Introduce a scale factor

For the per confirmation number tracking of data, introduce a scale factor so that in the longer horizones confirmations are bucketed together at a resolution of the scale.  (instead of 1008 individual data points for each fee bucket, have 42 data points each covering 24 different confirmation values.. (1-24), (25-48), etc.. )

Author: morcos

src/policy/fees.cpp

@@ -53,13 +53,14 @@ class TxConfirmStats
 
     double decay;
 
+    // Resolution (# of blocks) with which confirmations are tracked
     unsigned int scale;
 
     // Mempool counts of outstanding transactions
     // For each bucket X, track the number of transactions in the mempool
     // that are unconfirmed for each possible confirmation value Y
     std::vector<std::vector<int> > unconfTxs;  //unconfTxs[Y][X]
-    // transactions still unconfirmed after MAX_CONFIRMS for each bucket
+    // transactions still unconfirmed after GetMaxConfirms for each bucket
     std::vector<int> oldUnconfTxs;
 
     void resizeInMemoryCounters(size_t newbuckets);
@@ -73,7 +74,7 @@ class TxConfirmStats
      * @param decay how much to decay the historical moving average per block
      */
     TxConfirmStats(const std::vector<double>& defaultBuckets, const std::map<double, unsigned int>& defaultBucketMap,
-                   unsigned int maxConfirms, double decay);
+                   unsigned int maxPeriods, double decay, unsigned int scale);
 
     /** Roll the circular buffer for unconfirmed txs*/
     void ClearCurrent(unsigned int nBlockHeight);
@@ -113,7 +114,7 @@ class TxConfirmStats
                              EstimationResult *result = nullptr) const;
 
     /** Return the max number of confirms we're tracking */
-    unsigned int GetMaxConfirms() const { return confAvg.size(); }
+    unsigned int GetMaxConfirms() const { return scale * confAvg.size(); }
 
     /** Write state of estimation data to a file*/
     void Write(CAutoFile& fileout) const;
@@ -128,17 +129,17 @@ class TxConfirmStats
 
 TxConfirmStats::TxConfirmStats(const std::vector<double>& defaultBuckets,
                                 const std::map<double, unsigned int>& defaultBucketMap,
-                               unsigned int maxConfirms, double _decay)
+                               unsigned int maxPeriods, double _decay, unsigned int _scale)
     : buckets(defaultBuckets), bucketMap(defaultBucketMap)
 {
     decay = _decay;
-    scale = 1;
-    confAvg.resize(maxConfirms);
-    for (unsigned int i = 0; i < maxConfirms; i++) {
+    scale = _scale;
+    confAvg.resize(maxPeriods);
+    for (unsigned int i = 0; i < maxPeriods; i++) {
         confAvg[i].resize(buckets.size());
     }
-    failAvg.resize(maxConfirms);
-    for (unsigned int i = 0; i < maxConfirms; i++) {
+    failAvg.resize(maxPeriods);
+    for (unsigned int i = 0; i < maxPeriods; i++) {
         failAvg[i].resize(buckets.size());
     }
 
@@ -172,8 +173,9 @@ void TxConfirmStats::Record(int blocksToConfirm, double val)
     // blocksToConfirm is 1-based
     if (blocksToConfirm < 1)
         return;
+    int periodsToConfirm = (blocksToConfirm + scale - 1)/scale;
     unsigned int bucketindex = bucketMap.lower_bound(val)->second;
-    for (size_t i = blocksToConfirm; i <= confAvg.size(); i++) {
+    for (size_t i = periodsToConfirm; i <= confAvg.size(); i++) {
         confAvg[i - 1][bucketindex]++;
     }
     txCtAvg[bucketindex]++;
@@ -202,6 +204,7 @@ double TxConfirmStats::EstimateMedianVal(int confTarget, double sufficientTxVal,
     double totalNum = 0; // Total number of tx's that were ever confirmed
     int extraNum = 0;  // Number of tx's still in mempool for confTarget or longer
     double failNum = 0; // Number of tx's that were never confirmed but removed from the mempool after confTarget
+    int periodTarget = (confTarget + scale - 1)/scale;
 
     int maxbucketindex = buckets.size() - 1;
 
