added LWMA

src/cryptonote_basic/difficulty.cpp

@@ -33,6 +33,7 @@
 #include <cstddef>
 #include <cstdint>
 #include <vector>
+#include <boost/math/special_functions/round.hpp>
 
 #include "common/int-util.h"
 #include "crypto/hash.h"
@@ -162,4 +163,81 @@ namespace cryptonote {
     return (low + time_span - 1) / time_span;
   }
 
+// LWMA difficulty algorithm
+// Background:  https://github.com/zawy12/difficulty-algorithms/issues/3
+// Copyright (c) 2017-2018 Zawy (pseudocode)
+// MIT license http://www.opensource.org/licenses/mit-license.php
+// Copyright (c) 2018 Wownero Inc., a Monero Enterprise Alliance partner company
+// Copyright (c) 2018 The Karbowanec developers (initial code)
+// Copyright (c) 2018 Haven Protocol (refinements)
+// Degnr8, Karbowanec, Masari, Bitcoin Gold, Bitcoin Candy, and Haven have contributed.
+
+// This algorithm is: next_difficulty = harmonic_mean(Difficulties) * T / LWMA(Solvetimes)
+// The harmonic_mean(Difficulties) = 1/average(Targets) so it is also:
+// next_target = avg(Targets) * LWMA(Solvetimes) / T.
+// This is "the best algorithm" because it has lowest root-mean-square error between 
+// needed & actual difficulty during hash attacks while having the lowest standard 
+// deviation during stable hashrate. That is, it's the fastest for a given stability and vice versa.
+// Do not use "if solvetime < 1 then solvetime = 1" which allows a catastrophic exploit.
+// Do not sort timestamps.  "Solvetimes" and "LWMA" variables must allow negatives.
+// Do not use MTP as most recent block.  Do not use (POW)Limits, filtering, or tempering.
+// Do not forget to set N (aka DIFFICULTY_WINDOW in Cryptonote) to recommendation below.
+// The nodes' future time limit (FTL) aka CRYPTONOTE_BLOCK_FUTURE_TIME_LIMIT needs to
+// be reduced from 60*60*2 to 500 seconds to prevent timestamp manipulation from miner's with 
+//  > 50% hash power.  If this is too small, it can be increased to 1000 at a cost in protection.
+  
+// Cryptonote clones:  #define DIFFICULTY_BLOCKS_COUNT_V2 DIFFICULTY_WINDOW_V2 + 1
+
+
+  difficulty_type next_difficulty_v2(std::vector<std::uint64_t> timestamps, std::vector<difficulty_type> cumulative_difficulties, size_t target_seconds) {
+
+    const int64_t T = static_cast<int64_t>(target_seconds);
+    
+    size_t N = DIFFICULTY_WINDOW_V2;
+
+    // Return a difficulty of 1 for first 3 blocks if it's the start of the chain.
+    if (timestamps.size() < 4) { 
+      return 1; 
+    }
+    // Otherwise, use a smaller N if the start of the chain is less than N+1.
+    else if ( timestamps.size() < N+1 ) { 
+      N = timestamps.size() - 1;
+    }
+    // Otherwise make sure timestamps and cumulative_difficulties are correct size.
+    else {  
+      timestamps.resize(N+1);  
+      cumulative_difficulties.resize(N+1);
+    }
+    // To get an average solvetime to within +/- ~0.1%, use an adjustment factor.
+    // adjust=0.999 for 80 < N < 120(?)
+    const double adjust = 0.998;
+    // The divisor k normalizes the LWMA sum to a standard LWMA.
+    const double k = N * (N + 1) / 2;
+
+    double LWMA(0), sum_inverse_D(0), harmonic_mean_D(0), nextDifficulty(0);
+    int64_t solveTime(0);
+    uint64_t difficulty(0), next_difficulty(0);
+
+    // Loop through N most recent blocks. N is most recently solved block.
+    for (size_t i = 1; i <= N; i++) {
+      solveTime = static_cast<int64_t>(timestamps[i]) - static_cast<int64_t>(timestamps[i - 1]);
+      solveTime = std::min<int64_t>((T * 7), std::max<int64_t>(solveTime, (-7 * T)));
+      difficulty = cumulative_difficulties[i] - cumulative_difficulties[i - 1];
+      LWMA += solveTime * i / k;
+      sum_inverse_D += 1 / static_cast<double>(difficulty);
+    }
+
+    harmonic_mean_D = N / sum_inverse_D;
+
+    // Keep LWMA sane in case something unforeseen occurs.
+    if (static_cast<int64_t>(boost::math::round(LWMA)) < T / 20)
+      LWMA = static_cast<double>(T / 20);
+
+    nextDifficulty = harmonic_mean_D * T / LWMA * adjust;
+
+    // No limits should be employed, but this is correct way to employ a 20% symmetrical limit:
+    // nextDifficulty=max(previous_Difficulty*0.8,min(previous_Difficulty/0.8, next_Difficulty));
+    next_difficulty = static_cast<uint64_t>(nextDifficulty);
+    return next_difficulty;
+  }
 }