-and-cost dominance |
|---|---|---|
| sensory support | 31,415 | 10,983 |
| capacity provision | 30,414 | 16,035 |
| representation support | 45,951 | 1,404 |
| representation information | 51,891 | 7,344 |
| Tariff | Distinct values | Minimum | Median | Maximum |
|---|---|---|---|---|
| sensory support | 8 | 1/36 | 1/4 | 1/2 |
| capacity provision | 8 | 1/36 | 1/6 | 1/2 |
| representation support | 6 | 1/24 | 1/4 | 1/2 |
| representation information | 17 | ≈0.0833 | ≈0.3082 | ≈1.3247 |
Value of Information
Value of information studies whether acquiring information can improve decisions enough to justify its cost.
See all 39 papers