@@ -236,9 +239,9 @@ double TxConfirmStats::EstimateMedianVal(int confTarget, double sufficientTxVal,
             newBucketRange = false;
         }
         curFarBucket = bucket;
-        nConf += confAvg[confTarget - 1][bucket];
+        nConf += confAvg[periodTarget - 1][bucket];
         totalNum += txCtAvg[bucket];
-        failNum += failAvg[confTarget - 1][bucket];
+        failNum += failAvg[periodTarget - 1][bucket];
         for (unsigned int confct = confTarget; confct < GetMaxConfirms(); confct++)
             extraNum += unconfTxs[(nBlockHeight - confct)%bins][bucket];
         extraNum += oldUnconfTxs[bucket];
@@ -339,6 +342,7 @@ double TxConfirmStats::EstimateMedianVal(int confTarget, double sufficientTxVal,
         result->pass = passBucket;
         result->fail = failBucket;
         result->decay = decay;
+        result->scale = scale;
     }
     return median;
 }
@@ -358,15 +362,15 @@ void TxConfirmStats::Read(CAutoFile& filein, int nFileVersion, size_t numBuckets
     // Read data file and do some very basic sanity checking
     // buckets and bucketMap are not updated yet, so don't access them
     // If there is a read failure, we'll just discard this entire object anyway
-    size_t maxConfirms;
+    size_t maxConfirms, maxPeriods;
 
     // The current version will store the decay with each individual TxConfirmStats and also keep a scale factor
     if (nFileVersion >= 149900) {
         filein >> decay;
         if (decay <= 0 || decay >= 1) {
             throw std::runtime_error("Corrupt estimates file. Decay must be between 0 and 1 (non-inclusive)");
         }
-        filein >> scale; //Unused for now
+        filein >> scale;
     }
 
     filein >> avg;
@@ -378,22 +382,24 @@ void TxConfirmStats::Read(CAutoFile& filein, int nFileVersion, size_t numBuckets
         throw std::runtime_error("Corrupt estimates file. Mismatch in tx count bucket count");
     }
     filein >> confAvg;
-    maxConfirms = confAvg.size();
+    maxPeriods = confAvg.size();
+    maxConfirms = scale * maxPeriods;
+
     if (maxConfirms <= 0 || maxConfirms > 6 * 24 * 7) { // one week
         throw std::runtime_error("Corrupt estimates file.  Must maintain estimates for between 1 and 1008 (one week) confirms");
     }
-    for (unsigned int i = 0; i < maxConfirms; i++) {
+    for (unsigned int i = 0; i < maxPeriods; i++) {
         if (confAvg[i].size() != numBuckets) {
             throw std::runtime_error("Corrupt estimates file. Mismatch in feerate conf average bucket count");
         }
     }
 
     if (nFileVersion >= 149900) {
         filein >> failAvg;
-        if (maxConfirms != failAvg.size()) {
+        if (maxPeriods != failAvg.size()) {
             throw std::runtime_error("Corrupt estimates file. Mismatch in confirms tracked for failures");
         }
-        for (unsigned int i = 0; i < maxConfirms; i++) {
+        for (unsigned int i = 0; i < maxPeriods; i++) {
             if (failAvg[i].size() != numBuckets) {
                 throw std::runtime_error("Corrupt estimates file. Mismatch in one of failure average bucket counts");
             }
@@ -449,8 +455,9 @@ void TxConfirmStats::removeTx(unsigned int entryHeight, unsigned int nBestSeenHe
                      blockIndex, bucketindex);
         }
     }
-    if (!inBlock && blocksAgo >= 1) {
-        for (size_t i = 0; i < blocksAgo && i < failAvg.size(); i++) {
+    if (!inBlock && (unsigned int)blocksAgo >= scale) { // Only counts as a failure if not confirmed for entire period
+        unsigned int periodsAgo = blocksAgo / scale;
+        for (size_t i = 0; i < periodsAgo && i < failAvg.size(); i++) {
             failAvg[i][bucketindex]++;
         }
     }
@@ -490,9 +497,9 @@ CBlockPolicyEstimator::CBlockPolicyEstimator()
     bucketMap[INF_FEERATE] = bucketIndex;
     assert(bucketMap.size() == buckets.size());
 
-    feeStats = new TxConfirmStats(buckets, bucketMap, MED_BLOCK_CONFIRMS, MED_DECAY);
-    shortStats = new TxConfirmStats(buckets, bucketMap, SHORT_BLOCK_CONFIRMS, SHORT_DECAY);
-    longStats = new TxConfirmStats(buckets, bucketMap, LONG_BLOCK_CONFIRMS, LONG_DECAY);
+    feeStats = new TxConfirmStats(buckets, bucketMap, MED_BLOCK_PERIODS, MED_DECAY, MED_SCALE);
+    shortStats = new TxConfirmStats(buckets, bucketMap, SHORT_BLOCK_PERIODS, SHORT_DECAY, SHORT_SCALE);
+    longStats = new TxConfirmStats(buckets, bucketMap, LONG_BLOCK_PERIODS, LONG_DECAY, LONG_SCALE);
 }
 
 CBlockPolicyEstimator::~CBlockPolicyEstimator()
@@ -864,7 +871,7 @@ bool CBlockPolicyEstimator::Read(CAutoFile& filein)
 
             std::map<double, unsigned int> tempMap;
 
-            std::unique_ptr<TxConfirmStats> tempFeeStats(new TxConfirmStats(tempBuckets, tempMap, MED_BLOCK_CONFIRMS, tempDecay));
+            std::unique_ptr<TxConfirmStats> tempFeeStats(new TxConfirmStats(tempBuckets, tempMap, MED_BLOCK_PERIODS, tempDecay, 1));
             tempFeeStats->Read(filein, nVersionThatWrote, tempNum);
             // if nVersionThatWrote < 139900 then another TxConfirmStats (for priority) follows but can be ignored.
 
@@ -884,9 +891,9 @@ bool CBlockPolicyEstimator::Read(CAutoFile& filein)
             if (numBuckets <= 1 || numBuckets > 1000)
                 throw std::runtime_error("Corrupt estimates file. Must have between 2 and 1000 feerate buckets");
 
-            std::unique_ptr<TxConfirmStats> fileFeeStats(new TxConfirmStats(buckets, bucketMap, MED_BLOCK_CONFIRMS, MED_DECAY));
-            std::unique_ptr<TxConfirmStats> fileShortStats(new TxConfirmStats(buckets, bucketMap, SHORT_BLOCK_CONFIRMS, SHORT_DECAY));
-            std::unique_ptr<TxConfirmStats> fileLongStats(new TxConfirmStats(buckets, bucketMap, LONG_BLOCK_CONFIRMS, LONG_DECAY));
+            std::unique_ptr<TxConfirmStats> fileFeeStats(new TxConfirmStats(buckets, bucketMap, MED_BLOCK_PERIODS, MED_DECAY, MED_SCALE));
+            std::unique_ptr<TxConfirmStats> fileShortStats(new TxConfirmStats(buckets, bucketMap, SHORT_BLOCK_PERIODS, SHORT_DECAY, SHORT_SCALE));
+            std::unique_ptr<TxConfirmStats> fileLongStats(new TxConfirmStats(buckets, bucketMap, LONG_BLOCK_PERIODS, LONG_DECAY, LONG_SCALE));
             fileFeeStats->Read(filein, nVersionThatWrote, numBuckets);
             fileShortStats->Read(filein, nVersionThatWrote, numBuckets);
             fileLongStats->Read(filein, nVersionThatWrote, numBuckets);

src/policy/fees.h

@@ -82,6 +82,7 @@ struct EstimationResult
     EstimatorBucket pass;
     EstimatorBucket fail;
     double decay;
+    unsigned int scale;
 };
 
 /**
@@ -93,11 +94,14 @@ class CBlockPolicyEstimator
 {
 private:
     /** Track confirm delays up to 12 blocks medium decay */
-    static constexpr unsigned int SHORT_BLOCK_CONFIRMS = 12;
+    static constexpr unsigned int SHORT_BLOCK_PERIODS = 12;
+    static constexpr unsigned int SHORT_SCALE = 1;
     /** Track confirm delays up to 48 blocks medium decay */
-    static constexpr unsigned int MED_BLOCK_CONFIRMS = 48;
+    static constexpr unsigned int MED_BLOCK_PERIODS = 24;
+    static constexpr unsigned int MED_SCALE = 2;
     /** Track confirm delays up to 1008 blocks for longer decay */
-    static constexpr unsigned int LONG_BLOCK_CONFIRMS = 1008;
+    static constexpr unsigned int LONG_BLOCK_PERIODS = 42;
+    static constexpr unsigned int LONG_SCALE = 24;
     /** Historical estimates that are older than this aren't valid */
     static const unsigned int OLDEST_ESTIMATE_HISTORY = 6 * 1008;
 

src/rpc/mining.cpp

@@ -895,6 +895,7 @@ UniValue estimaterawfee(const JSONRPCRequest& request)
             "{\n"
             "  \"feerate\" : x.x,        (numeric) estimate fee-per-kilobyte (in BTC)\n"
             "  \"decay\" : x.x,          (numeric) exponential decay (per block) for historical moving average of confirmation data\n"
+            "  \"scale\" : x,            (numeric) The resolution of confirmation targets at this time horizon\n"
             "  \"pass.\"                 information about the lowest range of feerates to succeed in meeting the threshold\n"
             "  \"fail.\"                 information about the highest range of feerates to fail to meet the threshold\n"
             "  \"startrange\" : x.x,     (numeric) start of feerate range\n"
@@ -932,6 +933,7 @@ UniValue estimaterawfee(const JSONRPCRequest& request)
 
     result.push_back(Pair("feerate", feeRate == CFeeRate(0) ? -1.0 : ValueFromAmount(feeRate.GetFeePerK())));
     result.push_back(Pair("decay", buckets.decay));
+    result.push_back(Pair("scale", (int)buckets.scale));
     result.push_back(Pair("pass.startrange", round(buckets.pass.start)));
     result.push_back(Pair("pass.endrange", round(buckets.pass.end)));
     result.push_back(Pair("pass.withintarget", round(buckets.pass.withinTarget * 100.0) / 100.0));

src/test/policyestimator_tests.cpp

@@ -99,13 +99,10 @@ BOOST_AUTO_TEST_CASE(BlockPolicyEstimates)
             BOOST_CHECK(origFeeEst[i-1] <= origFeeEst[i-2]);
         }
         int mult = 11-i;
-        if (i > 1) {
+        if (i % 2 == 0) { //At scale 2, test logic is only correct for even targets
             BOOST_CHECK(origFeeEst[i-1] < mult*baseRate.GetFeePerK() + deltaFee);
             BOOST_CHECK(origFeeEst[i-1] > mult*baseRate.GetFeePerK() - deltaFee);
         }
-        else {
-            BOOST_CHECK(origFeeEst[i-1] == CFeeRate(0).GetFeePerK());
-        }
     }
     // Fill out rest of the original estimates
     for (int i = 10; i <= 48; i++) {
@@ -177,7 +174,7 @@ BOOST_AUTO_TEST_CASE(BlockPolicyEstimates)
         block.clear();
     }
     BOOST_CHECK(feeEst.estimateFee(1) == CFeeRate(0));
-    for (int i = 2; i < 10; i++) {
+    for (int i = 2; i < 9; i++) { // At 9, the original estimate was already at the bottom (b/c scale = 2)
         BOOST_CHECK(feeEst.estimateFee(i).GetFeePerK() < origFeeEst[i-1] - deltaFee);
     }